Date: 22 September 1981 18:31-EDT From: Richard M. Stallman To: INFO-EMACS-RECIPIENTS at MIT-AI Correction: EMACS manuals cost $3.25, not $3.75  Date: 21 September 1981 18:45-EDT From: Richard M. Stallman Subject: How to order EMACS manuals To: INFO-EMACS-RECIPIENTS at MIT-AI Send 3.75 per copy to Publications Department Artificial Intelligence Lab 545 Tech Square Cambridge, MA 02139 Be sure to say whether you want the ITS version or the Twenex version.  Date: 20 September 1981 02:37-EDT From: Richard M. Stallman To: INFO-EMACS-RECIPIENTS at MIT-AI Has anyone found the file EMACS;EMACS GUIDE useful lately? I speculate that it may be so long that it is not worth while for anyone to actually print it out, that it is always easier for everyone to get a printed manual, and that the file can be flushed. If you think the file is worth keeping, please let me know.  Date: 20 September 1981 02:30-EDT From: Richard M. Stallman Subject: Changes in EMACS 162, now NE on AI. To: INFO-EMACS-RECIPIENTS at MIT-AI *) You can now specify which command should exit an incremental search (and not do anything else when it exits the search) by setting the variable Search Exit Char to the desired 9-bit character. The default is Altmode. Commands whose definitions are to indirect to the specified character also exit the search. C-@'s default definition is now indirect to C-Space, so if you set Search Exit Char to 300 octal, both C-@ and C-Space will be search exit characters. I urge everyone to try out the use of C-@/C-Space as exit character, because eventually I'll conduct another poll and perhaps change the default. Try any other characters you might like, also. *) The new command Correct Spelling corrects spelling through the entire buffer. The new command Command to Spell passes its argument to the SPELL job as a command. *) The major modes for various assemblers, such as MIDAS, PALX, FAIL and MACRO-10, now each run their own mode hooks. Previously they all ran MIDAS Mode Hook. *) The completing reader has various extensions intended to allow completion on more kinds of things. Completion of buffer names and variable names may be available soon as an optional library. *) The new hook Exit to Superior Hook is executed before EMACS returns to its calling job. Unlike Exit Hook, this new hook is in addition to all normal actions.  Date: 14 September 1981 18:32-EDT From: Richard M. Stallman Subject: Search terminators To: INFO-EMACS-RECIPIENTS at MIT-AI Since searches already set the mark, why not use C-@ to terminate? Several people have said things like this, and I don't know what to make of it. Searches do (often) set the mark -- because (we hope) that mark is going to be useful. My best guess for what these people are saying is that there is no need for a special character to terminate the search, because the C-@ command can be used; the fact that it sets the mark as well is unimportant. However, this would make the mark left behind by the search inaccessible. Were you overlooking this? Or are you saying what you are saying because you think the mark left by the search itself is not useful? If so, do you think the search should not set the mark? (That can be turned off by setting Auto Push Point Option to a very large number.) Several other people suggested using ^G to terminate the search. Since ^G is now the way to abort a search, it seems this would either require a different way to abort a search (what would that be?) or else there would be no way to abort a search. C-] has also been suggested. I think that is asking for trouble: the same sort that Altmode causes, but more serious.  Date: 12 September 1981 05:29-EDT From: Richard M. Stallman Sender: RMS0 at MIT-AI Subject: Slight correction To: INFO-EMACS-RECIPIENTS at MIT-AI The new meaning of C-O that I am considering is what is now done by C-A C-O Tab, not C-A Tab C-O as I said.  Date: 12 September 1981 03:11-EDT From: Richard M. Stallman Subject: poll on C-S, C-O, and Return To: INFO-EMACS-RECIPIENTS at MIT-AI What do you think of these two suggested changes to EMACS? 1) Make Altmode not be ignored when it terminates a search, so that it would still function as a Metizer. Some other character would have to be provided instead as a way of terminating a search and doing nothing else. C-D was suggested, but I don't like it. I'd like to hear other suggestions. 2) Make C-O do what C-A Tab C-O does now. 3) Also, how many people now find it useful or desirable that Return eats up blank lines? This feature was invented in the days when terminals did not have insert/delete line.  Date: 6 September 1981 21:52-EDT From: Devon S. McCullough Subject: data compression and redundancy, or, why SUPDUP is like a finely crafted buggy in an age of autos. To: INFO-EMACS at MIT-MC, INFO-MACSYMA at MIT-MC Waiting for my screen to fill at 300 Baud takes so long, I forget what I was about to do! I like the idea of Huffman encoding the output through a STY. This requires knowledge of data characteristics on the character level. A simple frequency counting program can get this kind of knowledge. Encoding schemes that use knowledge of the data on the highest level should be even better. Intuition is the only way I know to get this level of knowledge. I am hacking a system where the host and the terminal both know about jobs and data contained in jobs. The terminal is a microComputer with "virtual memory" which gets "paged" in from the current job at 300 Baud, perhaps using Huffman encoding. Instead of filling the screen itself, the host sends a brief message saying "put datum X on the screen now". If datum X is not locally stored, the terminal will reply "send text of datum X", and keep a copy of datum X for possible future re-display. Also, the system could anticipate the user. Thus EMACS could send me more than the current screenful when I select a buffer, on the reasonable assumtion that I will want to see it soon. I suspect that anticipating transmission and eliminating repeated re-transmission wins more interaction speed than Huffman encoding does. All display is done locally. This means EMACS sends portions of it's buffers, not display codes, and MACSYMA sends craftily encoded SEXPR's, not pretty human-readable stuff. So the terminal has to know about EMACS and MACSYMA internal formats and display algorithms to be useful here. Now (at last) my questions: How can I deal with unsolicited type-out, without doing :GAG and ^_R? What are the display algorithms for EMACS and MACSYMA? What other programs or classes of programs should the terminal know about? Has anyone done (or thought about) anything similar? -- Devon -- [ps: there's a copy of this in MC:GUEST5;DEVON FLAME]  Date: 25 August 1981 02:51-EDT From: Richard M. Stallman Subject: Many new features To: INFO-EMACS-RECIPIENTS at MIT-AI These features are mostly due to people at the Canadian Defense Research Establishment Atlantic. MAZLIB is cleaner now, and well documented. Try it. MAZE.MAZE is the new data file for the maze. The SORT functions sort in reverse order if given a numeric arg. FORTRAN mode now provides for files with and without tabs after statement numbers, and for converting between the two formats. Also, it can insert or remove sequence numbers. It contains a Fortran Data mode for editing files of fixed-column data. SAIL mode now has an indenter, based on that of PASCAL mode. The new library MODE2 provides for having two mode lines: the actual EMACS mode line and one additional line stolen from the text area. You can select various options for information to display in that area. The new library NCOLUMNS can take single-column text and turn it into any number of columns, but cannot turn it back. This is different from the library COLUMNS, which can turn single-column text into two-column text or the opposite, but cannot handle more than two columns. Perhaps someone would like to write a new COLUMNS which replaces both of these. A new hook, Exit to Superior Hook, is executed before EMACS returns to its superior job. Twenex Only changes: The new library BUGHUNT modifies the comment commands to put your name at the front of each comment you edit. The new library MOVE contains functions for moving pages or paragraphs between files, supposedly more convenient than doing so by killing and switching buffers. The new library PERSONAL contains functions for keeping track of your projects, phone conversations, and notes to yourself. SYSTEM uses the superior EXEC if possible rather than an inferior, and has new functions Check Job, Check System Job, Check Log File, Logout, Enable, and Disable. EFORK: ^R Invoke Inferior can be called as a subroutine. The new library FDB is useful for examining occult parts of the status of a file, on Twenex. There is now a BASIC20 mode, and also a PCL mode for editing command files. The new library TALK contains functions for initiating and accepting links with other users. Only for Twenex. The VT132 terminal is supported. There is a library NVT132 for it analogous to NVT52 and NVT100. Two new hooks, Exit to Inferior Hook and Return from Inferior Hook, are executed respectively before EMACS runs an inferior and after EMACS starts up again. Not all invocations of inferiors call these hooks; those which are expected to be momentary and involve no keyboard input generally do not invoke the hooks.  Date: 21 August 1981 04:37-EDT From: Richard M. Stallman To: INFO-EMACS-RECIPIENTS at MIT-AI Who wrote the LABELS library? Where is the source?  Date: 14 August 1981 01:14-EDT From: Richard M. Stallman Subject: M-Y, keyboard macros To: INFO-EMACS-RECIPIENTS at MIT-AI Meta-Y now preserves the order of point and mark. If point was at the beginning of the un-killed text before M-Y, it will be at the beginning after the M-Y. C-U C-X ( now appends to the last defined keyboard macro, repeating its definition at the beginning. It is just like typing C-X ( and then retyping the previous keyboard macro definition yourself, but omitting the C-X ), so that you can add to the end of the definition.  Date: 6 August 1981 03:40-EDT From: Richard M. Stallman Subject: EMACS 161 To: INFO-EMACS-RECIPIENTS at MIT-AI EMACS 161 is now NE on AI. It has these new features: The defun commands, C-M-A, C-M-E and C-M-H, now believe that a defun is any character whose Lisp syntax is "(", appearing in column 0. It does not have to be literally a "(". The names for the built-in TECO functions are always accessible. You do not have to load the BARE library by hand to use them. All error messages pertaining to auto saving now say which file the problem was encountered with. In an incremental search on a printing terminal, you can type C-L to reprint the line which the search has reached so far, and it will not exit the search. On Twenex, a file's major mode will be defaulted from its extension even if the file has a local modes list (which fails to specify the major mode).  Date: 6 August 1981 03:40-EDT From: Richard M. Stallman Subject: EMACS 161 To: INFO-EMACS-RECIPIENTS at MIT-AI EMACS 161 is now NE on AI. It has these new features: The defun commands, C-M-A, C-M-E and C-M-H, now believe that a defun is any character whose Lisp syntax is "(", appearing in column 0. It does not have to be literally a "(". The names for the built-in TECO functions are always accessible. You do not have to load the BARE library by hand to use them. All error messages pertaining to auto saving now say which file the problem was encountered with. In an incremental search on a printing terminal, you can type C-L to reprint the line which the search has reached so far, and it will not exit the search. On Twenex, a file's major mode will be defaulted from its extension even if the file has a local modes list (which fails to specify the major mode).  Date: 17 July 1981 23:14-EDT From: Richard M. Stallman To: INFO-EMACS-RECIPIENTS at MIT-AI DEKKER would like to hire someone to do some TECO hacking for his company as a consultant. It relates to PASCAL mode.  Date: 25 Jun 1981 0737-PDT From: KENNARD at SRI-KL Subject: emacs, vt100 To: info-emacs at MIT-AI cc: kennard at SRI-KL i would like to be placed on the mailing list if possible? any info on vt100 wide screen usage and keypad definition would also be helpful. are there any available init packages for the vt100 kennard @ KL -------  NGAI@MIT-MC 06/24/81 23:08:04 To: INFO-EMACS at MIT-MC I understand this is a mailing list for people interested in Emacs. If so, please add my name to the list. Is there a package for VT100 keypads? thanks  Date: 30 April 1981 01:19-EDT From: Richard M. Stallman To: INFO-EMACS-RECIPIENTS at MIT-AI Is anyone interested in doing work on the search command in the PAGE library? I hear it could use some work, but I don't want to work on it.  Date: 18 April 1981 02:35-EST From: Richard M. Stallman Subject: M-$ ^R Correct Word Spelling To: INFO-EMACS-RECIPIENTS at MIT-AI The new M-$ command, implemented by Klotz, calls the spell program on the word before or around the cursor. If the spell program thinks the word is misspelled, EMACS prints a list of suggested correct spellings, numbered with digits, and you can ask to replace the word with one of them by typing the corresponding digit. For more info, see the function's self-documentation.  Date: 10 April 1981 16:32-EST From: Kent M. Pitman Subject: Updates to Emacs MODLIN library To: INFO-EMACS at MIT-AI If you are a user of the MODLIN library (defines an alternate Mode Line handler), you'll note some slight differences ... * The presence or absence of the keyword `NoWrite' in the mode line, which used to reflect the setting of Inhibit Write, has been discontinued. Instead, Visit File Save Old is reflected as follows... SaveOld = Emacs will automatically write the file under these conditions. (Visit File Save Old is negative) NoSaveOld = Emacs will assume without asking that this buffer is not something you want saved even if it is modified when you try reading a new file into this buffer. (Visit File Save Old is 0) (nothing) = Emacs will query if you try to read a new file into this buffer and buffer is modified. (Visit File Save Old is positive) * If buffer bounds are narrowed, the keyword `Narrow' will appear. * The occurrence of the keyword `ReadOnlyFile', `ReadOnlyBuf' or `ReadOnlyFile/Buf' signifies that the file, buffer, or both are readonly. * When visiting a ">" or "<" file, or a ".0" or ".-2" file on non-ITS, the true version number of the file will be shown after the filename as something like "=nnn". Eg, if KMP;FOO 3 were read in by naming FOO >, the mode line will display "KMP;FOO > =3" * Redundant redisplays of the modeline (ie, redisplay where nothing has changed to warrant a redisplay) should be less frequent. These changes are installed on ITS. Note to offsite Emacs maintainers: The source file MODLIN.EMACS.63 exists on XX for exportation to non-ITS sites, but is not installed as MODLIN.:EJ yet, since it will not run in versions of Emacs older than Emacs.154 and Emacs.150 is still the standard. When the new version of Emacs comes up on XX, MODLIN.:EJ.63 should be generated for use. -kmp  Date: 10 APR 1981 0158-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: default for Auto Save Filenames on ITS To: INFO-EMACS-RECIPIENTS at MIT-AI The default FN1 for auto save filenames on ITS is now _SAVnn, where nn is the buffer number of the buffer being saved. The buffer number is the number printed by C-X C-B at the beginning of the line describing the buffer. This is so that auto saves from different buffers do not clobber each other. In older versions of EMACS, which do not have the feature just installed to insert the nn, the default will be _SAV00. If you set Auto Save Filenames yourself you will not see the change unless you set the FN1 to something six characters long ending in 00.  Date: 7 APR 1981 0239-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Name changes coming up To: INFO-EMACS-RECIPIENTS at MIT-AI M-X Occur -> M-X List Matching Lines M-X How Many -> M-X Count Occurrences M-X Keep Lines -> M-X Delete Non-Matching Lines M-X Flush Lines -> M-X Delete Matching Lines M-X View Available Space -> M-X What Available Space The old names (on the left) will still be accepted.  Date: 6 Apr 1981 2117-EST From: WJN at MIT-DMS (Wayne J. Noss) To: INFO-EMACS at MIT-AI Cc: RobG at MIT-DMS Subject: RobG's inquiry Message-id: <[MIT-DMS].192991> I have answered it.  Date: 6 Apr 1981 1932-EST From: ROBG at MIT-DMS (Rob F. Griffiths) To: info-emacs at MIT-AI Message-id: <[MIT-DMS].192972> I have a few questions about things that I could not find in my manual: A) How do I set a TAB-STOP? (Say, to 5 places, as opposed to 8?) I couldn't find anything on that. B) How does one go about writing a library, and where is there a list of available directories? C) Where does one learn to use the minbuffer? (And WHAT is it used for?) D) Say I am trying to make a table, with a top line that looks something like this: :----------|----------|--------------|--------------|------: How could i repeat that design throughout the file, without re-typing it each time? E) What is this picture mode, that I have heard of? F) If I am looking at a file, and it goes over the last column on my screen, and wraps around, is there anything I can set to auto-matically drop the words and put them at the beginning of the next line? (My wrap around prints a '!' and I would like it to be all nice and neat. Thanx; -Rob.  Date: 30 MAR 1981 0349-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 155 (now NE on AI) To: INFO-EMACS-RECIPIENTS at MIT-AI *) The function names ^R Set Bounds Region, ^R Set Bounds Full, and ^R Set Bounds Page have been changed. The new function names are: ^R Narrow Bounds to Region, ^R Widen Bounds, and ^R Narrow Bounds to Page. The old names still work, for now. The purpose is to make the functions show up when expected in an Apropos. *) You can now turn auto save mode on and off safely by setting the variable Auto Save Mode. This is compatible with all other minor modes. *) The mode line says "narrow" when you have narrowed down.  Date: 11 MAR 1981 2356-EST From: KMP at MIT-AI (Kent M. Pitman) Sent-by: RMS at MIT-AI Subject: New Lisp Indentation Format To: INFO-EMACS-RECIPIENTS at MIT-AI In Emacs 154 (now NE on all sites), a new Lisp indentation style is in use. The changes have been provided for the following reasons: * To increase the ease with which stylized ``LispMachine-style'' Lisp indentation can be obtained in a user-programmable way. * To lessen loss of screen space due to indentation. This indenter is much more conservative about the amount of indent it gives to many kinds of expressions. * To enhance error checking. Many kinds of Lisp forms (DO, PROG, LET, LAMBDA, ...) have some number of leading forms, treated specially, followed by some trailing forms, called the body. Common forms of this type will now have expressions in their bodies indented slightly differently than their leading expressions to accentuate the separation. eg, (DO ((X 0 (1+ X))) ((= X 10.)) (PRINT X)) What follows is information about how customized indentation. * How Indentation is Decided The old indentation algorithm is first run, and a tentative indentation is computed. Then a hook is run to allow selection of a different indentation if appropriate. * User-Level Variables Lisp ... Indent where the '...' is replaced by the name of a Lisp operator, defines a special indentation method for that operator. Eg, Lisp PROG Indent controls how PROG's indent. The value of variables may be a string, in which case they will be called to compute the indentation (argument conventions and return values to be discussed later) or a number indicating one of several common operations. n description -4 or less undefined -3 don't indent any special way unless the value of Lisp Indent DEFanything dictates that we should. This is the default for unknown functions. -2 don't indent any special way. -1 indent as a DEF-form. If this is the 2nd line of the definition, indent specially. Otherwise, indent under preceding line. 0 or more indent as a special form, assuming that this form has n leading expressions before its body. (eg, LET and LAMBDA have 1 special expression for the bound variable list, PROGN has none, DO has 2). These special fields are indented normally. The first expression following these special ones is indented specially. Expressions after that are indented under the previous line as normal. The variable Lisp PROG Indent is initially defined to run code so that PROG tags can line up right-justified. Variables corresponding to several of the common Lisp special forms are set up with non-negative values corresponding to the indentation style suggested by their semantics. Lisp Indent DEFanything, if nonzero, will cause all operator names beginning with 'DEF' to default to definitional form style indentation (corresponding to a Lisp DEF... Indent variable of -1). Setting this to 0 will cause 'DEF...' operators to be indented specially only if they have associated Lisp ... Indent variables set up to cause such special indentation. Lisp Indent Offset (default 0), if nonzero, indent this many columns past the most recent unclosed "(" rather than worrying about indentation relative to arguments (this also bypasses special form hackery). If 0, normal Lispy indentation rules apply. NOTE: Although this variable has been present for a long time, it has changed slightly. It used to measure offset from the right side of the paren. Now it measures offset from the left of the paren. The change was made to provide greater flexibility since formerly it was impossible to say you wanted to indent just after the hanging open paren. Lisp Special Indent Offset (default 2), is the offset from the hanging "(" which special form indentation uses. Eg, with the default setting, Emacs will indent under the "E" of "DEFUN". * Low-Level Variables (primarily of interest to Emacs implementors) Lisp Indent Language contains a string (default "LISP") which is used to decide which lisp-like language is to be used. Other hooks key off of this string; ie, if the Indent Language is LISP, variables like LISP Indent Offset and Lisp PROG Indent will be looked for. If the language were Muddle, variables such as Muddle Indent Offset and Muddle PROG Indent would be looked for instead. Currently, Muddle mode doesn't use this feature, but perhaps it will in the near future. In the remainder of this description, only Lisp ... variables since they are the only ones which are currently useful to most users, but implementors should bear in mind that similar variables for other dialects may be defined if a local variable Lisp Indent Language is created with the appropriate dialect name as its contents. Lisp Indentation Hook holds the function which tries to determine how special forms should be indented. All of the effects described farther down in this description are dependent on its containing the default value. If you do not like the action of this hook, you may wish to put a null string in this variable to keep it from doing anything -- effectively nullifying most of the changes described in this text. * Extension Conventions (for sophisticated users only) Lisp Indentation Hook (or Muddle Indentation Hook or ...) must be defined and must contain a string. Its args (,) are the position in the buffer of the most recent unclosed "(" and the buffer position of the beginning of the line to be indented. The buffer position of the cursor at the time the tab was typed is stored as an offset from Z in qZ, so Z-qZ is that buffer position. The hook should NOT modify the buffer. If it returns 0 or no value, it will be assumed that the horizontal cursor position (given by fsshpos) is to be used as the indetnation level. Hence, to indent under the "O" in PROG, it is sufficient to jump to that character in the buffer and return. Replacement of the pointer to the appropriate line will happen automatically later. If it returns nonzero, that value will be assumed to be the desired indentation. Returning negative values is undefined. Lisp ... Indent variables containing strings to be executed observe the same argument conventions as the Lisp Indentation Hook would.  Date: 2 MAR 1981 1235-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Changes in EMACS 154 (now NE on AI) To: INFO-EMACS-RECIPIENTS at MIT-AI Incompatible changes (indicated by two *'s at beginning of item): EVARS file and local mode list format, Visit File Save Old vs. Inhibit Write, Meaning of argument to C-X C-V or C-X C-F, Meaning of Exit Hook, Format of Prefix Char List. *** Use *** **) Visit File Save Old now controls whether visiting a new file causes the old one to be saved. If this variable is zero, the old file is not saved. If it is -1, then the old file is automatically saved. If the variable is 1, then you are asked whether to save. The default value is 1. This variable replaces the old distinction between C-X C-R and C-X C-V, which are now identical commands. The variable Inhibit Write is no longer used for anything, though it still exists to keep some old programs from bombing. Find File Inhibit Write no longer exists at all. **) True read-only visiting is now implemented. An argument to the file-visiting command (C-X C-V or C-X C-F) is used to request this. Either the buffer or the file can be read-only. A read-only buffer means that you can't insert or delete anything. A read-only file means you can change the text in the buffer, but it will not be saved in the file unless you insist. To insist, you type C-X C-S and answer "Y" to confirm; absolutely nothing else will cause the read-only file to be saved. Normally what you get is a read-only file, whose buffer can be edited. The command C-X C-Q (^R Set Read-Only File) can be used to switch between the three modes (ordinary, read-only buffer, and read-only file). *) New self-documentation for recursive editing levels. When you are inside a recursive editing level, Help ? mentions this and tells you how to abort and how to exit, as well as listing the Help options available. A new Help option, Help R, gives the full description of the command that got you into the recursive editing level. *) If the variable Auto Save All Buffers is nonzero, each auto save will save whichever buffers need to be saved. It is up to you to make sure that the buffers are saved in different places, either by using Auto Save Visited File, or by giving each buffer a different value of Auto Save Filename (perhaps automatically in a Buffer Creation Hook). *) EMACS remembers the current version number of each visited file, and displays this in the mode line and in the printout from List Buffers. *) The Delete File command asks for confirmation if it is not sure that you know what default filenames were applied to the filename you specified. *) M-X View Available Space prints out how many pages of space are available in EMACS for more buffers or more libraries. *) The character "#" now has the syntax of a prefix character in Lisp mode. *) ^R Indent for Comment now does not move comments that start with a single semicolon in column 0. *** Customization *** **) The format of EVARS files and local mode lines has been changed to allow variable values to be multi-line strings. The value specified for a variable should now be either a number or a string enclosed in doublequote characters. Indentation before the number or before the string is ignored. To include a doublequote character in the contents of the string, quote it with a  character. Except for peculiar cases, old EVARS files and local modes lists still work. Peculir cases include string values which happen to begin with spaces followed by a doublequote, or spaces followed by a digit. The change applies only to values of variables. Lines which specify definitions for command characters are still single lines of TECO code. *) So that certain common forms of customization will not need to use FS flags, there are now variables which you can use instead of certain FS flags which people often set. Setting the variable sets the FS flag. Setting the FS flag explicitly also works, but the value of the variable does not change, so it is best not to mix the two styles of use. Here is a complete list of variables which control FS flags. Not all of these are new. Auto Save Interval FS ^R MDLY Bottom Display Margin FS %BOTTOM Case Search FS BOTHCASE Cursor Centering Point FS %CENTER Display Mode Line Inverse FS INVMOD Display Overprinting FS ^H PRINT and FS ^M PRINT Echo Area Height FS ECHO LINES End of Buffer Display Margin FS %END Error Messages in Echo Area FS ECHO ERRORS Fill Column FS ADLINE Overwrite Mode FS ^R REPLACE SAIL Character Mode FS SAIL System Output Holding FS TT PAGE (Twenex only) Top Display Margin FS %TOP *** Extension *** **) The format of the value of the variable Prefix Char List is changed. This should only affect people who create their own prefix characters. The purpose of the change is to allow for prefix characters whose dispatch tables are kept in variables rather than in q-registers. The new format of the value is: any number of lines, with a CRLF after each line, one line per prefix character. Each line contains the name to be printed for the prefix character, followed by two spaces, and the TECO expression which evaluates to the dispatch table for the prefix character. Look at the initial value of this variable if this is not clear. **) & Exit EMACS now calls the Exit Hook in addition to auto saving and optional explicit saving. The Exit Hook used to override them. *) The TECO feature FS SUPERIOR is now fully exploited by EMACS. Superiors can tell an EMACS to make space in the buffer, or to visit a file, by following the protocol described near the end of TECORD. *) M-X Describe can be given a precomma argument, which should be the character number of a ^R command being described. This is passed along to any documentation-macros contained in the documentation string. *) FS TAB WIDTH now controls the distance between tab stops on display. By setting this to 10, you can view a file the way it would look on Multics. *) On Twenex, Return from Inferior Hook is called on return from an inferior if Exit Hook is called before starting the inferior.  Date: 21 FEB 1981 0426-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Currently, Autoarg mode means that digits typed before a control or meta character count as an argument for it, but digits followed by a non-control non-meta character are inserted. Should Autoarg mode make the "-" character an argument character as well as the digits? Answers will go in EMACS;AUTARG ANSWER.  Date: 19 Feb 1981 1531-PST From: Ted Markowitz Subject: DT80 question To: info-emacs at MIT-AI Does anyone know how to get EMACS and a DT80 (the VT100 lookalike) to make up and be friends? At 9600 baud the XON-XOFF of the terminal loses badly (Search: ...argh). I checked the INFO-TERMS on this thing, but it didn't suggest a reasonable fix. Any thoughts? --ted -------  Date: 6 FEB 1981 1706-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: FS FLAGS To: INFO-EMACS-RECIPIENTS at MIT-AI In order to save beginners from having to know about FS flags, I'm thinking of creating EMACS variables to control those FS flags which a user might want to change in simple customization. I don't want to do it for all FS flags. That would be costly, and many FS flags are not useful except for people who are writing TECO code anyway. Can you tell me which FS flags you think should be candidates for this?  Date: 3 Jan 1981 (Saturday) 2106-EDT From: OTTO at WHARTON-10 (George Otto) Subject: C100 EMACS keypad library To: Info-emacs at MIT-AI, C100-Fans at MIT-AI cc: OTTO (George Otto) Has anyone implemented an EMACS library to utilize the keypad and function keys of the Concept-100 terminal? (up-arrow key moving the cursor up one line, etc.) George Otto  Date: 3 Jan 1981 1646-EST From: Jeff Shulman Subject: H19 EMACS PACKAGE To: BARMAR at MIT-MC cc: INFO-EMACS at MIT-MC THERE IS A GOOD EMACS H19 PACKAGE WRITTEN BY SOMEONE AT DARCOM-KA. I HAVE A COPY OF IT AND YOUR WELCOME TO FTP IT OR SEND ME A MESSAGE AND I WILL MAIL IT TO YOU. ONCE LOADED YOU INVOKE IT VIA M-X HEATH MODE$. IT IS PRETTY GOOD. THE ONLY BUG IT HAS IS WHEN YOU LEAVE EMACS OR PUSH OUT OF IT. WHEN YOU DO THAT TO CLEAR UP THE BUG INVOKE IT AGAIN. THE BUG IS JUST THAT THE DATE-LINE DOESN'T ALWAYS GO ON THE 25TH LINE. IT'S IN MY FILE AS H19.* SO JUST TO A MULTIPLE FTP OF H19.* . ANY PROBLEMS YOU WILL HAVE TO GO TO THE AUTHORS AS I MYSELF AM NOT A TECO HACKER. JEFF -------  Date: 3 Jan 1981 1316-PST From: Vanmelle@SUMEX-AIM Subject: Re: H19 To: BARMAR@MIT-MC cc: DMTGJR@AI, INFO-EMACS@AI In response to your message sent 3 Jan 1981 1301-PST I have a library to use the Heath19 keypad and function keys. To avoid conflicts with "regular" emacs keys, I implement them as prefix characters on ESC and ESC-?, and designate one of the keypad keys as a Meta-prefix to replace ESC. The library lets you assign any functions you want to the keys; the default puts the obvious things on the arrow keys. The library exists as [sumex]heath.*. I'll be happy to send sources and documentation to any requesting them. Bill -------  BARMAR@MIT-MC 01/03/81 15:21:58 To: DMTGJR at MIT-AI CC: INFO-EMACS at MIT-MC MSG: *MSG 440 DMTGJR@MIT-AI 12/21/80 18:24:41 DOES ANYONE KNOW IF THERE IS AN INIT FILE FOR EMACS THAT MAKES USE OF THE H19 KEYPAD AND ITS CURSOR CONTROLS. IF NOT, DOES ANYONE WANT TO WRITE ONE? FROM WHAT I READ IN THE DOCUMENTATION IT SHOULDN'T BE HARD FOR A GOOD TECO HACKER. THE H19 USES SIMPLE ESC (META) CODES. OR COULD IT EVEN BE DONE IN A EVARS FILE? THANX Unfortunately, I believe that the codes that the H19 special keys use would overide some of Emacs' standard key bindings (I think it uses ESC-? as its prefix). It could be designed, however, such that if the character following the ESC-? is not a valid H19 code, it is passed on to the Describe function. Barmar  Date: Tuesday, 9 December 1980 18:51-EST From: John A. Pershing Jr. To: INFO-EMACS at MIT-AI cc: EMACS-CONTACTS at MIT-AI Subject: New MKDUMP library A new version of the MKDUMP library exists on XX, as well as an INFO node. It is incompatible with the previous version in a number of ways: - The command M-X Dump My EMACS has been renamed M-X Dump EMACS. - The default dumped filename has changed from MYEMACS to DEMACS. - The variable Dumped EMACS Startup Time has been flushed. - The command M-X Prepare for Loading Libraries has been flushed in favor of . Bugs to JPERSHING@BBNA. EMACS maintainers note: this version depends on the most recent default init file also being installed in the EMACS directory. -jp  Date: 26 Nov 1980 1426-EST From: Wayne J. Noss Subject: Re: Algol mode To: LS.Barmar at MIT-EECS cc: info-emacs at MIT-MC In-Reply-To: Your message of 26-Nov-80 0253-EST Barry, I agree that it would be nice to have ALGOL mode do better things and that if someone implemented the features you mention, it would be put to good use. The problems are that the implementation would be of nontrivial difficulty and that someone would have to devote time to it. BEGIN and END cannot be recognized by TECO as brackets, as I understand it, so would have to be much hairier than for, say, LISP, in order to work right. (Besides which, there is no universal agreement about indentation anyway.) While I would be willing in principle to work on it, there is NO WAY I could assume such a responsibility at the present time, and I don't know when that situation will change. If you can find someone else to do it, I'll be pleased. Sorry. the WJN -------  Date: 26 Nov 1980 0253-EST From: Barry Margolin Subject: Algol mode To: ls.wjn at MIT-EECS cc: info-emacs at MIT-MC What if anything does Algol mode give me other than the fact that it says "Algol" on my mode line and binding TAB to indent-relative? I am trying to use it to edit my 6.001 labs, and it would be nice if I had an environment full of useful functions for editing Algol code, i.e. having TAB bound to the a function that knows how Algol code should look, and an Indent Algol Region command. I realize that there is not much demand these days for a hefty Algol programming environment, but the fact remains that every semester dozens of 6.001 students will have to use a crufty Algol Mode. I thought that "Emacs Makes ALL Computing Simple". Along the same lines goes the PASCAL Mode that the 6.030 students are using. As a 6.030 UTA, I find that it would make the structure of programs much clearer to the students if PASCAL Indent Region did the right thing. It doesn't seem to know about the PASCAL keywords and how they affect the formatting of a program. I have never successfully formatted a PASAL buffer with this; I usually have to step down the lines and use TAB, correcting it when it gets it wrong. Barry Margolin -------  Date: 18 NOV 1980 0321-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI The EMACS Manual For TWENEX Users is now out. You can order one by sending $3.25 to Publications Department Artificial Intelligence Lab 545 Tech Square Cambridge, MA 02139 Ask for AI memo 555.  Date: 22 OCT 1980 0002-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Lots of people are sending me messages askin me not to change what C-T does at the end of a line. I never had any idea of changing what C-T does at the END of a line. I sent a message asking what people think of changing what C-T does in th MIDDLE of a line. How people got the impression that I was asking about the end of the line, I can't imagine, but all the effort you put into taking about that is a complete waste.  Date: 16 OCT 1980 0242-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Right now C-T transposes the two characters around the cursor, except at the end of the line. In this it is compatible with M-T and C-M-T. However, it might be more useful for C-T to interchange the last two characters, always. Do you think this change should be made? Do you think that M-T and C-M-T should also be changed? Whichever ones were changed, would only be changed in the case when they have no argument. The current definitions of M-T and C-M-T have the property that repeating the command is useful, and is equivalent to giving the command an argument. So there is a conflict between utility of C-T, utility of M-T and C-M-T, and compatibilit between them.  Date: 22 SEP 1980 0249-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI, lyons at DEC-MARLBORO Please mail requests to add or remove names on INFO-EMACS to BUG-EMACS, not to INFO-EMACS, to avoid bothering all the users.  Date: 19 SEP 1980 1634-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: C-U C-X ; To: INFO-EMACS-RECIPIENTS at MIT-AI Just about everyone wants the effect to be permanent, so that's how it will stay. Some way of adjusting one comment without setting the comment column may be provided as well.  Date: 18 Sep 1980 2128-PDT From: KAPLAN at SRI-KL Subject: Heathkit H-19 EMACS package To: bboard at SRI-KL, bboard at SU-SCORE, bboard at SUMEX-AIM, bboard at SU-AI, bboard at MIT-AI, bboard at CMU-10A, bboard at RUTGERS, info-emacs at MIT-AI, bboard at BBNC cc: kaplan at SRI-KL With the aid of Bill VanMelle, I have developed a package that can be included in an EMACS.INIT file that adapts the Heathkit H-19 numeric keypad to common emacs functions. It works with minimal impact on bare emacs, i.e. with no special libraries loaded. Anyone who would like a copy is welcome to have one, by sending me a request, with two caveats: 1) I do not warrantee nor will I support or make changes in it. 2) It currently works with the (very) peculiar 1200/150 split speed modems we use here at Stanford. Specifically, this modem accepts characters from the terminal at 1200 baud, buffers two characters, and send them at 150 baud over the phone line. Since the heathkit keypad sends 3 characters per key, the middle character (the "?") gets eaten consistently by the modem. Consequently, the package looks for and recognizes the first and last letters of the keypad codes. I am sure that there is some simple change that will make the package consistent with more rational modems; however, having learned just enough of the opaque TECO language to implement what I needed, I am not particularly interested in figuring out what changes are required. Currently, the package is conditionally invoked upon executing EMACS if you are on an H-19 and on a 1200-baud line (presumed to be a dial-up). This condition can be deleted if desired. Incidentally, the keypad arrangement has proven quite convenient, and I recommend it to other itinerant high-tech tinkerers. A map follows: ___________________________________________________ | 7| 8| 9| |beginning of file| previous para | end of file | | | | | |beginning of line| previous line | end of line | |_________________|_______________|_______________| | 4| 5| 6| |delete back char | prev screen | delete char | | | | | |back character | next screen | forward char | |_________________|_______________|_______________| | 1| 2| 3| |kill back word | next para | kill word | | | | | |backward word | next line | forward word | |_________________|_______________|_______________| | 0| .| E| | | | | |fill paragraph | un-kill | | | | | | |_________________|_______________|_______________| -Jerry Kaplan Research Associate Computer Science Department Stanford University -------  Date: 18 SEP 1980 2051-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Right now, C-U C-X ; sets the comment column permanently. Do you think this is a good idea, or should it only adjust the comment on the current line and not affect further use of comment commands?  Date: 11 Sep 1980 1407-EDT From: Dave Lyons To: info-emacs at MIT-AI Subject: Addition to the EMACS mail list please Could you please add EMACS@DEC-MARLBORO to the mail lists for bothe the hacker and general mail lists for emacs. lyons@dec-marlboro hess@dec-2136 --------  Date: 6 SEP 1980 0316-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Do you think that it is right for C-U C-X C-F to leave the comment column permanently set?  Date: 17 AUG 1980 0144-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Another change to SCRIBE mode. To: INFO-EMACS-RECIPIENTS at MIT-AI "@;" at the front of the line will also make the line a paragraph-separator, just like @begin and @end. Put "@;" (the SCRIBE no-op command) at the front of any line (such as @quotation[, @example[, or a line of just a closing delimiter) which should not be filled in with the surrounding lines. Also, SCRIBE mode is now defined on ITS as well as Twenex.  Date: 13 AUG 1980 0219-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: SPLIT library for editing very big files. To: INFO-EMACS-RECIPIENTS at MIT-AI There is a new library SPLIT for operating on files too big to load into EMACS. It contains a command Split File which divides such a large file into small enough subfiles, and another command Unsplit File for putting the subfiles back together. On ITS, the subfiles made from FOO are called FOO 1, FOO 2, FOO 3, etc. Unsplit Files looks for the same names to recombine. On Twenex, the file FOO.anything.version is split into FOO.1.1, FOO.2.1, etc. After these have been edited to get FOO.1.3, FOO.2.2 (for example), Unsplit File combines those into a new version of FOO.anything The subfiles are made about 50,000 words long, and are divided at line boundaries. There is no limit to the size of file which can be split.  Date: 10 AUG 1980 0557-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI EMACS 150 has the function Compare Windows, which is what used to be in LSPUTL under the name Window SRCCOM. It's very useful. Also, now you can quit it at any time and the pointers in both windows are advanced as far as the comparison has gone. The character "@" at the front of a line no longer indicates a paragraph gap line by default. In SCRIBE mode (which exists on Twenex), the strings "@begin" and "@end" are recognized as paragraph-separating lines, but not other lines which start with "@". This is so that there is no problem with the many SCRIBE commands that are used within a paragraph.  Date: 31 JUL 1980 0242-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Twenex users: should Reap File and Clean Dir expunge the files they delete? I tend to think so, since their purpose is to make room. In fact I though they did, but apparently they don't. I plan to change it.  Date: 27 JUL 1980 0218-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: printing terminals To: INFO-EMACS-RECIPIENTS at MIT-AI TECO 1008 makes it possible to customize how cursor motion is indicated by EMACS on printing terminals. EMACS will still handle motion within one line or down only one screen line by backspacing or printing the characters moved over. Other cursor motion can be handled by the string of TECO commands stored in FS ^R TTM1. It is up to you to define it. I hope you will come up with useful ideas for what to put there and try them out. Then eventually I will look at what you've done and install something.  Date: 26 JUL 1980 0121-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI The library OUTLIN is now called OUTLINE on Twenex systems. The variable Mail Reader, which holds the name of the library to use for reading mail, is now called Mail Reader Library. On Twenex, there is also Mail Reader Program. Set the former if you use an EMACS library to read mail, or set the latter if you use a separate program to read mail.  Date: 25 JUL 1980 0405-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: LSPUTL library. To: INFO-EMACS-RECIPIENTS at MIT-AI Boyer@SRI has contributed a few useful functions. Window SRCCOM compares the text in window 1 with that in window 2, moving to the next difference. Useful with things other than Lisp also. Find Pat searches for lists that contain several specified strings. ^R Extract Sublist replaces a list with one of its sublists.  Date: 23 JUL 1980 1603-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI The obsolete duplicate function names Alter Options, Edit ..D, ^R Find File, and & Macro Get Full Name are being flushed. The new names (which already work) are Edit Options, Edit Syntax Table, Find File, and & Macro Name.  Date: 23 JUL 1980 0023-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS-RECIPIENTS at MIT-AI Reap File now only accepts one filename to reap. It used to accept any number of filenames, terminated by a null argument. The change is to make it more like most other commands.  KMP@MIT-MC 07/17/80 06:05:06 Re: TeXmac library To: INFO-EMACS-RECIPIENTS The TeXmac library has become obsolete. I have moved all of its macros (all 2 of them) into the TeX library and installed that. You should just load TeX library now (or it will load for you when you enter TeX mode). I will leave the TeXmac library around for about a week so that people's packages don't break (at worst you'll have two copies of the macro in your emacs if you explicitly load the TeXmac library) -- after that time, I'll assume everyone has changed over and will delete TeXmac. For those with short memories, TeXmac had the macro support for automatically generating \vbox's that know how to correctly preserve indentation of variable width font displays. -kmp  Date: 13 July 1980 22:14-EDT From: Eugene C. Ciccarelli Subject: Word Abbrev mode To: INFO-EMACS-RECIPIENTS at MIT-AI There is a new version of the WORDAB library with the following changes. This is now installed on AI, soon to move elsewhere. For details on new commands, see their on-line documentation. * M-X Word Abbrev Apropos has been replaced by M-X List Word Abbrevs with a string argument. M-X Insert Word Abbrevs can also take a string argument. * Keys (C-X C-A, etc.) are connected when the library is loaded, and stay connected outside of Word Abbrev mode: you can define and expand manually if the system is too slow for automatic expansion to be tolerated. Also, there is a new command M-X Expand Word Abbrevs in Region (queries like M-X Query Replace). * New option variable, Save Word Abbrevs, controls whether exiting EMACS will cause abbrevs to be saved if necessary. 1 => save all abbrevs on exit, -1 => just save incrementals, with default 0. * M-X Edit Word Abbrevs lets you re-edit to correct syntax errors. The error messages should be clearer too. * You can have word abbrev "macros", keyboard macros (or Teco code if you prefer) that can be "attached" to normal abbrevs, so that after those abbrevs expand, their attached macros are run. E.g. "textmode" could expand to "-*-Text-*-" and then do a M-X Text Mode automatically. Describe M-X Attach Word Abbrev Keyboard Macro for details. One final question: does anyone use abbrevs that are local to a file?  KMP@MIT-MC 07/12/80 03:53:25 Re: Minor changes to TAGS package To: INFO-EMACS-RECIPIENTS If qTags Find File is set to 2 instead of 1, it now means to force use of Find File always (rather than preferring to recycle the *TagSearch* buffer where possible). This allows a user to override the recent changes to the Tags Find File behavior. You are not recommended to make this change unless you find yourself annoyed by the default behavior. -kmp  Date: Tuesday, 8 July 1980 16:59-EDT From: SWERNOFSKY at BBN-TENEXD To: info-emacs-recipients at mit-ai cc: swernofsky at bbnd Subject: SCRIBE mode for EMACS Emacs people, I would like to hear from people who have written macros for use when editing SCRIBE source with EMACS. If you have any such macros, or know anyone who does, please contact me. If enough people would like to continue discussion on this idea, I will set up a mailing list for the purpose. -- Steve  Date: 1 JUL 1980 1814-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: No more unintended messages to INFO-EMACS. To: INFO-EMACS at MIT-AI Mail sent to INFO-EMACS will now claim to have been sent to INFO-EMACS-RECIPIENTS. That name is a dummy, and replies sent to it will go only to EMACS maintainers. If you want to send a message to all EMACS users, you should still use INFO-EMACS (or INFO-EMACS-NEWS if you are announcing a new feature and want the message to go to EMACS;EMACS NEWS).  Date: 30 Jun 1980 2154-PDT Sender: STEF at DARCOM-KA From: STEF at DARCOM-KA To: RMS at MIT-AI Cc: INFO-EMACS at MIT-AI Message-ID: <[DARCOM-KA]30-Jun-80 21:54:40.STEF> In-Reply-To: Your message of 30 JUN 1980 1911-EDT Happy Birthday Indeed! What an interesting msg to use for generating a reply from South Lake Tahoe to tell you it rained in Yosemite today, and that there is still snow in our campsite at ]toulumne Meadows, so we came on to Tahoe and got a nice Motel with HOT! TUB! Back to camping at Emerald BAy tomorrow. I hope the absence of substantive mail is not caused by my absence! My Next check in will probably be from Mammoth Mountain on July 3. Cheers - STef  Date: 30 JUN 1980 1911-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Answers to my questions should NOT be mailed to INFO-EMACS!  Date: 30 JUN 1980 0956-EDT From: KMP at MIT-MC (Kent M. Pitman) Subject: New Emacs library for TeX users -- TeXmac To: INFO-EMACS at MIT-MC CC: RLB at MIT-MC, INFO-TEX at MIT-MC, RIVEST at MIT-ML I have created and installed a library called TeXmac for use in Emacs to support useful operations on TeX code. Currently, the following macros are all it contains: [1] M-X Setup Indented TeX Display This macro will take a normally indented expression in a variable width region and set it up so that it will be output aligned similarly by TeX. *Only* the beginnings of lines will line up right. Example: The region must be placed around the area that is to be set up. This macro is primarily for setting up indented Lisp code in a variable width font. Requires a \noflash macro to be defined (see [3] below) [2] M-X UnSetup Indented TeX Display This macro will convert a region of the buffer which has had M-X Setup Indented TeX Display run across it and turn it back into normally indented text. The cursor may be placed generally anywhere touching the \vbox{...} which is to be converted back. [3] M-X Insert TeXmac Aux Macros Inserts macro definition for \noflash macro used by the output of M-X Setup Indented TeX Display. If the \noflash macro is found to be of general use to the TeX community, it may be moved into one of the TeX libraries, but for now you'll have to manually include its definition by running this macro at the top of your file. Comments to KMP@MC.  Date: 30 June 1980 10:16-EDT From: "Guy L. Steele, Jr." To: RMS at MIT-AI cc: INFO-EMACS at MIT-AI Date: 29 JUN 1980 0318-EDT From: RMS at MIT-AI (Richard M. Stallman) Suggested: change C-T to read two characters from the terminal and transpose the last place where those two characters appear together. I think this would be more useful than having to move the cursor there. Does anyoen object? I object, but propose that C-U C-T do the proposed operation. I suspect that few people use the numeric arg to C-T (though they may for M-T), or in any case seldom count off as muich as four rather than just repeatedly typing C-T however many times.  Date: 30 JUN 1980 0956-EDT From: KMP at MIT-MC (Kent M. Pitman) Subject: New Emacs library for TeX users -- TeXmac To: INFO-EMACS at MIT-MC CC: RLB at MIT-MC, INFO-TEX at MIT-MC, RIVEST at MIT-ML I have created and installed a library called TeXmac for use in Emacs to support useful operations on TeX code. Currently, the following macros are all it contains: [1] M-X Setup Indented TeX Display This macro will take a normally indented expression in a variable width region and set it up so that it will be output aligned similarly by TeX. *Only* the beginnings of lines will line up right. Example: The region must be placed around the area that is to be set up. This macro is primarily for setting up indented Lisp code in a variable width font. Requires a \noflash macro to be defined (see [3] below) [2] M-X UnSetup Indented TeX Display This macro will convert a region of the buffer which has had M-X Setup Indented TeX Display run across it and turn it back into normally indented text. The cursor may be placed generally anywhere touching the \vbox{...} which is to be converted back. [3] M-X Insert TeXmac Aux Macros Inserts macro definition for \noflash macro used by the output of M-X Setup Indented TeX Display. If the \noflash macro is found to be of general use to the TeX community, it may be moved into one of the TeX libraries, but for now you'll have to manually include its definition by running this macro at the top of your file. Comments to KMP@MC.  Date: 29 June 1980 19:12-EDT From: Earl A. Killian To: INFO-EMACS at MIT-AI The default EMACS init file has been changed to use Find File rather than Visit File to read in the file specified in the command line, if the variable Tags Find File is nonzero.  Date: 29 JUN 1980 1745-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Just about everyone is against the change to C-T, so it won't be made. I admit I'm surprised. There is so little disagreement that I don't think there's a need to store the answers anywhere.  Date: 29 Jun 1980 1237-PDT From: LARSON at SRI-KL Subject: C-T change To: info-emacs at MIT-AI I would object. It is currently consistent with M-T in the same way that C-F and C-B are with M-F and M-B. It also seems less useful. Sometimes I want to transpose two characters in a macro without actually knowing what they will be in general. The proposed change would ruin this feature. Alan -------  Date: 29 JUN 1980 0901-PDT From: THOMPSON at USC-ECL Subject: C-T change To: info-emacs at AI I typically use ^T while typing in. Being a computer user, my typing is fast but notoriously inaccurate. I find that I will realize that i have transposed two characters and i will ^T without even pausing for TECO to display all of the typin. Your suggested change sounds like it will ruin this useful feature. Mark -------  Date: 29 JUN 1980 0318-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Suggested: change C-T to read two characters from the terminal and transpose the last place where those two characters appear together. I think this would be more useful than having to move the cursor there. Does anyoen object?  Date: 28 Jun 1980 1926-EDT From: MALIS at BBN-TENEXE Subject: Using tabs in EMACS on the H19 To: info-emacs at AI, h19-people at AI cc: Malis Is there any way to get EMACS to use the physical tab stops in the H19 (columns 9, 17, 25, etc.) when typing text that contains tabs? I am using an H19 at 300 baud and notice that EMACS types spaces even when the text contains tab characters and the tab stops being used by EMACS are identical to those used by the terminal. Getting EMACS to type the actual tab characters in the text would seem to be a big help when one is limited to 300 baud. Emacs does seem to echo the tabs when the text is first typed in, but whenever it is re-displayed it then types 8 spaces per tab character in the text. Andy -------  Date: 27 June 1980 12:28-EDT From: Kent M. Pitman Subject: Modifications to TAGS To: INFO-EMACS at MIT-MC The following changes have been made in the Emacs TAGS library: People who do 1m.vTags Find File (default 0, same as before) will now find that TAGS search over large tag libraries will not run out of core. Now what happens if Tags Find File is nonzero is a buffer named *TagSearch* is used as a temporary during the search if the file being searched is not already in core. This affects only M-X Tags Search, C-., and M-X Tags Query Replace. (If you do a search that uses this facility, you are advised to use M-X Rename Buffer to change the buffer name if you don't want the next such search to recycle that space.) Doing 0m.vTags Search Verbose (default 1) will suppress the filenames from printing out as M-X Tags Search moves from file to file. M-X Tags Query Replace is not affected, since it is presumably more interactive anyway; if there are enough people who want the same or another flag to affect that macro, as well, it can probably be arranged. Some minor bug fixes. Tags will no longer accidentally err out due to finding something in an invalid field in the tags file. -kmp  Date: 6 JUN 1980 0555-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI PURIFY has been changed to allow blank pages and pages containing just comments in library source files. They are ignored. This should facilitate putting local modes lists in such files.  Date: 31 MAY 1980 2350-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I am working on a questionaire to go to all non-network Twenex sites that have EMACS. Do you have any suggestions for questions to ask? The questionaire text lives in AI:.TECO.;QUESTN >  Date: 19 MAY 1980 0700-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Have you tried journal files? I would like to know if anyone is actually using them, and whether there are any problems.  Date: 16 MAY 1980 1803-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: M-X Delete File To: INFO-EMACS at MIT-AI Do you think it should query "Delete FOO; BAR 69 (Y or N)?" before deleting?  KMP@MIT-MC 05/06/80 00:40:58 Re: MODLIN/TIME feature added To: INFO-EMACS at MIT-MC Users of the MODLIN mode-line handler can do 1 m.vTime Only in an init file if they would rather have only the time and not the date in the mode line. The default is still to get both date and time. For more info, send me net-mail. -kmp  Date: 2 MAY 1980 2139-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I am changing the name of the old EMACS-HACKERS mailing list to EMACS-CONTACTS, and creating a new mailing list EMACS-HACKERS@AI whose purpose is to contain everyone who might be interested in working on suggested EMACS improvement projects. If you know how to program in TECO and think you might be interested in working on medium-sized projects, ask me to put you on the list. Then, if you feel like implementing any of the suggestions you see, go ahead (but let me and the person who made the suggestion know).  Date: 15 April 1980 14:18-EST From: Charles Frankston Subject: How to count words To: RICH at MIT-AI, INFO-EMACS at MIT-AI Date: 15 APR 1980 1238-EST From: RICH at MIT-AI (Charles Rich) DICK@MIT-AI 04/15/80 11:12:13 Dividing the number of characters in the buffer (after the initial R start up stuff) by 6 is remarkably accurate. All such a macro would have to do is just skip over the R commands to improve the situation, since the definition of NUMBER OF WORDS, isn't how many words there happens to be, but the number of letters (ie not spaces or punctuation) divided by 5. Dick For those who would rather count words, this is more or less right, does not count most text justifier commands: bj 0[1 :<.-z; :fwl -3f= ."e l '"# fwl w%1 '>w q1:@= @ft words in buffer :fiw Admittedly, someone should probably put a better written one in Emacs somewhere.  Date: 15 APR 1980 1238-EST From: RICH at MIT-AI (Charles Rich) Subject: More on Counting Words. To: INFO-EMACS at MIT-AI DICK@MIT-AI 04/15/80 11:12:13 Dividing the number of characters in the buffer (after the initial R start up stuff) by 6 is remarkably accurate. All such a macro would have to do is just skip over the R commands to improve the situation, since the definition of NUMBER OF WORDS, isn't how many words there happens to be, but the number of letters (ie not spaces or punctuation) divided by 5. Dick  Date: 15 APR 1980 0928-EST From: RICH at MIT-AI (Charles Rich) Subject: Request for information. To: INFO-EMACS at MIT-AI Has anyone written a macro to count the number of words in a document?? (This should probably skip any lines text processor command lines). It seems like this might be a useful general utility. Thanks, Chuck.  Date: 15 Apr 1980 0819-EST From: Sross at MIT-XX (Sandor Schoichet) Subject: Re: New SLOWLY library To: RWK at MIT-XX cc: info-emacs at MIT-AI In-Reply-To: Your message of 15-Apr-80 0557-EST That sounds like a wonderful new feature. Thanks! Suggestion for the one other thing that bugs me about SLOWLY: Why doesn't the window narrowing happen when doing an incremental search&replace? Sross -------  Date: 15 Apr 1980 0557-EST From: RWK at MIT-XX Subject: New SLOWLY library To: info-emacs at MIT-AI There is a new version of the SLOWLY library, with an improved incremental search. This one only narrows the screen for searching when the item searched for is offscreen. This means that when you search for something onscreen, you don't end up doing a redisplay just to put the item found at the top of the screen. -------  Date: 12 APR 1980 1118-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Paragraphs To: INFO-EMACS at MIT-AI Should indented lines really start paragraphs? Paragraph Delimiter is normally set so that they do. Logically this is the correct decision for unindented text. However, making the paragraph commands ignore indentation would enable them to recognize paragraphs correctly in indented text. This would also benefit the sentence commands, which do not work well on indented text now that they pay attention to all paragraph boundaries. The cost would be that if anyone uses indentation to indicate paragraphs and has no blank lines between them, his paragraphs would not be recognized. However, he could change Paragraph Delimiter back to its old default value. Replies will go in EMACS;PARA ANSWER.  Date: 11 APR 1980 0605-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Journal file feature To: INFO-EMACS at MIT-AI EMACS can now write journal files which record all your commands so that they can be replayed later after a system crash or other lossage. Documentation for this feature can be found in AI:RMS;JRNDOC >. This file is in SCRIBE input format but is probably comprehensible.  Date: 6 APR 1980 0324-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI AI:RMS;GLOSS > is the glossary for the EMACS manual. Please read it and make suggestions, especially suggestions for terms which ought to be mentioned.  Date: 4 APR 1980 2234-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Does anyone ever use the feature whereby a buffer can be referred to by its "buffer number", to select or kill it?  Date: 4 APR 1980 0542-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI In EMACS 145, M-X and C-M-X will prompt for the command name with M-X and C-M-X rather than MM and :MM. this is so that the prompts will be understandable without reference to TECO, which users may not know anything about. These prompts can be overridden with the variables Read Command Prompt and Instant Command Prompt, as before.  Date: 3 APR 1980 0805-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 145 To: INFO-EMACS at MIT-AI EMACS 145 is now installed as NE on AI. For the most part it merely incorporates the patches for 144. However, there are these new minor features: View Buffer now takesthe name of the buffer to be viewed. A null argument means to view the selected buffer, as before. Also, Backspace moves back by screens, as in View File (which is used as a subroutine). A new command, Insert Buffer, inserts the contents of another buffer into the selected one. A new command, Make Space, tries to free up space. It offers to empty out the kill ring, to empty the undo memory, and to delete the EMACS buffers created automatically by TAGS, INFO, RMAIL, etc. (a separate offer for each such buffer). Even if you say "no" to each of these offers, it frees up some internal objects which you will not miss. Backtrace now has a new command, ".", which returns to any frame at any PC. First select the frame to return to, then type ".". You can then position point at the PC to return to, in a recursive editing level. The commands List Commands, List Subroutines and List ^R Commands have been flushed to save space, on the grounds that their output has become so verbose that nobody would want to use them. You can get a similar effect from doing Apropos with a null argument.  Date: 2 APR 1980 0145-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: TAGS feature To: INFO-EMACS at MIT-AI Going to a tag, when it does not have to switch files, now pushes point onto the mark ring first. You can go back with C-U C-@. It prints a notification in the echo area just as searching does.  Date: 31 MAR 1980 2255-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI DON'T mail to INFO-EMACS to ask to be taken off the list. Mail to me ONLY. There is no need to bother all the users with such messages.  Date: 31 Mar 1980 1004-PST From: MAINT.KENNARD at SU-SCORE Subject: PLEASE REMOVE ME FROM THE MAILING LIST To: INFO-EMACS at MIT-AI THANKS -------  Date: 31 MAR 1980 0801-EST From: DKM at MIT-AI (David K. Mellinger) To: INFO-EMACS at MIT-AI Please remove me from the mailing list. Thanks. -dkm  Date: 31 MAR 1980 0446-EST From: GLS at MIT-AI (Guy L. Steele, Jr.) Subject: Proposed new commands, and sundry remarks. To: (BUG ZWEI) at MIT-AI, (BUG EMACS) at MIT-AI CC: GLS at MIT-AI, INFO-EMACS at MIT-AI Comments solicited on the following suggestions (which are NOT from an EMACS or ZWEI implementor necessarily). I suggest that Help mention Undo explicitly, to help the naive user. Perhaps Help U should do Undo. On the other hand, I wish that Undo were by default on some single keystroke. I wish that some single keystroke (on a Knight keyboard) would do C-X B Return. This is a VERY common operation (I assert; has any thought been given to collecting statistics on which ^R commands (expecially C-X ones) are invoked most often, for possible redistribution to easier-to-reach keys?). Also, the Stanford E editor has an extremely winning feature whereby all the buffers (actually files) are ordered on a pdl, and C-H exchanges the top two on the PDL (like C-X B), but C-M-H rotates the top k (default k=3), n C-H rotates the top n, and n C-M-H rotates the top n and sets k to n. This allows easy running among three or more files. Specifying a file explicitly is like saying n C-H, where the file is n'th from the top of the pdl. Perhaps similar operations for the point/mark pdl and kill ring would also be useful. I want to yell some more for C-M-O => C-O, and C-M-O = ^R Stack List Vertically. I propose some commands which I believe would make LISP editing easier, and I wish I had them right now: (1) ^R Inject Arguments (suggested key: C-M->): n C-M-> means: n C-M-F ) C-B n C-M-B ( Space C-B [n positive] n C-M-B ( Space n C-M-F ) C-M-P C-B C-B [n negative] That is, take the following n s-expressions, wrap them up in a pair of parens, put a space after the left paren and the cursor before that. This prepares to make a new level of function call with the s-expressions as arguments. (2) ^R Extract Arguments (suggested key: C-M-<): n C-M-< means: n C-M-F C-Space n C-M-B C-W C-( C-M-K C-Y except that the kill ring is not affected (but Undo can undo it). This takes n s-expressions and deletes the s-expression containing them, then puts them back in place of it. Perhaps typing C-M-< after an n C-M-< should arrange to use the same n, i.e. allow incremental layers of list structure. Maybe a zero argument should mean all s-expressions from point to the enclosing ). (3) I think I've suggested these before: LET Replace (LET Query Replace?), and LET Displace (LET Query Displace). LET Replace expects the cursor to be within a LET binding clause, and replaces all occurrences of the variable in the LET body with the corresponding expression. Maybe if the cursor is in front of the set of bindings it does them all. LET Displace does the inverse operation. It reads an expression and a variable, replaces occurrences of the expression with the variable in the body, then creates a new LET binding clause.  Date: 28 MAR 1980 0854-EST From: ECC at MIT-AI (Eugene C. Ciccarelli) Subject: Word Abbrev Mode To: INFO-EMACS at MIT-AI Minor changes, mostly fixes and a few new variables to make things easier for users with init and EVARS files. Currently only on the ITS machines, but soon to move to others. 1. A second C-X U (^R Unexpand Last Word) undoes the first one. 2. M-X Word Abbrev Mode without an argument will toggle the mode. 3. M-X Read Word Abbrev File can read in two kinds of definition file formats: the default fast-loading kind, and the human-readable kind (the one you see with M-X List Word Abbrevs). M-X Write Word Abbrev File will write human-readable format files if the Readable Word Abbrev Files option variable is set to 1. 4. Setting the Only Global Abbrevs option variable to 1 will cause only global abbrevs to be used. Using only global abbrevs is faster, simpler and some people find it more appropriate. 5. You can specify additional characters, e.g. the digits, that should cause expansion. See the description of the M-X Make These Characters Expand command. (Users with AI terminals: you can make Top-characters be expanders too.) Please send bugs to BUG-WORDAB@MIT-AI.  ECC@MIT-MC 03/27/80 02:09:57 Re: WORDAB question To: INFO-EMACS at MIT-AI Currently the M-X Word Abbrev Mode command can also be called via the M-X WORDAB. Would anyone mind if that extra name went away, leaving only M-X Word Abbrev Mode? (For those worried about completion, this should in the default environment, allow M-X Word to complete unambiguously.)  Date: 23 MAR 1980 0156-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 144 up as NE on AI To: INFO-EMACS at MIT-AI *) New buffers will be created in Fundamental mode by default. However, this behavior is controlled by the value of Default Major Mode. It can either be the name of the mode to create new buffers in ("Fundamental", "Text", etc.) or it can be the null string, meaning create new buffers with the mode taken from the previous buffer. *) View Directory displays an entire directory in the system standard format. It takes an arg which specifies which directory to display. *) EMACS keeps better track of which buffers need saving and which ones need auto saving. A buffer will have a star if it needs a real save, even if it has just had an auto save. But even though it has a star, EMACS will realize that another auto save is not necessary. *) Q-register PDL overflow errors are no longer so disatrous. You can even get into Backtrace, although if you do then you will go back to top level when you exit. I am especially interested in having this version tested by people who use Auto Save mode.  Date: 21 MAR 1980 1723-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: PASCAL mode To: INFO-EMACS at MIT-AI The characters _ and ^ are now treated as alphabetic in PASCAL mode. The command C-M-? is no longer redefined. To get that function you must now type M-X Print Last Pascal Indenter. There are two new commands: C-M-+ turns the variable name before point into an assignment to increment that variable. C-M-' inserts a string full of blanks, length controlled by your numeric argument, and lets you edit the contents in using overwrite mode. The files AI:EMACS;PASCAL :EJ, AI:EMACS1;PASCAL > and AI:INFO;EPASC > implement this change.  Date: 21 MAR 1980 0328-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI @MIT-AI 03/20/80 11:30:54 Re: default major mode I definitely want the default mode for new buffers ... Otherwise I would have to put something in my init file to accomplish that. I very often receive messages like this. Why would someone tell me that he might customize his environment, in such a context? I look in replies for these things: indications of popular opinion, arguments pro and con, and information about how a command is usually used. Adding this statement to a message doesn't supply any information in any of those categories. One person who said such a thing explained later that it meant he wanted his opinion counted LESS, because he personally would be satisfied either way. However, it comes across to me emotionally as a threat, intended to influence my decision. If so, it is going wide of the mark, since EMACS is supposed to get customized; it doesn't hurt me that someone will do so. But I don't like the feeling that someone is trying to pressure me. I would like to ask you therefore to refrain from saying such things in the future. Instead, think of the conclusion you would like me to draw from the fact that you would customize your environment, and say that instead.  Date: 21 MAR 1980 0328-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI @MIT-AI 03/20/80 11:30:54 Re: default major mode I definitely want the default mode for new buffers ... Otherwise I would have to put something in my init file to accomplish that. I very often receive messages like this. Why would someone tell me that he might customize his environment, in such a context? I look in replies for these things: indications of popular opinion, arguments pro and con, and information about how a command is usually used. Adding this statement to a message doesn't supply any information in any of those categories. One person who said such a thing explained later that it meant he wanted his opinion counted LESS, because he personally would be satisfied either way. However, it comes across to me emotionally as a threat, intended to influence my decision. If so, it is going wide of the mark, since EMACS is supposed to get customized; it doesn't hurt me that someone will do so. But I don't like the feeling that someone is trying to pressure me. I would like to ask you therefore to refrain from saying such things in the future. Instead, think of the conclusion you would like me to draw from the fact that you would customize your environment, and say that instead.  Date: 21 MAR 1980 0328-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI @MIT-AI 03/20/80 11:30:54 Re: default major mode I definitely want the default mode for new buffers ... Otherwise I would have to put something in my init file to accomplish that. I very often receive messages like this. Why would someone tell me that he might customize his environment, in such a context? I look in replies for these things: indications of popular opinion, arguments pro and con, and information about how a command is usually used. Adding this statement to a message doesn't supply any information in any of those categories. One person who said such a thing explained later that it meant he wanted his opinion counted LESS, because he personally would be satisfied either way. However, it comes across to me emotionally as a threat, intended to influence my decision. If so, it is going wide of the mark, since EMACS is supposed to get customized; it doesn't hurt me that someone will do so. But I don't like the feeling that someone is trying to pressure me. I would like to ask you therefore to refrain from saying such things in the future. Instead, think of the conclusion you would like me to draw from the fact that you would customize your environment, and say that instead.  Date: 20 MAR 1980 0340-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: default major mode To: INFO-EMACS at MIT-AI Do you like having the default mode for new buffers be taken from the previously selected buffer, as it is now, or would you rather have new buffers created by default in Fundamental Mode? This is relevant only when the proper major mode cannot be determined from a -*-FOO-*- or a local modes list or (on Twenex) the filename. Answers to EMACS;MODE ANSWER.  Date: 19 MAR 1980 0250-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Mode line weirdness in EMACS 143? To: INFO-EMACS at MIT-AI If you have an init file which is a compiled library, and you find that in EMACS 143 your mode line is always surrounded by square brackets, you are probably neglecting to set Q..9 to the null string before calling the default init file. You should do :I..9 before you call the default init file from a compiled init file of your own.  Date: 14 MAR 1980 0612-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 143 is available as NE on AI. Please try it. I will be away until Tuesday night, but perhaps MMCM will fix some problems before then. *) M-X What Page will say what page and line number the cursor is on. What Cursor Position is the new name of ^R Where Am I, C-X =, which says where on the screen the cursor is, and the value of point and size of buffer. In completion, "W" is always a good abbreviation for "What". *) The new View File command, which allows Backspace to move backwards in the file, is installed. Note that the reason View File does not use a general recursive editing level is so as to avoid having to read the entire file into core at any time. This means it can start displaying much faster, and it can be used on files which will not fit in core. *) M-Q is once again allowed in all major modes. *) M-K kills to the end of the sentence. It accepts repeat count arguments, positive and negative. C-X Rubout, ^R Backward Kill Sentence, also accepts arguments. All the sentence commands now observe all paragraph boundaries. Blank lines and text justifier command lines are now considered between paragraphs, not part of any paragraph. M-] moves to the end of the paragraph, not the beginning of the next. *) C-K's arguments are now changed: -1 means kill to the previous line-beginning (the beginning of the current line, unless point is there, in which case it is the beginning of the previous line). An argument of 0 still means kill to the beginning of the current line. *) C-X T is ^R Transpose Regions. You mark both ends of one region, and put point and mark around anothe region, and this transposes them. Repeating the command undoes it. *) CR when you are typing a function name (in M-X, etc) now performs completion, internally, to prevent any paradoxical results from the distribution of various function's definitions among multiple libraries. *) M-^ no longer puts in a space when it joins a line onto a preceding blank line. *) "^R" will no longer appear in the mode line when you are inside a recursive editing level. Instead, just square brackets are used. *) Indent Tabs Mode is now a variable you can set as well as a function to toggle the setting. -1 means use tabs, 0 means don't. *) Overwrite Mode now leaves Meta commands unchanged. It appeared that the only people who liked having Overwrite Mode make Meta characters insert were those who also use E at SAIL. So it might make sense for Stanford alone to change this back. For customization only: *) Init files are now executed entirely within ^R mode. This should eliminate some anomalous behavior observed from certain functions when called from init files. *) The command C-X ? can now be redefined just like any other C-X command. *) The prompt used by C-M-X is now obtained from the variable Instant Command Prompt. *) ^R Return to Superior will execute Return from Superior Hook, if it exists, after EMACS is continued. *) 2,M.L FOO makes FOO a permanent local variable in the current buffer. This means that FOO will not lose its local definition if you switch major modes. This is intended to be used by libraries to store per-buffer internal information, not by the user, because making a variable both a permanent local and a major mode local does not attempt to do anything reasonable. *) The name ^R Modified Two Windows has been changed to ^R Visit in Other Window. This is the function on the C-X 4 command. *) M.C can be used to give a variable a default value (set it only if it does not already exist).  Date: 13 MAR 1980 0139-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Most people prefer to have C-K changed. However, an argument of zero will still mean to kill back to the beginning of the line, as it does now, for the sake of those with strong habits, since it would be a useless no-op under the new scheme.  Date: 12 MAR 1980 0642-EST From: RMS at MIT-AI (Richard M. Stallman) To: scott at SRI-KL CC: INFO-EMACS at MIT-AI I prefer to use Altmode when I give an example of TECO code so that anyone who examines the example closely or copies the code into an init file will not get screwed. I must respectfully consider it a bug when a mail-reading program does not display some character such as Altmode in a useful fashion.  Date: 12 Mar 1980 0331-PST From: Scott at SRI-KL (Scott J. Kramer) Subject: 's in messages To: bug-e at MIT-AI cc: info-emacs at MIT-AI Please use dollar sign to represent in Emacs-related messages, the doesn't display. Thanks. -sjk -------  Date: 12 MAR 1980 0622-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The VARG library is having its name changed to VT52. The names of the commands in it are also changed. Do M-X List LibVT52 to see the details.  Date: 11 MAR 1980 0437-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Hello, printing terminal EMACS users To: INFO-EMACS at MIT-AI Would you please read the file AI:EMACS;PRNTNG > and then tell me any useful techniques you have worked out for editing with EMACS on a printing terminal which are not in that file? These techniques will go in the manual.  Date: 10 MAR 1980 0550-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: C-K command arguments To: INFO-EMACS at MIT-AI Right now the C-K command is unlike the other killing and motion commands in the way it understands zero and negative arguments. Zero means kill back to the start of the current line, killing nothing if point is there already. Minus one means kill back to the beginning of the previous line. This is simply inherited from the TECO K command. If C-K were like the other commands, an argument of -1 would kill back to the previous line-beginning, which could be the beginning of the current line, or the beginning of the previous line if you were at the beginning of a line. Do you think this change would be a good idea?  Date: 10 MAR 1980 0534-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Sentence question answers To: INFO-EMACS at MIT-AI Most people want M-A and M-E to stay approximately the same as now, but want M-K to be changed to kill forward to end of sentence.  KMP@MIT-MC 03/10/80 04:12:16 Re: More on TeX mode To: INFO-EMACS at MIT-MC Apologies for all the mail you non-TeX folks are getting on this ... hopefully this library will stabilize shortly. If not, I'll create an Info-Tex-Mode mailing list, but I'd rather not bother ... (1) If you put 1 M.VInhibit TeX Dollarsign in your Emacs init, you will not get TeX's flashing dollarsigns when you do MM Tex Mode ... They still default to on, since I believe people are finding them useful. If you do have fancy $'s enabled, C-M-$ (Control-Meta-Dollarsign) will also get enabled to insert that funny barrier, keeping $ from scanning before that point to look for matching $'s. (2) The bug making \$ give a search-failed message instead of just self-inserting has been fixed. -kmp  KMP@MIT-MC 03/09/80 04:32:12 Re: TeX mode -- helpful hint To: INFO-EMACS at MIT-MC CC: DCB at MIT-MC Something which I guess I should have mentioned in the info about TeX mode is that in order to support flashing $...$'s, it is necessary to scan back as far as the last pending { or beginning of file if at toplevel. This can be very inefficient for long files and there is no way to get around it trivially from software. Hence, you may find it useful to add to your buffer at points where you know there is a division that you would like the package not to scan back past, the following character configuration: % {] This will tell TeX mode not to scan back past this { in looking for other $'s each time you type a $ ... It will not confuse the {...} balancing commands except at the instant you type the ] ... and finally, it won't come out in your output since % will comment it out. If this really causes people problems and they find $'s are too slow or that for someother reason they would like not to have it defaultly enabled, I'll create a variable you can set to make TeX mode not turn on flashing $'s by default. -kmp  Date: 8 MAR 1980 2331-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Carriage-return command To: INFO-EMACS at MIT-AI The answers to the question about the CR command are in EMACS;CR ANSWER. Almost everyone prefers the command the way it is now.  Date: 8 Mar 1980 0004-EST Sender: ADOWEN at BBN-TENEXD Subject: Re: Question on sentence commands From: Buz Owen To: RMS at MIT-AI Cc: INFO-EMACS at AI Message-ID: <[BBN-TENEXD] 8-Mar-80 00:04:16.ADOWEN> In-Reply-To: Your message of 7 MAR 1980 0558-EST I like M-a and M-e except that they presently ignore negative args. I think like they ought to move in the opposite to normal direction to the beginnings or ends of sentences respectively with negative args. This would provide a way to move to the beginning of the next sentence, i.e. to the end of the current one and past the whitespace. A function to skip over contiguous whitespace is another way to make the same thing possible with four keystrokes, (assuming a non-meta keyboard and the skip-whitespace function on a meta or control-meta key.) I think M-k should kill the current sentence with no arguments; kill forward or backward with one positive or negative arg, kill a range from arg1 beginnings of sentences back to arg2 ends of sentences forward with two args. Perhaps there ought to be killing of the space after or before which would be applied uniformly by commands which kill words and sentences and maybe larger units, but unless it were carefully worked out, I prefer things the old familiar way they are now.  Date: 7 MAR 1980 0607-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: CR command question To: INFO-EMACS at MIT-AI A year ago most people wanted CR to be changed not to gobble up the last blank line before a nonblank line. Now that it has actually been this way for a while, do you still think it is a good idea? Or would you rather have CR always use an existing blank line if there is one, even if the following line is not blank? Answers will go in EMACS;CR ANSWER  Date: 7 MAR 1980 0558-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Question on sentence commands To: INFO-EMACS at MIT-AI Do you like where M-A and M-E position the cursor? Or do you think that they should sometimes move past the inter-sentence delimiters? Do you like having M-K kill a whole sentence, or would you rather have it kill forward to the end of the sentence?  HAL@MIT-MC 11/29/79 18:31:20 To: INFO-EMACS at MIT-MC Could someone tell we what command I can add to my emacs init to cause new buffers to be created in text rather than fundamental mode?  KMP@MIT-MC 02/24/80 05:59:26 Re: TeX Mode To: INFO-EMACS at MIT-MC TeX mode now runs a special macro for $...$ matching which attempts to understand TeX's basic rules for when $'s are acceptable. If anyone has any troubles with odd configurations of {...}'s and $...$'s displaying an incorrect match, please let me know. -kmp  Date: 20 FEB 1980 2125-EST From: KMP at MIT-MC (Kent M. Pitman) To: INFO-EMACS at MIT-MC CC: INFO-TEX at MIT-MC I have created a TeX mode in Emacs. Do MM Load LibraryTEX to get in the library and the MM TeX Mode to get the macro turned on. It features $...$ is given the lisp syntax of |...| so that C-M- will treat it as a single object to be manipulated. {...} are given the syntax of (...) as is [...] and qPermit UnMatched Paren gets zeroed out so you get a beep when you mismatch your parens. \ gets the lisp syntax of / so that it doesn't interfere with C-M- commands when you do \$ or \{ or \}. % gets set up as the comment character. Tab gets set up to do indent relative and Rubout/Control-Rubout get turned around so that Rubout will rub backward changing tabs into spaces. There is also a macro named ^R TEX " which you can put on the " character if you want. It will insert `` or '' as appropriate according to context if you're the kind that is prone to forget to use these -- but that macro is not on any key by default. If anyone else has any good ideas for how to make TeX more livable via this editting mode, I'm open to suggestions tho' I won't make any guarantees. -kmp  Date: 13 February 1980 00:34-EST From: Eugene C. Ciccarelli Subject: WORDAB library To: Info-Emacs at MIT-AI Sometime in the near future I will be installing a new version of the word abbrev library with a few changes that need advance notice: 1. To be compatible with the convention now used by other EMACS minor modes, doing MM Word Abbrev Mode$ (i.e. no explicit numeric argument) will toggle the mode instead of turning it on. 2. The command MM Word Abbrev Apropos$filter$ will be going away, its functionality replaced by MM List Word Abbrevs: the commands MM List Word Abbrevs$ and MM Insert Word Abbrevs$ will take a filtering string argument so that some abbrevs will be listed/inserted. A null argument will mean list/insert all the abbrevs, thus giving the old behavior. Thus to prepare for these changes, if you use these commands (say in your init file), you can make the following replacements, which will work now and after the change: MM Word Abbrev Mode$ -> 1 MM Word Abbrev Mode$ MM List Word Abbrevs$ -> MM List Word Abbrevs$$ MM Insert Word Abbrevs$ -> MM Insert Word Abbrevs$$ If you use MM Word Abbrev Apropos$ in your init or some other function, could you tell me?  Date: 6 Feb 1980 1131-PST From: Denny Brown To: RMS at MIT-AI, INFO-EMACS at MIT-AI cc: csd.bth at SU-SCORE In-Reply-To: Your message of 3-Feb-80 0211-PST In order to give a consistent story to the secretaries here, I'd rather have to C-Q to insert any funny character. We use Meta characters for lots of things in our libraries. -Denny Brown -------  Date: 3 FEB 1980 0511-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI If you use Overwrite Mode on a terminal with a Meta key: Do you like the feature that Meta-characters insert and Altmode must be used to get the word commands? Or would you rather that Meta-characters continue to be the word commands and C-Q be necessary for insertion? I realize that the set of people to whom this is addressed may well be empty.  BLOTTO@MIT-MC 02/02/80 15:11:15 To: INFO-EMACS at MIT-AI PLEASE ADD ME TO YOUR MAILING LIST THANX, BLOTTO@AI  KMP@MIT-MC 01/28/80 05:27:11 To: INFO-EMACS at MIT-MC Anyone who uses my MODLIN or TIME libraries and has noticed that MM Query Replace and/or MM Replace have not been functioning with 100% reliability can rest a bit easier. I have tracked down and corrected what I believe to have been the cause of the problem. This fix is installed in ITS versions of these libraries. -kmp  Date: 27 JAN 1980 0432-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI there is a new View File command in the library VFILE on AI. It can move backwards in the file as well as forwards. Type Backspace to move back. Its possible disadvantage is that it may be slower. please try it and tell me what you think of it. If people think it is a win, I'll make it the standard version. It works by using random access IO to window part of the file into a buffer, then windowing the buffer onto the screen.  Date: 27 JAN 1980 0245-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI MM Tabify will be changed so that just two spaces will never be replaced with a tab. It was suggested that this might be done only at ends of sentences, but this would be extra difficulty, and could not eliminate problems if there were any, but only make them happen less frequently. It doesn't appear that failure to insert a tab would be a problem, anyway.  Date: 25 JAN 1980 1811-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Should MM Tabify refrain from using tabs to replace just a pair of spaces? The idea is that ends of sentences don't become garbaged.  JLK@MIT-MC 01/24/80 08:46:44 To: INFO-EMACS at MIT-MC When the DELIM library is used, the function of C-] is on C-M-G.  KMP@MIT-MC 01/23/80 09:07:45 To: INFO-EMACS at MIT-MC Users of the MODLIN package may note that I have modified how the mode line comes out ... lest ye be confused, please note -- 'DDT:' will no longer appear when Emacs is a toplevel job. When it is running inferior to another job (as in MAILT or LISPT) it will still try to say LISP: or MAIL: as it always has. This change was made to conserve room for the more useful data which some people have managed to run off the end of the existing mode line. -kmp  Date: 20 JAN 1980 0250-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI M-Q (^R Fill Paragraph) will now be allowed in programming language major modes. It used to be prohibited as likely to screw you, but with M-X Undo available it should not be much of a screw any more.  Date: 20 JAN 1980 0236-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI JLK@MIT-MC 01/15/80 03:31:49 When you revert a file from disk, it doesn't seem to run your Visit File Hook (it presumably should run it with an arg of 0, as it does when you first visit the file in a buffer). Can anyone see any problems with this idea?  Date: 17 Jan 1980 1034-EST From: Sross at MIT-XX (Sandor Schoichet) Subject: Spelling correction To: info-emacs at MIT-AI I second HAL's idea for integrating the spell program with emacs. -------  HAL@MIT-MC 01/16/80 22:29:21 To: INFO-EMACS at MIT-MC It would be really nice to have an emacs library which allowed one to do spelling correction on the buffer, say, by invoking the spell program or some other way. In fact, it could be considerably more useful than the current spell program if it also gave the user the option of going into a recursive edit mode at the appropriate point in the buffer in response to a spelling error. Any benevolent hackers out there?  MOON@MIT-MC 01/15/80 09:45:10 To: JLK at MIT-MC, INFO-EMACS at MIT-MC JLK@MIT-MC 01/14/80 16:45:41 To: info-emacs at MIT-AI I have a Visit File Hook which hides local modes info at the end of the file by setting the buffer bounds, and a command to toggle the buffer bounds so you can edit them if you want. This has the advantage that M-> does not put you after the modes info, and if that info is hairy, it doesn't make your file look so kludgy while you are editing it. Now, for editing public files I think it might go along way towards lowering people's resistance to having this modes information in their files if this feature were made standard. Let me point out that I think it would be an extremely bad idea to have anything in the standard Emacs that hid anything in the file from the user. This would only lead to difficulties and confusion when there was something wrong with whatever it chose to hide, or when it decided to hide the wrong thing. This happened in the past when Teco (or Tecmac?) had a misfeature by which it did not show extra control-C's at the end of a file. A student somehow got several hundred control-C's at the end of his thesis, which came out as garbage characters when text-justified. These control-C's could not be removed or even seen with the editor (nor with the standard file-printing command at the time, which had an even worse bug with control-C's). I think a better approach would be to have a command that moved you to just before the local modes, with the local modes just off the screen. But please don't change m-> to do this!!  JLK@MIT-MC 01/14/80 16:45:41 To: info-emacs at MIT-AI I have a Visit File Hook which hides local modes info at the end of the file by setting the buffer bounds, and a command to toggle the buffer bounds so you can edit them if you want. This has the advantage that M-> does not put you after the modes info, and if that info is hairy, it doesn't make your file look so kludgy while you are editing it. Now, for editing public files I think it might go along way towards lowering people's resistance to having this modes information in their files if this feature were made standard.  Date: 13 JAN 1980 2042-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: View Buffer To: INFO-EMACS at MIT-AI The way to exit from View Buffer while moving point into the screenful that is presently on the screen is now Return instead of C-C.  Date: 9 JAN 1980 0324-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The Set Key function will now ask for confirmation. Also, it really works to use prefix characters to specify the key to be redefined.  JLK@MIT-MC 01/07/80 11:05:46 To: info-emacs at MIT-AI I have been receiving dozens of inquiries about why "TXJ no longer works in the new EMACS". In the old EMACS TXJ was autoloaded so that simply typing M-X TXJ worked. Now you have to type M-X LOAD LIB$TXJ and then do M-X TXJ. I don't know why RMS flushed this without announcing it, but I am not surprised -- you will have to ask him. Perhaps some general scheme would be desirable whereby a failing M-X command would check for a library of the given name and try autoloading it. This would save a few characters for autoloading macro definitions in the default EMACS.  Date: 4 January 1980 01:03-EST From: Mark L. Miller Subject: Incompatible change philosophy To: RMS at MIT-AI cc: INFO-EMACS at MIT-AI Just for the record -- we have NOT been screwed by the chan}~es, and are grateful to keep receiving the new releases since we find most of the incompatible changes to be improvements worth adjusting to. we regard emacs as an important experiment in designing high productivity software develpment systems for expert programmers. We are also interested in systems for naive users, which EMACS is not -- in its current form -- but EMACS is more of a step, even toward that goal, than any other editor currently available. It got that way because of a refusal to maintain compatibility with the sins of the past. (I too believe in "upward compatibility" when possibld #(V3?M-E^3*l3*n3*phare in maintaining it.  Date: 23 DEC 1979 0258-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: paragraph commands To: INFO-EMACS at MIT-AI What appears to be wanted is: text-justifier command lines are treated as blank lines, and M-] stops at the end of the paragraph instead of the beginning of the next one. M-[ with a negative arg will go down to a paragraph beginning and M-] with a negative arg will go backward to a paragraph end. This will be implemented at some time in the future.  Date: 23 DEC 1979 0256-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Read only file visiting To: INFO-EMACS at MIT-AI EMACS;FILE PLAN contains a plan for changing the EMACS file visiting commands, designed to satisfy people's needs as expressed in their answers to my earlier questions. Please tell me your opinions of it.  Date: 15 Dec 1979 1109-PST From: Ted Markowitz Subject: Re: Standard exit command To: RMS at MIT-AI, INFO-EMACS at MIT-AI In-Reply-To: Your message of 13-Dec-79 0016-PST I think that C-\ would do the trick. --ted markowitz -------  Date: 13 DEC 1979 0316-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Standard exit command To: INFO-EMACS at MIT-AI What would you think of having C-\ as the standard exit command for all recursive editing levels and subsystems, just as C-] is the standard abort command? C-\ would replace things like C-M-C (or C-M-Z), Q, and C-Altmode. Unfortunately, C-M-C (or C-M-Z) can't be used because it requires a meta or control-meta prefix to type, and in subsystems there may be no prefix characters. C-Altmode has the same problem. Thus limited to ASCII characters, and ruling out printing ones and popular commands, C-\ is the only choice. It would still be imaginable to make C-M-C (or C-M-Z) the standard exit command by making the control-meta prefix work in all subsystems, but this would be painful to design in detail.  Date: 7 December 1979 05:06-EST From: Eugene C. Ciccarelli Subject: WORDAB problems To: Info-Emacs-News at MIT-AI The installation problem that Scott reported should be solved. If anyone else encounters any difficulties at all with WORDAB, please mail to ECC@MIT-MC. One thing mentioned before has been changed: the commands MM Define Word Abbrevs, MM Insert Word Abbrevs, and MM Kill All Word Abbrevs have not gone away since they are used more than I thought they'd be. Another note, geographical: the installations of WORDAB have occurred on the ITS machines, BBND, and MIT-XX.  KRAMER@MIT-MC 12/07/79 02:50:05 Re: Oversight To: INFO-EMACS at MIT-MC My mistake, I can just load OWORDA for now although I'd rather have to modify to get the new library. When new versions of Emacs come out, they're labeled experimental so it might be better if libraries were made this way until tested. Scott  KRAMER@MIT-MC 12/07/79 02:18:04 Re: Changes that screw up init files To: INFO-EMACS at MIT-MC Earlier tonight I received a message via INFO-EMACS-NEWS saying that changes had been made to WORDAB. Fine, but I think it is unreasonable to inflict (harsh word) these changes on people without some transition period so that they may make changes to init files or whatever. This is especially true for a person like myself who is not very familiar with Teco and may need help in making these modifications. Please consider this in the future. Thanks, Scott  Date: 6 December 1979 22:47-EST From: Eugene C. Ciccarelli Subject: New WORDAB release To: INFO-EMACS-NEWS at MIT-AI A new version of the WORDAB (Word Abbrev Mode) library has been installed. This new version is a minor change from XWORDA library which has been in use for quite a while among several users. The old version of the WORDAB library will be available for a while as OWORDA. The major changes that uses of the old WORDAB library will see are: 1. The format used for files of saved abbrev definitions is different, less human readable but much faster to load. For users with old-style format files, the MM Read Word Abbrev File will perform a conversion and rewrite the file in the new format. (The old file will be copied, just in case, to WORDAB ODEFNS on ITS, WORDAB.ODEFNS.0 on Tenex/Twenex sites.) 2. Multiple-line expansions are supported. 3. MM Edit Word Abbrevs has a slightly different format, to allow the more general expansions, e.g. multiple-line. Basically, if you have a double-quote character (") inside an expansion, it will be doubled (""), to distinguish it from the double-quotes that surround the entire expansion. 4. There is a new command, MM Word Abbrev Apropos, that will list some (not all) of the abbrevs currently defined. Like MM Apropos, it takes a string argument to match against the defined abbrevs: any abbrev containing that string, or having an expansion containing that string, will be listed. 5. Some commands went away, to leave a simpler looking user interface. These commands were: MM Define Word Abbrevs, MM Insert Word Abbrevs, MM Kill All Word Abbrevs, and MM Incremental Word Abbrev Apropos. If this bothers anyone, contact me. (There are some subroutine available to do the job, for instance.)  Date: 5 DEC 1979 2033-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI A clarification: the new mailing list is for people who are involved in maintaining EMACS up to date outside MIT, to inform them of when there is something new that they should copy to their own sites.  Date: 5 DEC 1979 1740-EST From: rms at MIT-AI (Richard M. Stallman) Sent-by: GJS at MIT-AI To: INFO-EMACS at MIT-AI There is a new mailing list called EMACS-HACKERS for people who are involved in maintaining EMACS or parts of it at various sites. If you would like to be on it, let me know.  Date: 1 DEC 1979 1153-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Twenex problems that screw EMACS To: INFO-EMACS at MIT-AI A list of Twenex problems that screw EMACS can be found in the file .TECO.;TWENEX LOSSES on AI or TWENEX.LOSSES on XX.  Date: 29 November 1979 00:05-EST From: Robert W. Kerns Subject: Change to PAGE and SLOWLY To: INFO-EMACS at MIT-MC To allow them to be used with each other, PAGE and SLOWLY have been changed. The functions ^R PAGE Reverse Search and ^R SLOW Display Reverse Search have been flushed .... ^R Reverse Search (the default binding for C-R) indirects through ^R Incremental Search. Thus now a search can be both SLOWLY and PAGE-ified....  Date: 25 November 1979 13:09-EST (Sunday) From: KRALEY at BBN-TENEXA To: info-emacs at mit-ai cc: Kraley at BBN-TENEXA Subject: source compare So many people have answered my query about source compare with what I feel to be irrelevant answers that I fear that I have not made myself clear enough. I am very familiar with the various "batch" source compares that exist (although I did learn about some new ones too) and it is precisely because of that that I sent my message. One of the problems with a source compare like that is you typically need to spread out your listings of the two versions, along with your listing of the differences, and then work hard to understand what really changed. Just like a display editor (like EMACS) is infinitely superior to a printing terminal editor (tenex teco) because you can understand what's happening, so would a "display" source compare that I envision. The key thing is that the files being compared would be VISIBLE and the changes can be directly observed and compared. The full power of the editor could be used to walk around in the files to see what is happening and possilby make some fixes. I point out the analogy to the compile feature mentioned in the multics emacs paper that lets you, in the context of the editor, look at the source file and error file in two windows and invoke a command to move the cursor inthe source file buffer to the "next" error mentioned in the error file. this allows one to be freed from the necessity to spread out those listings in the same way. Come on, ladies and gents, let's raise our horizons somewhat! ...Mike  Date: 25 Nov 1979 0645-PST From: NCP.Alpha at SU-SCORE (EGK) Subject: Re: source compare To: KRALEY at BBN-TENEXA cc: info-Emacs at MIT-AI In-Reply-To: Your message of 24-Nov-79 1609-PST There are two programs, namely SRCCOM and FILCOM which do an aquidite job of comparing files and giving you the differences. I dont see why anyone would waste his time writing an emacs macro to do this. Ed -------  Date: 24 November 1979 19:09-EST (Saturday) From: KRALEY at BBN-TENEXA To: info-emacs at mit-ai cc: Kraley at BBN-TENEXA Subject: source compare Has anyone ever written a source compare function for EMACS? It seems like such an obvious thing, ie. have the two files to be compared in two buffers and a user command to advance to the next difference which would then be displayed in two windows, perhaps with the differing lines flagged in one of the margins. Then you'd be able to invoke a ^R to walk around in the files. Anyone ever heard (or lust after) such a thing? ...Mike  Date: 23 NOV 1979 0614-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: INFO-EMACS-NEWS To: INFO-EMACS at MIT-AI To announce changes in EMACS or EMACS libraries, send mail to INFO-EMACS-NEWS@MIT-AI. This will go to INFO-EMACS and also to a file which logs change announcements, EMACS;EMACS NEWS. It will magically cause INFO-EMACS to appear in the header of the message, thus preventing any replies, complaints, etc. from getting into the EMACS NEWS file.  Date: 23 NOV 1979 0550-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: ^R Exchange Characters, etc. To: INFO-EMACS at MIT-AI In EMACS 141, all these functions will be renamed to things like ^R Transpose Characters. This should go better with commands like C-T, M-T, etc.  Date: 23 NOV 1979 0530-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 141 change To: INFO-EMACS at MIT-AI Revert File with an argument will revert to the last version of the visited file, ignoring any auto save files made since then if you are autosaving under separate filenames. A precomma argument will make it not require confirmation.  Date: 23 NOV 1979 0528-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 141 changes To: INFO-EMACS at MIT-AI In EMACS 141, there will be these changes: * Whether the parenthesis matching display feature beeps if there is no matching parenthesis is controlled by the new variable QPermit Unmatched Paren. If it is 1, unmatched parens are permitted (no beep) in all major modes. If it is 0, they are never permitted (always beep). If it is -1, they are permitted except in Lisp mode (because Lisp mode sets the variable to 0 if it was -1!). The initial setting is -1. Also, mismatched parens, as in "[)", will always beep if the matching display feature is turned on. * Help N prints out the file EMACS NEWS of EMACS change information. * C-X = will now print out the total size of the file in characters as well as its other information. It should be self-explanatory. * In M-X and C-M-X, just "I " as the start of a command will expand automatically into "Insert ". This is in addition to "V" for "View, "L" for "List", "K" for Kill and "E" for Edit. My intention is to rename the commands that do one of these sorts of things to start with the "standard" word for that type of operation, to make the command names easier to remember. * Get Date has been renamed to Insert Date. Alter Options has been renamed to Edit Options. Edit ..D has been renamed to Edit Syntax Table. The first two of these changes are examples of the policy described in the previous section. * The function Auto Save Mode will no longer set the number of versions to be kept or the number of characters between saves. Its argument will only say whether to turn the mode on or off. Those other parameters are the values of the variables QAuto Save Max and QAuto Save Interval. The first one already existed; the second one is new. * Write File and Set Visited Filename now run the Visit File Hook, if there is one, but give it an argument of 1 to distinguish themselves from an actual Visit File. Also, these changes have been made in other libraries: In KBDMAC, there is now C-U C-X Q. It causes a recursive editing level to be entered when the macro is defined and again when the macro is run. Thus, at that point in the macro, you can do a different thing each time around. In PAGE, there are several changes: * ^R Go to Page is now ^R Goto Page, so as to resemble the other commands, and lives by default on C-X C-P instead of on M-G. M-G and C-M-G are no longer changed by the package. * C-X C-P when you are looking at the whole buffer will now narrow to just the page that point is in. * Print File Directory has been renamed to View Page Directory. (After all, a file directory is something entirely else). Make File Directory has been renamed to Insert Page Directory. * Trying to move to a nonexistent page will get an error, but will leave you in the first or last page with the correct page number in the mode line.  Date: 23 NOV 1979 0416-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Reap File and Clean Directory To: INFO-EMACS at MIT-AI Giving a nonzero precomma argument to these will make them ignore the don't-reap bit.  Date: 23 NOV 1979 0239-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Reap File and Clean Directory To: INFO-EMACS at MIT-AI These ITS-only functions will now never delete any file whose don't-reap bit is set.  Date: 15 Nov 1979 1744-PST From: NYBERG at USC-ISIE (Karl A. Nyberg) Subject: EMACS at ISIE To: info-emacs at MIT-AI Can anyone tell me who has put EMACS on ISIE, and how I could get them to get a newer version out here? I would even be glad to help if there is any way that I could. Thanks. Karl -------  Date: 15 Nov 1979 1920-EST Sender: ADOWEN at BBN-TENEXD Subject: Re: the change to M-K From: Buz Owen To: DDM at MIT-AI Cc: INFO-EMACS at MIT-AI Message-ID: <[BBN-TENEXD]15-Nov-79 19:20:20.ADOWEN> In-Reply-To: Your message of 15 NOV 1979 1711-EST The logical symmetry of M- and C- A, E and K seems to me much more important than left-right hand symmetry which is not part of the rest of the key assignments. The fact that ^R delete sentence doesn't work analogously to C-K is a second order issue which ought to be fixed someday. buz  Date: 15 NOV 1979 1731-EST From: PONDY at MIT-AI (Peter Andreae) Subject: the change to M-K To: DDM at MIT-AI CC: INFO-EMACS at MIT-AI personally, I reckon that you should change that in your init file - having two keys doing the same thing just for left and righthandedness seems a bit wasteful of keys. PONDY.  Date: 15 NOV 1979 1711-EST From: DDM at MIT-AI (David D. McDonald) Subject: the change to M-K To: INFO-EMACS at MIT-AI My appologies for not using NE and finding this problem earlier, but I've just had my first experience with the new assignment of M-K to ^R delete sentence and I don't like it. Deleting single words is something I do extremely often and having a symetric set of keys assigned to it was very convenient (i.e. M-D and M-K allowed you to delete words while holding down your choice of meta keys ...very useful to us lefties). Is there anyone out there who agrees and would like the default changed back to ^R delete word ??  Date: 14 NOV 1979 1620-EST From: rms at MIT-AI (Richard M. Stallman) Sent-by: ___021 at MIT-AI To: INFO-EMACS at MIT-AI By "read only" I mean that it would be impossible to modify the buffer. I suppose that an alternative might be that M-X Save All Files would not offer to save the buffer and C-X C-S would ask for confirmation before saving it. Sorry for not having explained this before. In fact, I'm sorry I asked, because all I found out is that nobody understands how things work now. Most of the reasons people gave for doing things the way they do are based on factual errors about how things work. So the main need seems to be to stop having C-X C-R and C-X C-V as nearly identical commands so that people will not be: confused about the difference.  Date: 14 NOV 1979 0154-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI It seems some people don't know what C-X C-V is and how it differs from C-X C-R, and therefore don't understand what the actual consequences of making C-X C-R mean "read only" would be. Right now C-X C-V and C-X C-R are identical except that when you visit a file with C-X C-V, change it, and then visit another file in that buffer, EMACS will offer to save the original file. If you had visited the original file with C-X C-R, EMACS would not offer to save the changes. Aside from an effect on auto saving, this is the ONLY difference. When a file is visited in a separate buffer with C-X C-F, it does a C-X C-V type of visit rather than a C-X C-R type of visit, unless you change a flag (Find File Inhibit Write).  Date: 14 NOV 1979 0102-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X C-R To: INFO-EMACS at MIT-AI People have suggested that C-X C-R make the file read-only. What do you think of this, and why? If you are one of the people who thinks that C-X C-V is "dangerous", do you ever use C-X C-F? If so, do you set QFind File Inhibit Write in your init file? If not, you are effectively using C-X C-V whenever you use C-X C-F. Do you still think it is dangerous?  EJS@MIT-MC 11/11/79 01:04:38 To: INFO-E at MIT-MC hello  EULER@MIT-MC 11/10/79 10:32:56 To: INFO-E at MIT-MC If anyone iss interested, I have written a screen editor for BASIC programs. The editor sits in memory from 0 to 137 hexadecimal, and it can be initialized by PRINT USR(0) from BASIC. The editor has the standard cursor moving commands plus ^d (delete) and ^i (insert). If you are interessted, send mail to EULER@MC.  Date: 10 NOV 1979 0302-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 140 To: INFO-EMACS at MIT-AI EMACS 140 now exists. C-X C-S does a real save now even in auto save mode. C-U C-X C-S is the way to ask for an "auto" save. C-M-X (^R Instant Extended Command) won't actually start the command until you type a CR. Before, it was unpredictable whether a Space or Altmode would actually start the command. A library can now contain a macro & Kill Library to be run when that library is killed (by Kill Libraries). Twenex users: if the system doesn't know a terminal type for your terminal, you can use the Set Terminal Type command to specify it to EMACS.  RMS@AI Here is what has changed in version 139 since version 135: * Tab Stop Definitions works differently now. The value of this variable is now just a string, not a buffer, so you can set it more easily in an init file or with a local modes list. Also, the two lines it is supposed to contain have been interchanged. This means that now it is the second line which can usually just be blank. When that happens, you are allowed to omit the CRLFs, which is an additional convenience in setting the variable. You can still set it with Edit Tab Stops. * C-X C-T now runs ^R Exchange Lines. * The variable Lisp ) Hack has been renamed to Display Matching Paren. It is now -1 by default. Text and TECO modes turn it off locally, but you can have a Text Mode Hook or TECO Mode Hook to turn it back on if you like. Whenever the feature is on, it will now apply automatically to all characters which have the syntax of ")" in ..D. This uses the new FS ^R PAREN feature of TECO. * There is now a new standard way to get out of subsystems and recursive editing modes: the character C-]. All recursive editing modes (except minibuffers) will automatically support C-], and all subsystems ought to support C-]. For example, C-] while editing a message to send will exit and not send the message. If there is some way to resume the command you aborted, C-] will tell you. For example, after aborting a message, it will tell you how to resume editing it so that you do not lose what you had typed in. * The new Undo command will undo many sorts of gross changes to the buffer. You do not have to give it immediately after the change occurs. However, only the last gross change can be undone in this way. Not all changes can be undone. However, Undo always tells you what sort of change it plans to undo, and asks for confirmation. Also, Undo can be undone by a second Undo, so it is safe to try it out. * The C-X C-E command has been deleted. To get auto-save to save as the visited filename, set QAuto Save Visited File nonzero and use C-X C-V. A zero setting of QInhibit Write now means that C-X C-V was used. * Auto save can now be used when you are not visiting any file, but it will never be on by default at such times. You must use M-X Auto Save to turn it on. If you do C-X C-S when you are not visiting a file, it will ask you for filenames to save under. * Using Rubout or C-D with an argument now saves on the kill ring. * A CRLF will now appear to be a single character when you move over it with C-F or C-B. * M-K is now ^R Kill Sentence. It kills the sentence containing point, or the one following point if point is between sentences. * C-M-X is now an alternative version of M-X which does not read any arguments for the command itself. As soon as you terminate the command name the command is executed, and it reads its own arguments. This has the advantage that any special processing on these arguments (if they are command names or filenames) can take place. It has two disadvantages: you can't rub out from the arguments back into the command name, and the command is not saved in the ring buffer of commands for access with C-X Altmode. * EVARS files are now better documented, and allow you to specify C-X subcommands and to include comment lines. * The mode line will no longer contain the name of the selected buffer when that is the same as the first filename of the file you are visiting. * When you ask for the documentation of a command, and it tells you about another command, it will tell you what that command is in your current environment, rather than what it is in the standard environment. For example, when M-Q tells you about the ^R Set Fill Column command, it will normally say "C-X F", but if you move that command elsewhere it will say wherever it actually is. * The SORT library contains commands for sorting the region line by line, paragraph by paragraph, or page by page. The BCPL library implements BCPL mode. The BLSLIB library implements BLISS mode. Unfortunately, there is no documentation for either of those modes outside of their self-documentation. * The Apropos command will omit mention of subroutines if you give it a numeric argument. Help A invokes it this way. This is to avoid confusing new users with them. * ^R Just One Space is now available but not on any key by default. It is like the M-\ command but leaves one space (or how ever many the argument is). * The commands Make Local Q-register and Kill Local Q-register now allow you to make TECO FS flags local. To make the flag FS CTLMTA local, just specify "FS CTLMTA" as the "q-register name" argument. * You can now set any variable conveniently with the Set Variable command, and put commands on characters conveniently with the Set Key command. * C-M-V with just a minus sign as argument scrolls the other window a full screen backwards. * M-X Revert File now asks for confirmation. * Commands M-X Delete File, M-X Copy File and M-X Rename File now exist for performing those operations. They take one or two string arguments, as in M-X Copy File. These should not be needed in TECO code since the ED, E_ and EN commands exist. Changed Names and Eliminated Commands * The commands ^R Set Column, ^R End Comment, ^R MM Via Mini and ^R Mini Visited File have been flushed. * The commands Make Variable, Set Variable Comment, Make Dispatch String and Make Prefix Character now have "&" in front of their names. All commands names which used to start with "^^" now start with "&" instead. & Macro Get Full Name is now called & Macro Name. * The command List One File has been flushed. List Library can now be used to list the contents of libraries already loaded as well as libraries not loaded. The command Variable Apropos has been flushed. The command List Variables now takes a string argument which is a substring to match on. If the argument is null, all variables are listed. Thus, List Variables can be used for either purpose. * The commands RMAIL and ^R RMAIL have been flushed in favor of the Read Mail command, which uses a variable to decide which library to use for reading the mail. * ^R Execute Completed MM Command has been renamed ^R Extended Command, but the old name has been kept around for the sake of libraries which refer to it.  MACRAK@MIT-MC 11/07/79 15:06:46 To: INFO-EMACS at MIT-MC Does anyone have a sort program which will sort 10000 records (total 600000 chars) in Teco by some trick like file merging? Teco itself goes URK! (not surprisingly...)  Date: 6 NOV 1979 0212-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI C-X C-S will be changed to do a real save always. C-U C-X C-S will explicitly request an auto save.  Date: 5 NOV 1979 0311-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X C-S with auto saving on. To: INFO-EMACS at MIT-AI Does anyone find useful the feature of C-X C-S in auto save mode whereby it does just an auto save type save unless given an argument? Or would you rather that it did a real save with no argument?  Date: 2 NOV 1979 2341-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Paragraph commands To: INFO-EMACS at MIT-AI Right now, a line which is a text justifier command is treated as a paragraph unto itself. It has been suggested that text justifier command lines be treated as blank lines which only separate paragraphs. Another alternative would be to treat any consecutive set of text justifier command lines as one paragraph. What is your opinion? Answers will go in EMACS;PARA ANSWER.  KMP@MIT-MC 11/02/79 17:55:47 To: INFO-EMACS at MIT-MC For those of you with VT100 terminals if you run in ANSII mode, the VT100 library, which you can get via MM Load LibraryVT100 will set up the keypad to do pretty reasonable things as follows: (1) Keypad digits and minus are args to the next command automatically. [Normal digits in main key area are self-inserting still]. (2) Comma will eventually be an arg separator for advanced commands that take precomma args. I haven't coded this yet. (3) Dot runs ^R Documentation like ^_H or Top-H would normally do. (4) Arrows go vertically up, down, back and forth, independent of what you define ^B,^F,^N,or ^P to do ... (5) PF1-4 run macros you can define. They are initially undefined. (6) Enter followed by a PF-key will read characters like the ^X(...^X) does up until the next ^X), C-M-C, or Enter that you type. You can abort a definition by doing ^]. Sorry for the length of the message, but hopefully this will be of interest to enough of you to make it worthwhile. Interestingly enuf, this will not work on AI. No, I am not boycotting -- I just don't know what the new Emacs over there has done and I am nearly sure that my stuff is wholely incompatible at this point ... when the new emacs works its way back over to MC (which is happening slowly now), I will work on an all-ITS compatible version of the handler. -kmp  Date: 2 NOV 1979 1752-EST From: rms at MIT-AI (Richard M. Stallman) Sent-by: SHIM0 at MIT-AI Subject: Tab Stop Definitions To: INFO-EMACS at MIT-AI The value of this variable is now just a string, not a buffer, so you can set it more easily in an init file or with a local modes list. Also, the two lines it is supposed to contain have been interchanged. This means that now it is the second line which can usually just be blank. When that happens, you are allowed to omit the CRLFs, which is an additional convenience in setting the variable. You can still set it with Edit Tab Stops.  Date: 2 NOV 1979 0227-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Lisp ) Hack on by default in EMACS 139. To: INFO-EMACS at MIT-AI The variable Lisp ) Hack has been renamed to Display Matching Paren. It is now -1 by default. Text and TECO modes turn it off locally, but you can have a Text Mode Hook or TECO Mode Hook to turn it back on if you like. Whenever the feature is on, it will now apply automatically to all characters which have the syntax of ")" in ..D. This uses the new FS ^R PAREN feature of TECO. Also, C-X C-T is now ^R Exchange Lines.  Date: 30 OCT 1979 0428-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Practically everone wants the ) hack to be on by default. Look in EMACS;) ANSWER.  Date: 29 October 1979 18:56-EST (Monday) From: Winston Edmond To: RMS at MIT-AI, INFO-EMACS at MIT-AI Subject: Lisp ) Hack I think the default value for $Lisp ) Hack$ should be -1 instead of 0. Many new users don't find out about this winning feature for a long time, and then have to make init files just for this reason. What do you think? I think it's a neat hack. I never use it. If you make it the default for LISP mode, I think it would be winning. If you make it the default for all modes, I will have to change my init file. -Winston  Date: 28 OCT 1979 0112-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Lisp ) Hack To: INFO-EMACS at MIT-AI I think the default value for $Lisp ) Hack$ should be -1 instead of 0. Many new users don't find out about this winning feature for a long time, and then have to make init files just for this reason. What do you think?  Date: 22 OCT 1979 0148-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X O, C-X 2 To: INFO-EMACS at MIT-AI People are very divided on this question, and there is no clear mandate, but it seems many people do use the current meaning of C-X O when there is only one window (which surprised me) and do not want it to be changed, so I will leave it alone. The answwers are in AI:EMACS;^XO ANSWER.  Date: 21 OCT 1979 0425-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X C-T To: INFO-EMACS at MIT-AI I am planning to flush this command, so the code is in AI:EMACS1;MINI VISTED in case anyone wants to copy it.  Date: 19 OCT 1979 0303-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Two window mode question To: INFO-EMACS at MIT-AI What would you think of having one command, either C-X 2 or C-X O, to mean enter two window mode, or switch windows if already in two window mode, and flushing the other of the two commands? If you like the idea, which of the two should it be?  Date: 18 OCT 1979 0237-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X C-T (^R Mini Visited File) To: INFO-EMACS at MIT-AI Does anyone use this command?  Date: 18 OCT 1979 0104-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: C-X C-E and C-], C-G. To: INFO-EMACS at MIT-AI Nobody has objected to eliminating C-X C-E, so I will. Most people prefer to have separate C-G and C-] commands. You can read people's comments in AI:EMACS;^G ANSWER and ^X^E ANSWER.  Date: 15 OCT 1979 1741-EDT From: rms at MIT-AI (Richard M. Stallman) Sent-by: BRIAN at MIT-AI Subject: C-] vs C-Altmode To: INFO-EMACS at MIT-AI Several people have asked mewhy we needed C-] when we have C-Altmode (or, C-C C-C). What I propose for C-] to do is not the same as C-Altmode. C-Altmode returns from a recursive edit. C-] will quit the command containing the recursive edit. Thus, C-Altmode while editing a message would send it, but C-] would not send it. The same is true of what C-G would do instead of C-] if the second alternative is preferred to the first. The issue is, is it better to enable novices to get out of all unusual states by repeated use of C-G, or is it better to make it more reliable to choose between quitting a partly typed or running command and quitting a recursive editing mode.  Date: 14 OCT 1979 1112-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I'm considering two possible plans for making quitting out of recursive editing levels more standardized. One is that ^G will always only quit as far as the innermost recursive edit, and ^] will quit out of one recursive edit up to the next edit level. ^G will still flush partially typed and running commands, whereas ^] will be an ordinary command that can't be used in the middle of a command. It will ask for confirmation before quitting. The other plan is that ^G serves both functions. When there is a partially typed or running command, or an argument, it will just flush that. Otherwise, inside a recursive edit, it will ask whether to quit one level up and then maybe do so. If you try to quit while it is asking that question, it could interpret that as "yes" (I'm not sure whether this is a good idea). Please tell me as much as you can about what you think of these plans and why.  Date: 14 OCT 1979 1057-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I am thinking of getting rid of the C-X C-E command. Right now, there are these differences between C-X C-E and C-X C-V: If you visit a file with C-X C-E, and then visit another file on top of it, the old file is saved automatically. If it was visited with C-X C-V, you are asked whether to save it. If you visit a file with C-X C-E, and use auto save, the auto saves go under the same name you visited. If you use C-X C-V, then auto saves go under the "auto save" filenames. I propose to eliminate the C-X C-E command, leaving C-X C-v the same, and add a flag that could be use to get the old C-X C-E type of autosaving while using C-X C-V. Opinions?  Date: 11 OCT 1979 1841-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: FS TTY MAC$ and dumped EMACS environments To: INFO-EMACS at MIT-AI If you use FS TTY MAC$, make sure you put an impure string in it. Strings in libraries don't work, because FS TTY MAC$ is run at startup time before the libraries are loaded.  Date: 3 OCT 1979 2349-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Ok, in the next version Auto Fill Mode, Atom Word Mode and Overwrite Mode commands will toggle the modes. Should the same be done to Auto Save Mode?  Date: 1 OCT 1979 2024-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Do you think that commands like Auto Fill Mode, Overwrite Mode, etc. should toggle the modes if given no argument instead of turning them on with no argument? Explicit arguments of 1 would still turn them on and 0 or - would turn them off. This might be especially useful for putting them on keys.  Date: 18 SEP 1979 1856-EDT From: MOON at MIT-AI (David A. Moon) Subject: New DIRED installed (applies to ITS only) To: INFO-EMACS at MIT-AI The changes are: A new command, M, which asks for a device (and/or directory) and moves the current file onto that device or copies it onto that directory. With a numeric argument does that many files. The C command now looks at the variable SRCCOM switches which if defined is the options to be given to srccom. Various minor bug fixes.  Date: 6 SEP 1979 1550-EDT From: BRIAN at MIT-AI (Brian C. Smith) Subject: Phrase macros To: INFO-EMACS at MIT-AI I have put a simple library of PHRASE macros (meant to be analogous to the word and sentence macros) onto EMACS;PHRASE :EJ on all ITS machines. Loading PHRASE gives you ^R Forward Phrase -- on Control-comma ^R Backward Phrase -- on Meta-comma ^R Mark Phrase -- on Control-Meta-comma ^R Delete Phrase -- on Control-X comma A phrase is taken to be text plus the following punctuation. They all do sensible things with arguments. The search string defining a phrase boundary is kept in the user modifiable variable qPhrase Delimiters. See the auto-documentation. Bugs to BRIAN@AI.  HAL@MIT-MC 09/01/79 21:38:23 To: INFO-EMACS at MIT-MC some suggested minor changes: (1) when a file is opened with ^X^R and auto saving is on, typing ^X^C should NOT write out the file automatically. Rather, EMACS should ask "write out changes in file ..." (2) when a file is opened with ^X^E the mode line should contain an (ED), just as it now contains a (RO) for ^X^R.  Date: 1 SEP 1979 0218-EDT From: CMR at MIT-MC (Clayton M. Ryder) To: INFO-CRTSTY at MIT-MC, (BUG CRTSTY) at MIT-MC To: (BUG EMACS) at MIT-MC, INFO-EMACS at MIT-MC This information may not be of interest to all of those who this was mailed to, but I feel someone might have some answers. I have tried using emacs thru crtsty and :tctyp with the t1061 and dm2500, over the network. problems i have encountered with these terminals are as follows: datamedia 2500 when thru :tctyp (assuming a direct network dial-up thru a tip) echo random alpha characters when inserting text with emacs. it apears some escape sequence of sorts is being sent, to possistion the cursor, however it loses, as if it is being repeated (uneccasryily). this is the only problem i have encountered (i was using both :tctyp dm and :tctyp datamedia setting ) This makes me belive that the lossage is from emacs rather than :tctyp however, I don't know enough about either to know if this is correct. Te;Teleray 1061: When supduping or telenetting from a foreign site, emacs doesn't display properly the control characters, they are not represented byt ^a ^h etc but seem to be echoed as the ascii value and this is quite confusing (as they are therefore not visable on the screen) I don't know the cause for this crock if it is telnet or supdup or what. I hope this information has been o value and if any one knows the answers or can possibly explain this.. I would appreciate an answer thanks, Clay Ryder  Date: 18 AUG 1979 2038-EDT From: AGRE at MIT-AI (Philip E. Agre) To: RMS at MIT-AI, MARKS at MIT-AI CC: INFO-EMACS at MIT-AI, AGRE at MIT-AI Isn't there AI:EMACS;EMACS GUIDE, a manual one can print out on one's own terminal?  Date: 18 AUG 1979 2032-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI There isn't any printed EMACS reference manual. The only printed documentation is the primer. You can order a copy of that from PUBLICATIONS@AI. I may someday get around to generating a printed manual from EMACS GUIDE, but that will not be trivial to do.  JMSK@MIT-MC 08/18/79 16:31:08 Re: MARKS msg on manual To: INFO-EMACS at MIT-MC I would be interested in hearing reponses to the above query . Thanks  MARKS@MIT-MC 08/18/79 03:26:12 To: INFO-EMACS at MIT-MC is there any way for me to get an emacs manual? i'm on a 300 baud line and it's real rough to use emacs... i was told there are some tricks like shortening page lengths and such that help response time... thanks...marks s.  Date: 13 AUG 1979 1140-EDT From: MCLURE at MIT-AI (Stuart M. Cracraft) To: INFO-EMACS at MIT-AI There's a new version of the EMACS reference manual on AI:EMACS;EMACS GUIDE which reflects the recent changes to the INFO documentation of EMACS.  PGA@MIT-ML 07/18/79 21:48:03 To: info-emacs at MIT-MC I find it somewhat annoying that exiting from emacs RMAIL, DIRED, ALTER OPTIONS, EDIT WORDAB, and other similar modes should not be the same command. C-M-C would be my choice. Is this worth standardizing? Is it difficult?  Date: 26 June 1979 10:42-EDT From: Charles R. Davis Subject: Optimum M-\ To: INFO-EMACS at MIT-MC I use a slightly different tack. I, too, have a binding for M- which (by default) leaves one space at the cursor. However, given an argument, it leaves ARG number of spaces. Thus, at the end of a sentence, I often say M-2 M-, and I very seldom use M-\ as I find M-0 M- to be far more mnemonic.  Date: 26 June 1979 00:31-EDT From: "Guy L. Steele, Jr." Subject: Optimum M-\ To: INFO-EMACS at MIT-AI, CBF at MIT-MC Date: 23 June 1979 03:51-EDT From: Charles Frankston I disagree with making M-\ leave one space. I happen to use a private binding of M- for that purpose, which I think is much more mnemonic. Ie. M- forces there to be one space at the cursor regardless of what (including none) whitespace existed there before. This is very useful when you've typed ahead and don't remember if you've typed a space or not, or if there's the possibility of tabs etc. I use both M-\ and m-Space> usefully threfore. Right! Very good idea. I in aact use META-SPace for something else, but this idea is so good I'd be willing to switch.  Date: 25 JUN 1979 1628-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The change to save as a new version on Twenex has been approved. You will still be able to save as a fixed visited version if you use C-X C-W . Answers to the question are in EMACS;VERSIO ANSWER.  Date: 24 JUN 1979 1336-EDT From: SROSS at MIT-AI (Sandor R. Schoichet) Subject: M-\ To: INFO-EMACS at MIT-AI I would like to agree with CBF about M-\. I like the fact that it gets rid of all space, and also like the idea of having a command that specifically make one space. Along these lines, I would like to make another suggestion: how about having C-x C-o remove all vertical blank lines instead of leaving one if there were more than one to start (a feature that I sometimes find very annoying), and create a new command, perhaps C-x C-space, that would always make there be exactly one blank line.  Date: 24 Jun 1979 0004-PDT From: Meyers at SRI-KL (Harris A. Meyers) To: RMS at MIT-AI, INFO-EMACS at MIT-AI In-reply-to: Your message of 23-Jun-79 2052-PDT That would seem to be to be the more useful defauld. harris -------  Date: 23 JUN 1979 2352-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Tenex and Twenex people: would it be a good idea for C-X C-S and Save All Files to clobber the version number to 0 before saving? The effect of this is that if you visit a file specifying an explicit version number, then saving it will make a new version instead of rewriting the same version visited.  Date: 23 June 1979 03:51-EDT From: Charles Frankston Subject: Optimum M-\ To: info-emacs at MIT-AI I disagree with making M-\ leave one space. I happen to use a private binding of M- for that purpose, which I think is much more mnemonic. Ie. M- forces there to be one space at the cursor regardless of what (including none) whitespace existed there before. This is very useful when you've typed ahead and don't remember if you've typed a space or not, or if there's the possibility of tabs etc. I use both M-\ and m-Space> usefully threfore.  Date: 22 Jun 1979 2202-PDT From: Rubenstein at SUMEX-AIM Subject: Optimum M-\ To: info-emacs at AI How about a combination: ^R commands' arguments generally default to one; why not make M-\ remove all space around point except spaces, and just fix its default to one? If you want the old version, do M-0 M-\. Stew -------  Date: 22 JUN 1979 0355-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Suggestion: in modes which have a comment terminator, CR given just before a comment terminator at the end of a line would move over the comment terminator before doing what it does now. LF would be affected through CR. Is this a good idea?  Date: 19 JUN 1979 1519-EDT From: RICH at MIT-AI (Charles Rich) Subject: Outline Mode Genealogy To: INFO-E at MIT-AI I left out Dick Waters, who made a major design effort.  Date: 19 JUN 1979 1511-EDT From: RICH at MIT-AI (Charles Rich) Subject: outline mode To: PGA at MIT-AI, INFO-E at MIT-AI CC: brian at MIT-ML Such a package has been evolving for a while. The first version was written by Mitch Marcus. I repackaged it as an EMACS library, and Brian Smith has been improving it since. I have copied a middle-aged version to EMACS;OUTLIN :EJ and the source to EMACS1;OUTLIN > for those who want to try it out. Get documentation by the usual method (MM apropos$outline). I am not currently working on this macro package, so please do not send any bug messages to me. I would be quite happy to put this project up for adoption, but before you go charging off I recommend checking with Brian Smith. I think his experimental version is superior in many ways to the one I have released. Yours in hacking, Chuck Rich.  pga@MIT-ML (Sent by GYRO@MIT-ML) 06/19/79 14:47:47 To: info-emacs at MIT-MC COMSAT@MIT-ML 06/19/79 14:45:30 Re: Msg of 14:45:25 To: GYRO at MIT-ML CC: MAIL-MAINTAINERS at MIT-ML WARNING - "INFO-EMACS" at MIT-ML is an unknown recipient. Your message is being returned: ------- OUTLINES I) Has anyone considered (or is there hiding somewhere) an EMACS submode for outline writing? A) It would know about 1) Outline levels (I am on level I.A) 2) Entry numbers (I am on entry number I.A.1) A) Commands might be: 1) Increment Outline Level a) Sets outline down one level b) Chooses the correct character string from the heading string and sets the left margin so auto-fill does the right thing 2) Decrement Outline Level a) Sets outline up one level b) As above 3) Start Outline Entry a) Outputs the string which preceeds an entry, in this case ' a) ' b) Invoked by tab? 4) Set Heading String, sets a variable which provides an appropriate string for each level. 5) Set Indent Increment, sets a variable which decides how far to increment the indent on each change in outline level 6) Insert New Entry, does a Start Outline Entry, but relabels all the entries on the current level by incrementing them so that the new entry can be inserted. B) That's the rough idea. Any comments? ^_  Date: 21 MAY 1979 0231-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI People are overwhelmingly in favor of the changes to M-], C-F and C-B.  Date: 21 MAY 1979 0012-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI What do you think of changing M-] (^R Forward Paragraph) so that it stops at the end of the paragraph rather than before the beginning of the next? Specifically, if there are several blank lines between the paragraphs, it would stop on the first blank line instead of the last. This might make it more useful for going back to the place at which you are inserting new text, after making a correction back up in the paragraph. What do you think of making C-B and C-F treat a CRLF as one character?  Date: 10 MAY 1979 1917-EDT From: rms at MIT-AI (Richard M. Stallman) Sent-by: ___002 at MIT-AI To: INFO-EMACS at MIT-AI The change to incremental search echo area display has been unanimously approved.  Date: 10 MAY 1979 0122-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Do you think it would be a good idea for incremental search to display its direction and status in the echo area only and leave the normal mode line in place? Ie, EMACS (Mode) ... Failing I-Search: FOO  Date: 18 APR 1979 0251-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The latest NE on AI has been observed to have a bug whereby the mode line displays information about a buffer which is not actually selected. When this occurs, please try to remember how you switched buffers last, and any unusual circumstances. Switching buffers again will bring the mode line back into synch. However, the bug cannot be found unless someone can find a clue to how the losing state is entered.  Date: 18 APR 1979 0051-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Answers to my question about auto saving will be put in EMACS;AUTOSV ANSWER on MIT-AI.  Date: 18 APR 1979 0050-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Do you think that auto saving should occur when you switch buffers? The motivation is that if you deselect a buffer which has changes in it and don't come back to it, you might lose the changes if the system crashed. When you answer, please tell me whether you use auto saving.  Date: 12 APR 1979 0319-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: EMACS 133 is NE on AI To: INFO-EMACS at MIT-AI This version is not going to be installed on ML, MC or DM. There are significant changes: *) The self documentation features now know about names for the built-in TECO ^R command definitions. For example, Apropos of "Line" will now show ^R Open Line on C-O. You can't actually call those commands by those names, unless you load the library BARE; this library is temporarily loaded by the documentation commands that need it. MM& Load BARE loads BARE and causes BARE to remain loaded until its caller returns. *) C-U C-Y differs from plain C-Y in that it leaves point before the un-killed text and the mark after. *) C-U C-L causes the line point is on to be redisplayed. Actual numeric arguments (not just C-U) still specify a line to display point on. *) C-; and M-; (^R Indent for Comment) now inserts the comment terminator as well as the comment starter, leaving point in between them. M-N and M-P (^R Down Comment Line, and .. Up ..), have been modified for this behavior. *) M-G and M-Q (^R Fill Paragraph and ^R Fill Region) now by default remove excess spaces such as are produced by filling. *) Return now will not gobble a blank line followed by a nonblank line. *) C-K will now treat a blank line as if it were empty, killing through the CRLF instead of to it. More specifically, if point is followed by a CRLF with only whitespace in between, C-K will kill through the CRLF. *) M-[ and M-] (^R Backward Paragraph, and ... Forward ...), now treat "\" at the front of a line just like "." and "-", for the sake of TEX files. This special treatment is that if those characters are declared (using QParagraph Delimiter) to start paragraphs, then whenever a paragraph starts for that reason, that paragraph is only one line long. *) You can now put MM Query Replace on a character. It will read its arguments in the echo area. *) Errors in auto-saving will type out "(Auto Save Error!)". *) ^R Indent Nested (the definition of Tab in certain modes) now does something useful with negative arguments. It indents the line more than the previous line, (* (ABS ARGUMENT) (MAX 1 QIndent Increment)) spaces more. QIndent Increment is not initially defined. *) There is now an autoloading definition for C-X ( which loads in KBDMAC and transfers to it. People need no longer load KBDMAC in their init files. You will now see "Def" in the mode line while defining a keyboard macro. *) Exiting from the top-level ^R invocation now returns to EMACS's superior. To exit from it back to TECO command level, do 1MM Top Level. This will advise you to re-enter ^R mode by doing :M..L. Returning to the superior will now clear out the mode line completely. *) Creating a new buffer initializes it in the same mode as the previous selected buffer. This uses the new @F^G command in TECO. *) Making variables like QAuto Fill Mode local now works straightforwardly; it is no longer necessary to deal with the crock Switch Modes Process Options, which no longer exists. These variables have been equipped with macros which are run whenever their values change. Thus, simply doing 1UAuto Fill Mode is enough to turn on the mode. MM & Process Options no longer exists. A side effect of this is that buffer switching is much faster. This uses the new FS VAR MACRO feature of TECO. QProcess Options Hook no longer exists; instead, you must make macros to be run when your variables' values change. This will be documented in CONV later. *) Instead of calling MM & Set Mode Line, macros should simply do 1FS Mode Change, which will cause MM & Set Mode Line to be called when the mode line comes up for redisplay. Thus, multiple reasons for calling it will not cause it to be called more than once. QSet Mode Line Hook still exists and works as before.  Date: 9 APR 1979 0541-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI There is a new NE, on AI only. Please tell me if it blows up on you. Some init files may have to be changed because the Switch Modes Process Options crock has been flushed. I'll describe the changes in another message in a few days. This version is still somewhat experimental.  Date: 9 APR 1979 0309-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Does any of you make use of & Toplevel ^R? (other than the use EMACS normally makes of it) Does any of you use ever the ability to escape into normal TECO by exiting from the top level ^R mode?  Date: 7 APR 1979 0456-EST From: JBROWN and CMR at MC at MIT-MC Sent-by: JBROWN at MIT-MC Subject: Tailor made Emacs inits To: INFO-EMACS at MIT-MC We have a fairly winning Emacs init for the ADM2 terminal that allows use of the terminal's built-in keys (up-down-left-right keys, etc) If anyone is interested in it, contact one of us. We would be interested in creating tailor-made Emacs inits for other terminals (or personal preferences for that matter)... Send mail if you would be interested... Jordan Brown (JBROWN) and Clay Ryder (CMR)  Date: 3 APR 1979 0215-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI People's responses to my latest questions can be found in EMACS;CKCRMQ ANSWER. My decision to make the change to M-Q was partially based on the fact that, when I looked at the code, I discovered that it did indeed know completely about ends of sentences.  Date: 2 APR 1979 2350-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI All three changes, to CR, C-K and M-Q, have been approved. Those of you who use C-K for getting rid of blanks should read about M-\.  Date: 2 APR 1979 0443-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI When you answer my question about the CR command, please say whether your usual terminals have insert/delete line and whether you think that has anything to do with your decision. I would like to determine whether this is a matter of different types of terminal or different personal styles of editing. Also, do you think that M-Q should by default flush excess spaces? This would be a little slower, and might sometimes do the wrong thing to double spaces at ends of sentences.  Date: 2 APR 1979 0344-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI What do you think of these changes, suggested by EAK: 1) C-K would not consider blanks when deciding whether it was at the end of a line. In other words, C-K when point was followed by blanks and a CRLF would kill through thr CRLF. 2) CR would not gobble up a blank line followed by a nonblank one. In other words, typing text with CR's at the end of a paragraph would not use up the blank line separating it from the next paragraph. I can imagine that 2) might bother people who on terminals without insert/delete line are used to inserting several lines of text by first making the lines with C-O and then filling in the text. Do people actually use that technique?  Date: 6 MAR 1979 2049-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Everyone agrees that "(AUTO SAVE)" should be printed out, in general. What about specifically on printing terminals? It has been advanced that on printing terminals printing "(AUTO SAVE)" gets in the way of editing.  Date: 5 MAR 1979 2247-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Do people like having (AUTO SAVE) printed out when there is an auto save?  Date: 5 Mar 1979 1514-PST From: Rubenstein at SUMEX-AIM Subject: Changes to PAGE library To: info-emacs at AI The go to next and go to previous page commands have been moved to C-X ] and C-X [, respectively. These commands were previously page handling commands but are now replaced by the commands from the page library. Also, if you like ^R String Search instead of the default Incremental Search, the PAGE library remembers that. If you use some other search, or you put your search on a character other than C-S or M-S, then you will get a message saying that PAGE cannot find your search. In that case, you must do M.M^R My Search$ M.VReal Search Macro$ for the PAGE search to work right. Then you can put ^R PAGE search$ and ^R PAGE Reverse Search$ on whatever characters you want. Also, if ^R PAGE Search is given a precomma arg, it macros that instead of Q$Real Search Macro$. Thus, other macros can use this facility to do things to the whole buffer, closing the bounds when finished. Comments, questions, etc. to Rubenstein @ Sumex-AIM -------  Date: 1 Mar 1979 1430-PST From: Rubenstein at SUMEX-AIM Subject: New PAGE library To: info-emacs at AI I've just finished a library of EMACS macros to handle pages in an intelligent manner (it turns out to be look very similar to the way E (SAIL's display editor) handles them). The page number appears in the mode line, and commands are available to go to a specific page, go to next or previous page, insert pagemark, and Join pages. The file is AI:EMACS;PAGE :EJ, and documentation is in the INFO tree, Topic EMACS.Page handling (INFO;PAGE >) Bugs, comments -=> Rubenstein@Sumex or STEW@MIT-* -------  Date: 15 FEB 1979 2017-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: Comment commands in text mode To: INFO-EMACS at MIT-AI Date: 14 Feb 1979 1603-PST From: Rubenstein at SUMEX-AIM The documentation for Text mode states that there are no comments. This is good. However, a new user types M-N and gets immediately very confused. Why doesn't MMText Mode disable comment commands??? My philosophy on text mode was that nothing should automatically recognize any comments, but that commands that are specifically for comments, such as M-;, C-M-; and M-N, might as well still do what the user asked. I am willing to change this if there is a real concensus. Do you ever use those commands in Text mode?  MOON@MIT-MC 02/10/79 18:03:38 Re: Meta-G (CWH) To: INFO-EMACS at MIT-MC I must point that Carl is confused. EINE's actual scheme is to keep track of whether the region is "really" there, and to display it with underlining or inverse video when it is "actually there". This is hard for Emacs to do, as it depends on keeping the mark (the region) and the point-pdl (old places you may want to go back to) separately. Emacs does ask you before saving the region on meta-G and similar commands, if the region is very large (a certain number of characters, I forget exactly how many.)  CWH@MIT-MC 02/10/79 17:46:40 Re: Meta-G To: INFO-EMACS at MIT-MC, MOON at MIT-MC I don't think the problem is so much the current binding of Meta-G to Fill Region, as in unintentionally keeping a large region around for a long period of time. I think that we should adopt EINE's scheme whereby the mark follows the point whenever a self-inserting charater is typed (or perhaps a non-cursor-moving character -- I'm not exactly sure what the mechanism is). This will not only solve the Meta-G problem, but will keep naive users from losing so much from Control-W and other commands which affect the region.  MOON@MIT-MC 02/09/79 02:38:59 Re: Meta-G To: INFO-EMACS at MIT-MC It has been pointed out that typing meta-G accidentally can screw you. This is true, however the old text is saved on the kill-ring, and you can get it back with control-Y. MMcM and I just fixed a bug so that you can type meta-Y immediately after a meta-G, and the command will be fully un-done. This fix is in the Emacs source code but is not yet installed. This also applies to control-X control-L and control-X control-U (lower and upper case region). If anyone has opinions about whether this makes the meta-G command livable-with on its present key, or not, mail them to BUG-EMACS.  Date: 31 JAN 1979 1721-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI How about if Select Buffer, in two-window mode, would switch windows if the desired buffer were in the other window? This would make everything that switches buffers, including C-X C-F and Tags, behave that way. It would still be possible somehow to get the same buffer in both windows.  JLK@MIT-MC 01/25/79 11:37:07 To: INFO-EMACS at MIT-AI I have flushed the TEMACS macro library. Instead there are now the libraries: SCRLIN - macros for moving over screen lines instead of text lines. Automatically sets up ^A, ^E, ^N, ^P. On non-display consoles, these macros revert to text line movers. DELIM - macros for moving over balanced pairs of []{}()<>"'. Sets up C-[, C-], C-(, C-), etc. to move over pairs (forward or back). JLKMAC - Mostly defines setup hooks for JLK. Anybody currently using TEMACS probably wants to simply load SCRLIN or DELIM. Other people who wanted to try out screen line moving or delimiter moving, but didn't want to get screwed by JLK's hacks, now have a chance.  OTA@MIT-MC 01/16/79 05:09:01 Re: auto save mode To: INFO-EMACS at MIT-MC It seems like the major problem with auto save mode where output is going to the regular filename is that it sometimes deletes files it didn't output. Couldn't is just keep around the version numbers of all the files it wrote out by itself. This should not include ^X^S files. Then it would leave files that were ^X^Sed, or otherwise written out, behind as the auto saved version numbers increased. It would only have to keep 2 or 3 numbers around. If this really worked there would be no reason to keep around lots and lots of them. You could ocasionally ^X^S when you get to a good spot and those would stay around. It seems to me that this is essentially equivalant to the separate filename mode but much handier.  Date: 13 JAN 1979 1851-EST From: HENRY at MIT-AI (Henry Lieberman) To: INFO-EMACS at MIT-AI, (BUG LISPM) at MIT-AI A suggestion for a feature: When in two window mode, when I want to edit the code in the other window, I often make the mistake of typing Control-X B rather than Control-X O, which is what I'd do with a single window to edit that code. How bout a command which took a buffer name, and if that buffer is in the other window, switch windows, otherwise switch buffers.  Date: 15 DEC 1978 0309-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I would like to flush ^R Make Insertion Command. Who uses it? It is not on any key in the default environment.  Date: 14 DEC 1978 0306-EST From: MOON at MIT-AI (David A. Moon) Subject: New DIRED installed (ITS only) To: INFO-EMACS at MIT-AI The variable DIRED Long Format, if defined and non-zero, inhibits the multiple-files-per-line compressed format in the confirmation phase of DIRED. This should make KLH happy. Also in the compressed form the $ (don't reap) flag is now displayed; think twice about deleting anything with this flag set on it.  Date: 10 DEC 1978 0019-EST From: MOON at MIT-AI (David A. Moon) Subject: DIRED 152 ([DIRE] 96) installed (ITS only) To: INFO-EMACS at MIT-AI A new version of DIRED has been installed. The following changes exist: The $ command complements the don't-reap attribute of the file, which displays as a $ to the left of the creation date. The X command now does what meta-X does in the default environment, use Q to exit. The Help key now gives information on Dired. Files whose FN1 begins with an underscore are considered temporary; this applies to MM Clean and MM Reap also. The H command been's changed a bit; I think it still works.  Date: 29 NOV 1978 2025-EST From: MOON at MIT-AI (David A. Moon) Subject: New NE exists To: INFO-EMACS at MIT-AI A new TS NE has been installed which contains changes relevant to users of ^R Character Search and ^R String Search (the default incremental search is not affected.) The bugs in the ^F feature have been fixed, and Word Search (entered by ^W in String Search) has been introduced; try it. Better ways to get into it are solicited.  Date: 21 NOV 1978 0244-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Do you think that Rubout and ^D, the only commands which can delete arbitrary text and do not save it, ought to be limited as to how big an argument they can be given? Ie, giving either one an argument larger than 16 might be an error. It would still be easy to get rid of more characters than that by setting the mark, moving, and killing, but that would save the killed text. Please reply to me, not to INFO-EMACS. All answers will be put in AI:EMACS;RUBOUT ANSWER.  Date: 20 NOV 1978 0207-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI C-X 4 improved yet again: if you are already in two-window mode, it will go to the other window and select the buffer, file or tag. Also, C-X 4 never displays anything in the other window until you have finished saying what you want selected there.  Date: 19 NOV 1978 0203-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 129 is now NE. There is a new command ^X4 which enters two window mode and then lets you select a specific buffer, file or tag. Follow the ^X4 with B and a buffer name, F or ^F and a file name, or T or . and a tag name. Meta-' (^R Upcase Digit) now works on all keyboards. The first time you use it, it asks you to type the digits 1, 2, ... 9, 0 holding down the shift key so it can tell what your keyboard looks like. For then on it works automatically. If you don't recall, this command finds the last digit on the current or previous line and shifts the character. It is for when you forget to use the shift key.  ELLEN@MIT-MC 11/16/78 21:56:06 Re: Slow terminal speeds To: INFO-EMACS at MIT-MC A special interest group does exist for people who use slow terminals, called INFO-300. However, this question bears on all EMACS, since if the INFO-300 crowd set things up their way, the high baud crowd would complain (as someone pointed out this thing does have elements of baud rate boundaries). So the argument belongs quite rightly in INFO-EMACS. For strictly low baud rate and non-controversial EMACS options, however, do use INFO-300.  Date: 16 NOV 1978 2136-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I find that the people who like the ^X2 change are those who usually use ^X^F, most of whom set Tags Find File to 1. The people who don't like the change are those who do not use ^X^F, and they never set Tags Find File. So I will make ^X2 depend on Tags Find File. However, if you don't use ^X^F, I recommend that you start. I will also put in a command that does ^X2 and ^X^F, some time.  KWC@MIT-ML 11/16/78 15:30:46 To: info-emacs at MIT-MC Perhaps there should be a new mailing list for those you that care about slow terminals.  Date: 16 NOV 1978 1528-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Please do not mail answers to my questions, or discussions pro and con, to INFO-EMACS. A lot of people dislike even the questions. It is not fair to burden them with all the discussions. I put the discussions on these topics in files on EMACS; so that people can see them.  EAK@MIT-MC 11/16/78 15:00:16 Re: 2 To: INFO-EMACS at MIT-MC I dislike KLH's proposal. More often than not I use  rather than B after entering two-window mode the first time. I would find something that forced a B upon me to be obnoxious. One alternative would be to do an automatic  since that often can substitute for a B. However it wouldn't work all the time, so I advocate leaving it as it is, or making 2's actions user-controllable.  GSB@MIT-ML 11/16/78 14:04:04 Re: ^X2 on slow terminals To: JLS at MIT-AI, INFO-EMACS at MIT-AI I find i am just as likely to use 2 windows at 300 baud as i am at 9600, if not more so. At 9600 baud i tend to switch buffers more and use a full screen so that i can get more context; if i am single-buffer hacking at 300, i generally increase the number of echo lines to something like 10, or maybe just leave myself with 2 windows, one of them possibly empty.  RWK@MIT-MC 11/16/78 12:04:03 Re: JLS's note re ^X2 vs. Slow terminal usage. To: INFO-EMACS at MIT-MC The SLOWLY library is shortly going to have exactly what JLS was looking for. I.e ^R Narrow Display which will take an argument to say how much of the screen to use. A negative argument will mean to use the bottom of the screen instead of the top. Also, SLOWLY has ^R Slow Incremental Search, which uses a narrowed part of the screen for the search (say, a single line, or a couple lines) for it's redisplay of what's been found so far, and then displays the rest of the context when the search is exited. It will also have a version of & Minibuffer which uses the bottom or the top of the screen as you wish. So you can use the top of your screen for editing, and the bottom part of your screen for minibuffers. Since I'm trying to get this stuff ready to go with the tapes RMS is going to take with him to the West Coast, I'm going to need volunteers to test this stuff thouroughly before then, on both ITS and TWENEX. If you use slow terminals and would like to volenteer, send me a note; I should have it finished and documented sometime tonight.  Date: 16 NOV 1978 1115-EST From: JLS at MIT-AI (Jim Stansfield ) To: RMS at MIT-AI, INFO-EMACS at MIT-AI WITH REGARD TO THE TWO WINDOW MODE CONTROVERSY. I THINK THE PROBLEM AROSE BECAUSE OF A CONFLICT BETWEEN THOSE WHO WANT TO USE TWO WINDOWS AND THOSE WHO WISH TO AVOID LONG PRINTOUTS BECAUSE OF SLOW TERMINALS. I NEED BOTH OF THESE MODES SINCE I SOMETIMES USE A DATAPOINT AT HOME AT 300 BAUD. IF THIS IS THE SITUATION A FIX WOULD BE TO HAVE ^R GROW WINDOW WORK IN ONE WINDOW MODE SO THAT WITH A NEGATIVE ARGUMENT PEOPLE COULD SHRINK THE SINGLE WINDOW TO ANY SIZE DESIRED. THEN THERE WOULD BE A CLEAR LOGICAL SEPARATION BETWEEN THE TWO USES. I DOUBT ANYONE WOULD REALLY WANT TWO WINDOWS IF THEY WERE USING A SLOW TERMINAL.  ECC@MIT-MC 11/16/78 04:19:20 Re: WORDAB library To: INFO-EMACS at MIT-MC A few bugs have been fixed (tab echoing, some bugs inhibiting proper upper- casing) and a small feature installed regarding upper-casing: If you type an abbrev all in caps, e.g. "FOO", and it expands to more than one word, e.g. "find outer otter", then the default is the old action: just capitalize each word, e.g. "Find Outer Otter". However, if the variable WORDAB All Caps is non-zero, the entire expansion is upper-cased, e.g. "FIND OUTER OTTER", which may be much more convenient for those editing all-uppercase code for instance. Please let us know in the next couple of days if there are any bugs in the new fixes/feature as RMS wants to make a tape of working libraries for the outside world.  Date: 16 NOV 1978 0030-PST Sender: GEOFF at USC-ISI Subject: Re: ^X2 From: Geoff at SRI-KA (Geoffrey S. Goodfellow) To: KLH at MIT-AI Cc: INFO-EMACS at MIT-AI Message-ID: <[USC-ISI]16-NOV-78 00:30:43.GEOFF> In-Reply-To: Your message of NOVEMBER 16, 1978 i think your solution/compromise to the new vs. old ^X2 problem is the best i've heard thus far, esp. wrt the superfluous creations of the 2nd window with the contents of the first over the net and on slow devices. it would just mean one extra character for those that like the new way to type (CR), and give us others who liked the old way better a way for us to type in 3 extra chars (W2 CR). g  Date: 15 Nov 1978 2326-PST From: Mclure at SRI-KL (Stuart McLure Cracraft) Subject: Re: ^X2 To: KLH at MIT-AI, INFO-EMACS at MIT-AI In-reply-to: Your message of 15-Nov-78 2315-PST I think that is a reasonable compromise that provides for winnage. I'm not one to advocate a lot of hairy questions asked whenever meta-mumble or whatever is invoked but it's preferable to what exists now where you will want to get another buffer anyway. -------  Date: 16 NOV 1978 0215-EST From: KLH at MIT-AI (Ken Harrenstien) Subject: ^X2 To: INFO-EMACS at MIT-AI I think everything about ^X2 currently is slightly wedged. For the record, I liked NEITHER the old nor the new schemes. What I really want is some way of specifying, when I ask for two-window mode, which buffer I want in the second window. In other words it should also be a buffer select and ask for a buffer name; a simple would default the name to whatever the new scheme uses (i.e. depending on ^U argument). Thus to initially effect the current "new" scheme, one types C-X 2 CR; for the old, C-X 2 W 2 CR (to get/create buffer W2). My primary reason for this is that over the network or on a slow speed terminal it is agony to watch EMACS re-display something that you don't want, which is often; being able to arbitrarily select the buffer before display starts will avoid this. I've tried both the old and new schemes, and get screwed by both of them (more by the new than the old, even); the change I propose seems very cheap and simple and flexible, and avoids the screwage problems I have. Do other people see any problems in it? (I don't see the extra character as a problem).  Date: 16 NOV 1978 0058-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: ^X2 To: INFO-EMACS at MIT-AI Thank you for your replies. People are overwhelmingly in favor of the change. However, non-MIT people are mostly opposed. I have a few more questions: 1) Do you normally use ^X^F to visit files, or do you normally visit a new file in the same buffer? 2) Do you set the value of Tags Find File to 1 in your init? 3) Did you like the change? (so I don't have to match up your new answers with your old answers). 4a) If you liked the change, did you know about the ^U^X2 command? 4b) If you did not like the change, do you know that ^X2 and ^U^X2 are NOT now identical, except for the first time you enter two-window mode? Do you know about the ^XB command for switching buffers? In case you don't know, ^X2 puts you into two window mode showing in window two whatever buffer was in it the last time; showing the same buffer as in window one, if there was no last time. ^U^X2 puts you into two window mode and always shows in window two the same buffer that was in window one when the command was given. 5) Now that you know that, what is your opinion of the change?  Date: 15 NOV 1978 1825-EST From: JERRYB at MIT-AI (Gerald R. Barber) To: INFO-EMACS at MIT-AI With respect to: Date: 14 NOV 1978 2324-EST From: MRC at MIT-AI (Mark Crispin) To: INFO-EMACS at MIT-AI The latest version of EMACS released by MIT has a new misfeature that ^X2 puts the second window in the same buffer instead of creating a new one. Apparently some people at MIT pressured for this change. A lot of people, myself included, do not like it and want it reverted back to the way it used to work. If you've been screwed by this change and haven't complained yet, or you're at one of the lucky sites which hasn't gotten this new version but agree it will be a loss, complain to BUG-EMACS@MIT-AI. I don't agree! I think this is much better than it was. There is a problem tho: The documentation for ^R Two Window states that with an arg > 1 the same buffer is chosen, in fact whatever the argument, the same buffer is chosen. I do think that with a non-zero arg it should ask for a buffer, similar to ^X B.  Date: 14 NOV 1978 2324-EST From: MRC at MIT-AI (Mark Crispin) To: INFO-EMACS at MIT-AI The latest version of EMACS released by MIT has a new misfeature that ^X2 puts the second window in the same buffer instead of creating a new one. Apparently some people at MIT pressured for this change. A lot of people, myself included, do not like it and want it reverted back to the way it used to work. If you've been screwed by this change and haven't complained yet, or you're at one of the lucky sites which hasn't gotten this new version but agree it will be a loss, complain to BUG-EMACS@MIT-AI.  Date: 14 NOV 1978 0037-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: New KBDMAC features To: INFO-EMACS at MIT-AI A repeat count of zero means repeat forever (until an error). C-X) with a repeat count executes the macro that many times right away, counting defining the macro as one time. C-X Q lets you define macros that work like Query Replace. All are documented in INFO under EMACS KBDMAC.  Date: 10 NOV 1978 0119-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI When making a variable local its name must now be written in full. This change is a result of fixing a bug that made it impossible to create a variable whose name was an abbreviation of some other variable's name.  Date: 9 NOV 1978 0437-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Aside from KBDMAC and DIRED, INFO;EMACS > is now up-to-date as far as I know. I would appreciate it if people would read through it and see if they can find any errors.  Date: 8 NOV 1978 0207-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI How would people like ^X^Q to make the current buffer read-only and unmodified, instead of turning off auto-writeback? ^U^X^Q would make it writeable again. Answers will be put in AI:EMACS;^XQ ANSWER.  Date: 8 NOV 1978 0157-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI There is a new NE, version 128, which fixes all known bugs in 127. Everyone please try it. I am in a hurry to get it well debugged.  Date: 6 NOV 1978 0341-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI It now works well to use the name of a machine instead of DSK: in source filenames in a tag table file. This is because the TECO feature that canonicalizes them, so that EMACS will realize that DSK:FOO;BAR > and AI:FOO;BAR > are the same file when you are on AI.  Date: 6 NOV 1978 0339-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: New in EMACS 127 To: INFO-EMACS at MIT-AI 1) ^X[ and ^X], new, move to the previous and next page boundary. ^X^P will now always mark the page that point is in. 2) Query Replace and Replace String, when Case Replace is set, will now preserve case only if the string to be replaced contains no upper-case letters. Thus, you can use Query Replace to change something to lower case without turning off Case Replace. You can also use it to convert something to upper case, since the case-preservation works by upcasing only. Anything you specify as upper case in the string to replace with will always go in as upper case. 3) Meta-' is now ^R Upcase Digit, which finds the last 7, 8 or 9 on the current line before point, or the previous line, and replaces it with a ', ( or ). 4) ^Z and ^X^C, ^R Return to Superior, will now do an auto save first if auto saving is on. 5) ^X Altmode, ^R Re-execute Minibuffer, will now print out the command that is to be re-executed and ask you for confirmation in the form of a Y or N. 6) M-G and M-Q, which use ^R Fill Region, will now work better when a Fill Prefix has been established (with ^X.). 7) It now works to set Underline Begin and Underline End to strings longer than one character. 8) Visiting a file now calls the value of Visit File Hook if that value is nonzero. The call is done after all the normal work of visiting. 9) Local modes in a file can now include local ^R commands. Simply use, for example, .^RA instead of the variable name to bind control-A.  Date: 6 NOV 1978 0240-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS in TECO 709 is now installed on all machines. All init files have been renamed. A new EMACS version (127) will appear as NE soon. Everyone, please do not make any changes to the EMACS sources for the next month without checking with me in advance. A new distribution tape is being prepared.  Date: 6 NOV 1978 0206-EST From: rms at MIT-AI (Richard M. Stallman) Sent-by: HIC at MIT-AI To: INFO-EMACS at MIT-AI With NE, the "right" name of the default init file is EMACS;* EMACS. EMACS;.EMACS (INIT) will continue to exist, but you should switch your init files over to referring to the file by the new name.  Date: 3 NOV 1978 0840-EST From: RMS at MIT-AI (Richard M. Stallman) Subject: KBDMAC library To: INFO-EMACS at MIT-AI Use ^X( to start defining a keyboard macro, and ^X) to end the definition. ^XE executes the last keyboard macro defined. An argument serves as a repeat count. To save a keyboard macro for a longer time, use MM Name Kbd Macro. It takes as a string argument the name to use for an MM command to define to run the macro, and reads from the terminal a ^R command to define. Type Return or Rubout if you don't want to define a ^R command. Use a null string argument if you don't want to define an MM command (though you should always define one, even if your intended use is as via the ^R command). View Kbd Macro will print out the contents of a keyboard macro, reading the ^R command from the terminal. You can give ^XE as the ^R command to view the last macro defined. ^X commands can be used as the ^R commands to put macros on. However, you aren't allowed to clobber any already defined command this way, unless the definition is another keyboard macro or similar thing.  Date: 1 NOV 1978 1706-EST From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI FOR NE, LIBRARIES CALLED .EMACS :EJ SHOULD BE CHANGED OVER TO LIBRARIES WHICH GO UNDER THE SAME NAME AS INIT FILES. THIS INVOLVES CHANGING THE & SETUP MACROS IN A SPECIFIC WAY DOCUMENTED IN EMACS;CONV >  Date: 1 NOV 1978 0022-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI NE now resides in TECO 708. This means that you can use the library KBDMAC to get keyboard macros (your keystrokes are remembered and replayed). Also, the new init file scheme is installed, so your init file must be called EMACS on your home directory.  Date: 27 OCT 1978 2207-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: ^U^U^@ - pop and discard mark To: INFO-EMACS at MIT-AI Does anyone use this command? It has been suggested that it pop the mark twice instead. ^U^U^U^@ would pop three times, etc.  Date: 24 OCT 1978 0006-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Please try NE, which is the same EMACS in TECO 699.  Date: 20 SEP 1978 1710-EDT From: EAK at MIT-AI (Earl A. Killian) Subject: new function in TAGS: MM Args To: INFO-EMACS at MIT-AI TAGS now contains MM Argstag which prints all the comments before tag. For assembly language programs this is usually a subroutine description giving the arguments and results. Someone might want to hack it to know about LISP mode and print the argument list in addition to the comments.  Date: 13 SEP 1978 0143-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: New command Compile for compiling your program. To: INFO-EMACS at MIT-AI MM Compile will compile the program you are editing. Unless given a numeric argument, it first offers to write out any files that have been changed. Then it compiled the visited file in whatever way is appropriate for the major mode you are in. This method of compilation can be overridden by setting the variable Compile Command to a TECO command string to perform the compilation. It can expect the name of the visited file to be in Q1 and in the default file names. Most compilation methods will involve valretting commands to DDT. They normally leave you in DDT so you can load up or test the program. A string argument to MM Compile is stuck into the compilation command so as to be useful for specifying extra switches or an output file. Thus, in a MIDAS program, MM Compile/T would give MIDAS the /T switch, and MM CompileFOO;_ would put FOO;_ at the front of the command string, thus putting the binary on FOO;. User-supplied compilation methods can't be vouched for, but ought to try to do this. MM Compile is available only in NE, as yet. It will not self-document properly until the next EMACS version.  Date: 13 SEP 1978 0059-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI 1) You can now have local character redefintions in files. They look almost like ordinary local modes, except that you use the name of tha character (such as, "...^RD") as the variable name, and some TECO code to come up with the value instead of a literal string. For example, this ;Local Modes: ;Mode:MIDAS ;...D: M.M ^R Down List ;End: willset up a mode like MIDAS mode except that C-M-D retains its normal definition instead of MIDAS mode's redefinition (This mode is good for editing Lisp machine microcode). 2) Dump Environment has been greatly improved by EAK. The documentation in CONV has not been updated yet, but its self-documentation is informative. 3) Compile One Macro has been renamed TCompile.  Date: 2 SEP 1978 2140-EDT From: MOON at MIT-AI (David A. Moon) Subject: Changes to ^V, string-search To: INFO-EMACS at MIT-AI In the next Emacs, the following changes will exist. If you define the variable Next Screen Context Lines, it will be the number of lines of overlap used by c-V (^R Next Screen) and m-V (^R Previous Screen). The default remains 2. In Character Search and String Search (note that these are different from the default search, Incremental Search), the ^F feature now will put the cursor Next Screen Context Lines down from the top instead of at the very top. If you don't like this, complain and I'll flush it. String Search now obeys its numeric argument (a long-standing bug fix). String Search has two new features inspired by Multics Emacs. Typing ^Y will yank in the default without flushing what you've already typed. Compare this with ^D. Typing ^W will put you in word-search mode, in which the words you type will be found regardless of what punctuation and white-space characters lie between them. meta-F's definition of "word" is used, thus Atom Word Mode applies, but in Text Mode punctuation is considered part of the adjacent word. Currently word-search only finds exact matches, (e.g. searching for "to and fro" doesn't find "grotto and frobs"). If anyone thinks it should be the other way, I can change it.  Date: 26 AUG 1978 1952-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: New NE To: INFO-EMACS at MIT-AI There is a new NE with these changes: 1) The word case-conversion commands do not move the cursor when given a negative argument. 2) C-M-V scrolls the other window, just like C-V. No argument scrolls a full screen; an argument is the number of lines to scroll by. 3) C-X 2 the first time gives you the same buffer in both windows, instead of creating a buffer named W2. 4) M-X now no longer echoes if you type everything ahead.  Date: 23 AUG 1978 0249-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Often, people want to specify the exact name of a tag instead of just a part of the name, because they know there will be ambiguity. I am thinking of arranging to have a character which would mean the beginning or end of the tag name when typed in the argument to Meta-.. It would probably be impossible to use that character itself in a tag. Do you like this idea? What character would you like?  Date: 18 Aug 1978 1120-EDT From: CICCARELLI at BBN-TENEXD (Eugene Ciccarelli) Subject: Auto Save Mode To: INFO-EMACS at AI Earl Killian and I are designing a new Auto Save Mode (probably to be put in the TMACS library). Anyone interested (e.g. uses my version in the CSRLIB library, is dissatisfied with other Auto Save modes, or is just a hacker) is invited to read and comment on the design ideas, which are in the following files: MC: ECC; AUTOSA IDEAS [BBND]Auto-Save.Ideas.0 [BBNE]Auto-Save.Ideas;0 Or, if accessing one of these files is impossible/inconvenient, I can mail a copy to anyone requesting. Comments or suggestions about these ideas can be mailed to ECC@MC, EAK@MC, Ciccarelli@BBND, or EKillian@BBNE. - Gene Ciccarelli -------  Date: 16 AUG 1978 0117-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI RICH@MIT-AI 08/15/78 09:13:31 Re: ^R Mark Page To: (BUG EMACS) at MIT-AI I think it is a bug rather than a feature that ^R Mark Page marks the -next- page if it is at the top of a page. Could you please gather opinions on this an change it if the vote is in favor. Ok, what do you think?  Date: 9 AUG 1978 0306-EDT From: ECC at MIT-MC (Eugene C. Ciccarelli) Subject: New Word Abbrev Mode ^R Command To: INFO-EMACS at MIT-AI CC: Ciccarelli at BBN-TENEXD The WORDAB and XWORDA libraries now have a new command, which was suggested by KMP: ^R Inverse Add Mode Word Abbrev, which is put on ^X - (previously unused in the default EMACS environment). I think people will find this a very useful command to complement ^X^A. This command has an opposite sense from ^X^A (^R Add Mode Word Abbrev), namely: you define an expansion for the word before point. (For instance: you thought you had an abbrev "foo", but "foo" didn't expand. So you type ^X -, "Find Outer Otter", return, and thus define the abbrev "foo".) You can give ^X - a numeric argument, which means the abbrev to be defined is n words back. ^X - will end by expanding the abbrev.  Date: 1 AUG 1978 1249-EDT From: KLE at MIT-AI (Edward Jay Kleban) Subject: Case conversion commands To: INFO-EMACS at MIT-AI, RMS at MIT-AI, BSG at MIT-AI, CWH at MIT-MC CWH@MIT-MC 07/31/78 06:17:42 Re: Case conversion commands To: INFO-EMACS at MIT-AI, RMS at MIT-AI, BSG at MIT-AI How about making M-U and M-L apply to the current or previous word. However, if the current or previous word is already uppercased (or lowercased), then apply the command to the word following the cursor. Foo! How about making M-U anc M-L do something reasonable that is predictable (at least as far as where you will end up afterward) and independent of what words on the line already look like as far as case.  Date: 31 JUL 1978 2141-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The library NCASE contains the experimental word case conversion commands, and puts them on M-L, M-C, and M-U unless you supply an NCASE Setup Hook. They apply to the previous word when the last character typed was not control, meta, or altmode. In addition, when given in the middle of a word, they apply to the whole word. Please take the trouble to try them out. If you try them long enough to get used to them, I'll be interested in hearing how each of those two changes affects their convenience of use.  Date: 31 JUL 1978 2017-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Answering questions about changing EMACS. To: INFO-EMACS at MIT-AI When I ask for people's opinions on a possible change to EMACS, I don't just want to get a "yes" or a "no". Such answers don't really help me. I don't regard the answers as votes, and I don't just count the "yes" and "no" answers. What I am looking for by asking these questions is an idea of which segments of the user community will be helped and which will be hurt. I want to know what it is about a user that determines what his opinion will be. For this, I need more information in answers. In addition, a lot of people just say "Don't change it. I don't think anything should ever be changed". Such a statement does not address the question I have asked, which is whether the changed command would be better than the existing one. Since I have already decided that it is better to make improvements than never change anything, such statements are wasted breath. Some replies to my latest question (about case-conversion commands) which did supply some information are those which say "Don't change them. Typing M-B is easy enough". These replies generally come from people who use Meta-keyboards, and it is logical that they should have that opinion. A few came from people whose keyboard usage I do not know, but would like to find out, since how easy people on non-Meta keyboards regard the typing of M-B is clearly the center of the issue. I will probably make the changed case commands available in a library for people to try out.  CWH@MIT-MC 07/31/78 06:17:42 Re: Case conversion commands To: INFO-EMACS at MIT-AI, RMS at MIT-AI, BSG at MIT-AI How about making M-U and M-L apply to the current or previous word. However, if the current or previous word is already uppercased (or lowercased), then apply the command to the word following the cursor.  Date: 30 JUL 1978 0109-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Case conversion commands To: INFO-EMACS at MIT-AI CC: BSG at MIT-AI The current way that the word case conversion commands (M-C, M-L, M-U) work is good for converting a lot of upper case to mixed cases, but isn't so good for when you have just typed a word in the wrong case. At such a time, it would be better for the case conversion commands to apply to the previous word instead of the next one. What would you think of making these commands apply to the previous word if and only if the last character typed was a printing character (including Space)? I realize that this introduces a measure of unpredictability, but only very little, since it almost always does the right thing. Use with numeric arguments will be unchanged, of course. People have in the past investigated compromises in which the commands would act differently at the front of a word and at the back of a word, but such schemes all do the wrong thing frequently.  Date: 25 JUL 1978 0016-EDT From: RWK at MIT-MC (Robert W. Kerns) Subject: Incremental search and redisplay To: INFO-300 at MIT-MC CC: INFO-EMACS at MIT-MC The SLOWLY library is a library (maintained by me, more or less) for slow terminals. Currently it contains just one feature, a version of ^R Incremental Search and ^R Reverse Search which limit re-display while searching to a fixed number of lines at the top of the screen. The default is to use only one line, but it may be changed by giving Q$Slow Search Lines Used$ a value other than 1 (use M.V to create it). If MM Load Library$SLOWLY$$ is used to load this library, it will put these macros on ^S and ^R IFF your OSPEED is unknown or slower than 1200. For more info, see me. If you have any other ideas, I might be willing to implement them, so send them too...  Date: 20 JUL 1978 1836-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: Final two-window change plans To: INFO-EMACS at MIT-AI It seems that what people generally want is for ^X 2 to ask for the name of a buffer. If you type just , you will get the same buffer that was in window 2 previously, or some default the first time. ^U^X2 will continue to put you in window 2 with the same window 1 buffer in both windows. In addition, C-M-V will scroll the other window by the specified number of lines (positive or negative).  Date: 17 JUL 1978 2235-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 125 is now E on all ITS systems.  Date: 16 JUL 1978 0151-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: ^X2 To: INFO-EMACS at MIT-AI Some people have suggested that ^X2 ask for a buffer name only the first time it is used. I think that this is a bad idea. When a user starts to use two window mode for a given purpose, he may or may not have been using it for another purpose earlier in the same EMACS. Whatever the user wants ^X2 to do then, it must be independent of whether such prior use had taken place. But this suggestion would make ^X2's behavior depend on that, so it must be wrong. Also, this is the sort of question which you would want to type ahead the answer to; but often you would be surprised to find that ^X2 had been used before and the question was not asked. So, if ^X2 is going to ask a question, it should always ask it.  Date: 15 JUL 1978 1616-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI It has been suggested that ^X2 always ask for the name of the buffer to select in the second window, or that when creating the second window for the first time it should contain the same buffer as the first window until you select some other one. What do you think of this idea? I will put your answers in AI:EMACS;^X2 REPLY.  Date: 15 JUL 1978 0358-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 125 is now NE on all ITS machines. The only visible differences are: C-= and C-X = now run a different command which prints, not just the hpos, but the vpos, the octal code for the next character, point both absolutely and as a percentage of the total size, and the virtual buffer boundaries if any. C-T at the end of a line with no argument will exchange the previous two characters.  Date: 13 JUL 1978 2224-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI .EMACS (INIT) files which are libraries work fine in EMACS 124. These libraries MUST have & Setup .EMACS Library commands in them. Full details are in EMACS;CONV > and are accessible through INFO.  Date: 12 JUL 1978 0000-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI The default EMACS init file wil now set Inhibit Write from Find File Inhibit Write, and if a file is visited under JCL control that too will use Find File Inhibit Write.  Date: 6 JUL 1978 0445-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI NE has become E on all ITS systems. An additional new feature, for printing terminals, is that Help M prints what on a display would be in the mode line.  Date: 4 JUL 1978 0212-EDT From: rms at MIT-AI (Richard M. Stallman) Sent-by: ___046 at MIT-AI To: INFO-EMACS at MIT-AI Who uses either ^R Character Search or ^R String Search and does not load MOON;LUNAR :EJ?  Date: 3 JUL 1978 0440-EDT From: RMS at MIT-AI (Richard M. Stallman) Subject: .EMACS (INIT) files which are :EJable libraries To: INFO-EMACS at MIT-AI These files will have to do all the initialization themselves, so their & Setup macros will have to load in and jump to the default init file. In addition, the feature may require some tweaking to get working right, for which I would appreciate the help of someone who has such an init library, since I don't.  Date: 3 JUL 1978 0151-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI Does anybody use C-X : (^R Set Column)? Does anybody use M-( or M-) (^R Make () and ^R Move over )?)  Date: 3 JUL 1978 0138-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 124 is now NE on AI. Tabular Text Mode and Indented Text Mode are now gone as major modes. Instead, there are Edit tabular Text and Edit Indented Text, which are commands that call ^R recursively with the appropriate things rebound. They are the first examples of "submodes", which are commands that call ^R recursively on the buffer with auto-saving still turned on. When you are inside a submode, the name of the submode appears inside square-brackets after the name of the major mode. A few commands concerning q-registers, which were not in the default enviroment but used by TECMAC, have been deleted. The library purifier now no longer converts uparrow-L to a control-L. You were already told to change the uparrow-L's to control-L's, with preceding ^]^Q's when they were at the beginnings of lines. *** If you have a start-up library instead of just an init file, you can now call that library the same thing you would call the init file. EMACS can tell what kind of file it is. Soon, EMACS will stop looking for .EMACS :EJ. Please wait for me to move this version off of AI.  ECC@MIT-MC 07/01/78 20:48:32 Re: Word Abbrev Mode To: INFO-EMACS at MIT-MC When defining an abbrev for the last N words, ^X ^A (^R Add Mode Word Abbrev) now works slightly differently (better, I think -- a good suggestion by KMP): the abbrev's expansion includes only up to the end of the last word -- and not any break characters following it (but preceding point). This lets you not worry about whether you've typed a space or something after the word you want to abbreviate. For the times when you want to include break characters before or following a string of words, you can give ^X^A an argument of 0, telling it to use the region. (This is as it always was.)  JLK@MIT-MC 06/30/78 11:16:30 To: info-emacs at MIT-AI People that have init files which check FS LISPT"N to see if they should load LISPT, should check FS %OPLSP "N and the XJNAME of the job instead.  Date: 30 JUN 1978 0131-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI If your init file sets any variable whose effect is to redefine some command character, you must run MM & Process Options to make the setting take effect. This was always true, but in the past visiting a file or switching buffers would do it for you. Now, to save time, they call & Process Options only when they think it is necessary, so you must do it yourself when you make it necessary. These are the variables in question; Atom Word Mode Auto Fill Mode Autoarg Mode FS CTLMTA Lisp ) Hack Overwrite Mode It will not work to put any of these variables in a file's local modes unless you also put in "Switch Modes Process Options:1". However, these variables all relate to a user's taste in editing, rather than to the contents of a file, so they belong in your init file rather than in the local modes (imagine that someone else looks at the file; does he want Atom Word Mode set?). Fill Column is an example of something that does belong in the Local Modes; it specifies how the file should look, rather than just how certain commands should be changed. By the way, there has not been any connection between Lisp ) Hack and Lisp Mode for a year and a half.  Date: 29 JUN 1978 1124-EDT From: rich at MIT-AI (Charles Rich) Sent-by: HES at MIT-AI Subject: New E To: INFO-E at MIT-AI If you have an .EMACS :EJ file make sure that the setup macro is called & Setup .EMACS Library The word Library is compulsory in the new emacs, and it was not necessary in the old version. Many people have been losing on this.  Date: 28 JUN 1978 1642-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI I would like to make ^R Execute Completed MM Command available from non-^R subsystems such as RMAIL, INFO, Backtrace, etc. Since Meta-prefixes aren't available, M-X can't be the way to get it. It has to be put on an ASCII character. Does anyone have a suggestion for what character to use? The one that makes sense to use is "X", but it has the problem of meaning "eXit" already. Since "Q" also means exit, it would be possible to reuse "X". Does that sound like a good idea?  Date: 28 JUN 1978 1635-EDT From: RMS at MIT-AI (Richard M. Stallman) To: INFO-EMACS at MIT-AI EMACS 123 is now installed as E.  ECC@MIT-MC 06/26/78 03:01:09 Re: Word Abbrev Mode To: INFO-EMACS at MIT-MC NWORDA and XWORDA versions of Word Abbrev Mode (only runnable in NE) are allow much faster buffer- and mode-switching now. Also, there is a new, special kind of abbrev expander for Tab which should help LISP and TECO users who have had trouble: it won't expand if Tab is not inserting space at point. Did anyone use the $Untouched by Word Abbrev$ or $Word Abbrev Hook$ variables? If so, please contact me.  RMS@MIT-AI 06/25/78 03:20:15 Look in EMACS;INFO OEMACS for the rest of this.