|
|
|
.emacsrc File
(Also see the original MicroEMACS 3.9e documents in the
help
directory;
emacs-doc.txt
is the full reference manual.)
Last Update: Thu Nov 23 21:38:04 2023
This was MicroEMACS 3.9, written by Dave G. Conroy and substantially modified by Daniel M. Lawrence (with contributions from others). The last entry in the change log is dated November 4, 1987. The software was frozen for the version 3.9e USENET release according to the last log entry.
In September 1987, I moved from 5 years of heavy VAX/VMS experience to a small prototype project using an early Sun M68K Unix workstation. One person used the workstation, two others used their PCs connected via serial lines, and I naturally gravitated to a DEC VT220 terminal (also connected via a serial line).
I initially configured vi(1) to work like DEC's EDT editor (Wikipedia), mapping the VT220 keypad escape sequences to commands that would take VI out of insert mode if necessary and perform the commands equivalent to EDT's keypad functions. This made VI workable for me, but it still was just not the same as my heavily customized EDT (and then TPU in the last couple of VMS years).
Note: I originally included the following paragraph here,
but I just happened upon a comment dated January 4, 1988
about a change I made in estruct.h, so I must have
downloaded MicroEMACS before getting a short-lived copy
of UniPress EMACS. My memory is playing tricks on me!
"As a subcontractor, I somehow convinced our contractor manager to buy a copy of UniPress EMACS. However, he was only willing to buy the binary for $300 and wouldn't splurge on the source code for $600 total. I went to town with EMACS Lisp and soon had my beloved editing environment once again. And then we got the first Sun SPARC workstation — Goodbye, M68K EMACS!"
Somehow I heard of MicroEMACS and downloaded the source code from Quantum Link's CP/M archive for the Commodore 128 or from GEnie (General Electric Network for Information Exchange).
Quantum Link eventually became AOL. Speaking of the old pre-internet days, should I call EMACS by its original moniker, EMACS, or should I follow the not-so-new fad of calling it Emacs? UNIX Unix? FORTRAN Fortran? SNOBOL Spitbol? You tell me!
I figure it cost me over $40 to obtain MicroEMACS. Both Q-Link and GEnie were run on corporate computers and hourly rates went up after 7:00 or 8:00 in the morning. My 1200 baud download took longer than expected and I got socked with that $40 charge!
Anyway, with some tweaking, I built MicroEMACS on our SPARCstation and I was back in business. As a result I've been using a heavily customized EDT editor since 1982, beginning with the EDIT/EDT editor itself and, since 1987, continuing with MicroEMACS!
Customized how? I have included my EDT initialization file,
edtini.edt,
and my TPU file,
eveini.tpu.
The files have
December 1993 dates because, in 1992-1993, I worked on a
project for which we ported our SunOS software to VAX/VMS
and I got to use EDT and TPU again.
When I moved from EDT to TPU (in 1985?), I found DEC's TPU version of EDT too limited, so I instead used TPU's EVE editor with my own EDT-inspired customizations.
When recently checking my MicroEMACS version of EDT against
the DEC EDT reference manual (downloaded from the internet),
I was struck by how incredibly awkward EDT's keypad method of
substitution using "GOLD ENTER" was. Early on, I must have
implemented my "GOLD s" method (possibly learned from another
more experienced developer).
HTML/PDF versions of the EDT, TPU, and EVE reference manuals can be found at https://www.digiater.nl/openvms/doc/alpha-v8.3/ (EDT's is PDF only).
If you're reading this and using GNU Emacs, you might be
interested in GNU's
"Emacs
EDT emulation" and/or
tpu-edt.el.
These seem to postdate my need for MicroEMACS (and would not have
been easily accessible at the time in any case).
I added the following new macro commands:
alx-center-line
edt-delete-char
edt-undelete-char
edt-delete-bol
edt-delete-eol
edt-delete-line
edt-undelete-line
edt-delete-word
edt-undelete-word
edt-next-word
I added the following function:
&DRS wildcardFileSpec indexReturns the I-th (1..N) file name that matches the wildcard file specification. Wildcards can only appear in the file name and not in the preceding directory names, if any. A blank string, "", is returned if no file names are matched or if the index is out of range.
The result is stored locally in alx_drs() (found in
alex.c), so it is a good idea to not use
&DRS more than once in an expression. The result
is stored as a dynamically allocated string, so there is no length
limit as seen with the built-in ampersand string functions (see
"Warnings About the Macro Language" below).
(I originally modified the various file input functions
in MicroEMACS's file.c to perform the wildcard matching
internally. Circa 2001, I added the same functionality
to JASSPA MicroEMACS for Windows by implementing the
&DRS function. In June 2023, I figured I might
as well use the same technique in this MicroEMACS.)
I modified the following function:
&CHR numberReturns the character whose code is the specified number.
This modification allows numbers to be specified in decimal,
hexadecimal ("0x..."), and octal ("0...").
I added the following macro environment variable:
$curoffByte offset (0..N-1) in the line of the current cursor position.
Note that this is different than "$curcol", which
takes into account tab settings and returns the cursor column as
rendered on a virtual screen. For example, if the cursor is positioned
on "a" in a line beginning with
"^Iabc...", $curcol
of "a" is 9 (its column on the screen) and
$curoff is 1 (the offset in the line of "a"
following the tab character). [READ ONLY]
Additional software changes included bug fixes, expansion of predefined size limits, use of the termios(3) subsystem instead of the obsolete termio(7) subsystem, clean-up of some of the code, ...
My efforts debugging a somewhat complex "!if" expression revealed
two implementation details of the MicroEMACS ampersand string functions that
the user should be aware of:
The ampersand string functions that manipulate and return
strings store their results in the same, statically allocated
buffer, result, in eval.c's
gtfun() function. If more than one function using
the result buffer appears in an expression, then the buffer
will be unwittingly overwritten by intermediate values.
The functions in question are:
&CAT &CHR >K &IND &LEFT &MID &RIGHT
To avoid conflicts, break the expression into smaller pieces
and store intermediate values in user variables
("%name"). MicroEMACS stores the value of
user variables in dynamically allocated strings. (The storage
will be freed and allocated anew the next time the expression
is evaluated.)
The result buffer in gtfun() is of a fixed size, 2048
(2*NSTRING, where NSTRING is defined
in estruct.h). Consequently, the results of the
ampersand string functions are limited to 2047 characters.
Exceeding this limit is entirely possible; for example, when
manipulating a long line of text. (I have modified the functions to
truncate results and not write past the end of the buffer.)
I added the ability to generate debug output to a file. First, the following
global variables were added to edef.h:
int dbgIsOn ; - 0 = debug off, ~0 = debug on
FILE *dbgFile ; - Debug file, opened in mesetup() in "main.c"
Second, the following command-line option was added to main.c:
-dfile - Enable debug and open "file" for debug output
Debug output can then be generated wherever desired with:
if (dbgIsOn) fprintf (dbgFile, ...) ;
My MicroEMACS configuration file,
.emacsrc,
currently depends on DEC VT1xx/VT2xx application mode keypad escape
sequences. Ways to get these escape sequences from an IBM PC/AT keypad include:
Under Windows, I use Tera Term to access (i) Cygwin on the host machine and (ii) Linux/BSD on a remote machine. The Tera Term keyboard setup file I use to map the PC keypad presses to VT escape sequences is:
VT-ibm-keypad-teraterm-AlexEdt-CNF.txt
Note that keycodes in the setup file are not Windows
keycodes, but the values displayed by the keycode.exe
program included in the Tera Term distribution. (See
Tera
Term Keyboard Setup for more information.)
The connection to Cygwin on the host machine is accomplished using
Tera Term's cygterm+ program in conjunction with Tera
Term. See the "Command line" and "Setup files" entries in Tera Term's
Help Index.
I used to use an AutoHotkey script with Cygwin's mintty(1) terminal program under Windows 7 (and possibly under Windows XP before that?), but at some point it stopped working. I have included the AutoHotkey script in case someone else prefers to use it and can figure out the problem:
Under X Windows in Linux or BSD, I use xterm(1) and this xmodmap(1) keypad map file:
VT-ibm-keypad-xterm-xmodmaprc.txt
(The Xterm FAQ has more information.)
With either of the two methods above, the VT keypad PF2 key is dropped and the PC keypad is mapped to the VT keypad as follows:
Num Lock -> PF1 (the GOLD key!)
/ -> PF3
* -> PF4
- -> Keypad "-"
+ -> Keypad ","
Enter -> Keypad Enter
. -> Keypad "."
0 .. 9 -> Keypad digits
(PF2 in the EDT editor was the "help" key, which I never used.)
Visually, the VT keypad keys are laid out on the PC keypad like this:
.=====.=====.=====.=====.
| PF1 | PF3 | PF4 | - |
+=====+=====+=====+=====+
| 7 | 8 | 9 | |
+=====+=====+=====+ |
| 4 | 5 | 6 | , |
+=====+=====+=====+=====+
| 1 | 2 | 3 | |
+=====+=====+=====+ |
| 0 | . | Ent |
`===========^=====^=====+
In years long past, I did use other non-PC/AT keyboards on PCs and
Unix computers and I successfully managed to mimic VT-style keypad
use, even when there was no keypad! (You can see remnants of the
handling of these keyboards in my
.emacsrc
file; I should probably clean that up someday.)
README.txt
.emacsrc - My MicroEMACS configuration file, which depends on ...
- ... (See keypad section above)
VT-ibm-keypad-mintty-autohotkey.txt
VT-ibm-keypad-teraterm-AlexEdt-CNF.txt
VT-ibm-keypad-xterm-xmodmaprc.txt
VT-edtini-edt.txt - My EDT initialization file ("edtini.edt").
VT-eveini-tpu.txt - My TPU initialization file ("eveini.tpu").
Makefile.bsd - Using on FreeBSD 13.2 Makefile.cygwin - Using on Cygwin64 (Windows 10) Makefile.linux - Using on Linux Mint 20
alex.c - Original, c. 1988 EDT functions. eprotos.h - ANSI C prototypes for all MicroEMACS C functions. Directory: orig - Original source for modified MicroEMACS files. Directory: ots - on-the-shelf MicroEMACS code; i.e., not used.
These files are largely used to support my &drs function.
Circa 1990, this functionality was implemented using only a
few files, but this changed over the years as I reworked the
code to be more generalized and more robust. Not to mention
the fact that I ported MicroEMACS, again over the years, to
UNIX (SunOS, HP/UX, Solaris, etc.), Linux, BSD, and Cygwin
platforms. The LIBGPL code itself is portable to various
systems including VAX/VMS and Windows Visual Studio 2022.
pragmatics.h - The hand-crafted equivalent of an autoconfig file. iel_util.h - Simple debug logging and error reporting. aperror.c aperror.h drs_util.c - Directory scanning package, which requires the drs_util.h source files below, including the REX parser get_util.c down through the UTF package. get_util.h meo_util.c meo_util.h nnl_util.c nnl_util.h rex_util_y.bison - Regular expression parser and lexer for reference. rex_util_y.c - Bisonized C code for the parser. rex_internals.c rex_internals.h rex_tokens.h rex_util.c - Actual matching/sustitution code. rex_util.h sdx_util.c - Handles dynamically growing strings in sustitution. sdx_util.h sto_util.c - Length-limited S-to-N functions for the parser. sto_util.h sto_util_template.h str_util.c - String manipulation functions. str_util.h utf_util.c - UTF-8, -16, and -32 conversion functoins. utf_util.h
These files were part of the original MicroEMACS 3.9e distribution.
Directory: cmd - command stuff. Directory: help - help information. basic.c bind.c buffer.c crypt.c display.c ebind.h edef.h efunc.h epath.h estruct.h eval.c evar.h exec.c file.c fileio.c input.c isearch.c line.c main.c random.c region.c search.c spawn.c tcap.c termio.c window.c word.c
.emacsrc
File
%edt-advance-forward - Direction of EDT "next" movements:
forward (TRUE) or reverse (FALSE).
%edt-mark-is-active - Is EDT's mark active? (This is independent
of the always active MicroEMACS mark.)
%edt-paste-was-appended - Set to FALSE by EDT cut; set to TRUE by
subsequent EDT replace(s). Affects deletion
of previous characters in append and paste.
%edt-search-mode - Cursor position after successful, forward
EDT FIND/FNDNXT commands. The default for
MicroEMACS is to place the cursor after the
last character in the match. By setting
this variable to "BEGIN", the cursor will
be backed up to the beginning of the match.
(The cursor will be advanced one character
before searching for the next match.)
%edt-name-buffer - Names of various buffers.
%my-name[-number]-buffer
...
EDT-PASTE - EDT cut text is copied here for subsequent EDT pasting.
EDT-DELETE - Replaced text is copied here during EDT replace.
ALTERNATE - Temporary buffers for whatever use by typist ...
ALTERNATE/4
ALTERNATE/5
ALTERNATE/6
ALTERNATE/7
ALTERNATE/8
ALTERNATE/9
For completeness, in 2023, I added EDT functions RESET
(GOLD DOT), APPEND (KP9),
REPLACE (GOLD KP9), CTRL/U (delete to BOL),
DEL EOL (GOLD KP2), LINE (KP0),
and OPEN LINE (GOLD KP0). Since I have no
memory of using these functions before, I implemented
and tested them based solely on the descriptions in the
EDT Reference Manual. I suggest you check them out to be
sure they behave as you expect. I have not implemented
SUBST (GOLD ENTER). (Yet, maybe?)
Naming convention:
GOLD - PC NumLock, VT PF1 key
PF3 - PC keypad "/"
PF4 - PC keypad "*"
MINUS - PC keypad "-"
COMMA - PC keypad "+"
ENTER - PC keypad "Enter"
DOT - PC keypad "."
KP0..KP9 - PC keypad digits
Actions (ordered by top row, right column, digits 7-9, 4-6, 1-3, and 0).
Actions tagged with "[DIRECTION]" work forwards or backwards
depending on the %advance-forward setting, which is set/reset
with KP4/KP5:
GOLD PF3 - EDT prompt and search (FIND)
PF3 - EDT search next (FNDNXT)
PF4 - EDT delete line (DEL L)
GOLD PF4 - EDT undelete line (UND L)
MINUS - EDT delete word (DEL W)
GOLD MINUS - EDT undelete word (UND W)
COMMA - EDT delete character (DEL C)
GOLD COMMA - EDT undelete character (UND C)
ENTER - (Unimplemented: EDT ENTER) Use keyboard's regular
"Enter"/"Return" key to finish EDT FNDNXT and
COMMAND (GOLD KP7) prompts.
GOLD ENTER - (Unimplemented: EDT SUBST) I use "GOLD s",
described below, to invoke MicroEMACS's own
query-replace-string command.
DOT - EDT begin/cancel selection of text (SELECT)
(Selected text is NOT highlighted.)
GOLD DOT - EDT reset selection and direction (RESET)
KP7 - EDT next page (next form feed) (PAGE) [DIRECTION]
GOLD KP7 - Prompt for and execute MicroEMACS named command.
KP8 - EDT next section (scroll) (SECT) [DIRECTION]
GOLD KP8 - EDT fill paragraph (FILL)
KP9 - EDT append selection to paste buffer (APPEND)
GOLD KP9 - EDT replace selection with paste buffer (REPLACE)
KP4 - EDT set forward direction (ADVANCE)
GOLD KP4 - EDT move to bottom of file (BOTTOM)
KP5 - EDT set backward direction (BACKUP)
GOLD KP5 - EDT move to top of file (TOP)
KP6 - EDT cut selected text to PASTE buffer (CUT)
GOLD KP6 - EDT paste text from PASTE buffer (PASTE)
KP1 - EDT next word (but always move forward) (WORD)
My variant. Explanation by example:
Given "while ((i > 0) && ...", the next word
after "while" is the first left parenthesis,
not "i". And the next word after that is "i",
not the second left parenthesis.
GOLD KP1 - (Unimplemented: EDT change case in region)
KP2 - EDT next line (but always move forward) (EOL)
GOLD KP2 - EDT delete to end-of-line (DEL EOL)
KP3 - EDT next character (but always move forward) (CHAR)
GOLD KP3 - EDT insert special character (SPECINS)
KP0 - EDT next beginning of line (LINE) [DIRECTION]
GOLD KP0 - EDT open line (OPEN LINE)
GOLD k - Begin keyboard macro
^K - End keyboard macro
^E - Execute keyboard macro
^A - Insert line of asterisks
^I - Insert 4 spaces (NOT a tab)
GOLD t - Insert genuine tab character
^L - EDT insert form feed (CTRL/L)
^N - Insert number line (useful for positioning text)
^R - EDT refresh editor screen (CTRL/R)
^U - Use "GOLD u" for EDT CTRL/U (delete from beginning-of-line)
GOLD m - Switch to main buffer
GOLD p - Switch to EDT CUT's "PASTE" buffer
GOLD D - Switch to EDT REPLACE's "DELETE" buffer.
GOLD a - Switch to "ALTERNATE" buffer
GOLD 4 - Switch to "ALTERNATE/4" buffer
GOLD 5 - Switch to "ALTERNATE/5" buffer
GOLD 6 - Switch to "ALTERNATE/6" buffer
GOLD 7 - Switch to "ALTERNATE/7" buffer
GOLD 8 - Switch to "ALTERNATE/8" buffer
GOLD 9 - Switch to "ALTERNATE/9" buffer
GOLD 1 - Return to one window in editor
GOLD 2 - Split current window vertically
GOLD ` - Switch cursor to next window
GOLD " - Convert smart quotes, etc. to ASCII characters
GOLD M - Convert CR/LF to LF (CR=^M)
GOLD b - Bottom of buffer
GOLD B - Prompt for and switch to another buffer
GOLD c - Prompt for and insert control (A-Z) character
GOLD C - Center current line
GOLD d - Delete (not cut) current selection
GOLD D - Generate directory listing
GOLD e - Save file and exit MicroEMACS
GOLD q - Quit MicroEMACS without saving
GOLD f - Go to matching fence (e.g., braces, brackets, etc.)
GOLD F - Prompt for and set fill paragraph width
GOLD g - Prompt for and go to line in file
GOLD G - Get (open) another file for editing
GOLD i - Prompt for and insert contents of a file
GOLD I - Toggle insert mode
GOLD j - Join paragraph (multi-line to single-line)
GOLD l - Display current line number
GOLD L - Prompt for and execute MicroEMACS command file
GOLD r - Prompt/interactive substitution (with regular expression)
GOLD s - Prompt/interactive substitution (without regular expression)
GOLD T - Trim trailing blanks from all lines in buffer
GOLD u - EDT delete from beginning-of-line (CTRL/U)
GOLD U - Expand tabs to 4 spaces
GOLD w - Prompt for and write current buffer to file
GOLD x - Prompt for and edit one or more shell commands.
When the commands are ready, press "GOLD x" again
to execute the shell commands and then switch to
the shell output buffer.
GOLD X - Creates and drops you into a shell subprocess.
This is NOT in an editor window. Enter shell
command "exit" to return back to the editor.
GOLD z - Insert 3-byte UTF-8 BOM
GOLD Z - Remove 3-byte UTF-8 BOM