roeder@sbsvax.UUCP (Edgar Roeder) (07/18/89)
Today i have submitted my port of GNU Emacs for Atari ST computers to the atari binaries group. The port has been done with gcc V1.30 . The only missing features are those related to multitasking/multiuser environments like asynchronous processes. 'call-process' is implemented through system()-calls. With the help of bert@sbsvax.uucp (Norbert Reithinger) we have also an implementation of dired written with emacs functions. The executable program runs on 1MB of memory but more is recommended. - Edgar
mark@ccvr1.uucp (Mark Boyd) (07/25/89)
I got the final parts of the gnuemacs distribution yesterday and got it un-zooed ( zoo //x in c:/usr/lib/gnuemacs ) with no problems. Thanks for the real emacs! I love it. When I ran it I discovered one minor problem: Some of the files in the lisp subdirectory have underlines in the file names, e.g. text_mod.elc, but the program looks for dashes, not underlines, e.g text-mod.elc. The solution is simple, just change all the underlines to dashes in the file names. There are only four or five that need changing, but one of them is the site-ini file, so you really do need to make the change! The only other problem is that my 1 megabyte machine isn't big enough. I just ordered 2 Megabytes of memory chips this morning. They are dowm to about $100 / megabyte, so there is no economic reason not to have enough memory for FSF software on an ST! Mark Boyd Computer Science, UNC asheville, Asheville, NC mark@ccvr1.ncsu.edu mbunca@ecsvax.uncecs.edu
andyc@hplsla.HP.COM (Andy Cassino) (07/28/89)
I've also gotten the gnuemacs port running on my Mega-2. It's VERY
impressive. Watching REAL gnuemacs come up on my Atari was a REAL thrill.
It is just like using gnuemacs at work! GREAT JOB!
I do have a couple of problems to work out. The first is, I can't get
"M-x info" to work. After succesfully finding a file or two, I get the
following message:
"Info file /usr/lib/GNUemacs/info/dir does not exist."
Nowhere in the emacs.zoo or elisp.zoo archives did I find a file named "dir",
so is something missing?
Also, I am running emacs from gulam, and "M-x suspend-emacs" or C-z just
cause emacs to bomb off (2 bombs). I end up in gulam again and everything
seems ok, but when I type exit, I go back to the desktop, not back into
emacs. Any hints? (Sure would be nice if this worked... gnuemacs takes awhile
to get under way, seems that executing loadup.el is the biggest bottleneck)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Andy Cassino %
% uucp: hplabs!hplsla!andyc domain: andyc%hplsla@hplabs.hp.com %
% Hewlett-Packard Lake Stevens Instrument Division %
% 8600 Soper Hill Road Everett, WA 98205-1298 %
% (206) 335-2211 %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%saj@chinet.chi.il.us (Stephen Jacobs) (07/28/89)
I know, we were told not to expect much from GNU EMACS with only 1 Meg, but
after determining that it only used about 300 K text+data+bss I gave it a
shot. Just start it, read some stuff in and push a few keys, nothing serious.
Impressive hardly describes it. Looks just like on the VAX, only it starts
up faster. But I have a couple of ???-s.
The start-up seems to crash rather than finishing normally (something about
invalid argument stringp null, as I remember). Control Z gives a kind of
no-op (the bottom line flashes 'program terminated' ever so briefly, then
I'm back in EMACS. And dired seems to be getting confused, probably by the
$HOME environment variable in \ form (rather than / form) which is really
there for the use of the shell, msh, rather than EMACS.
But the program runs fast, and there's a tremendous amount there. This is
going to be fun to explore.
I hope the diffs from a widely available source distribution show up on
the net.
Steve J.roeder@sbsvax.UUCP (Edgar Roeder) (07/31/89)
In article <5440043@hplsla.HP.COM>, andyc@hplsla.HP.COM (Andy Cassino) writes: > I do have a couple of problems to work out. The first is, I can't get > "M-x info" to work. After succesfully finding a file or two, I get the > following message: > > "Info file /usr/lib/GNUemacs/info/dir does not exist." > > Nowhere in the emacs.zoo or elisp.zoo archives did I find a file named "dir", > so is something missing? > You are right: there is something missing. Since i did not modify any of the documentation files in /usr/lib/GNUemacs/info, i did not include them in the distribution of GNU-Emacs in comp.binaries.atari.st. You can get any documentation on Emacs, GCC and some other GNU programs from a host running GNU emacs. You would have to look for the info directory and copy those files. I have about 2.5 MB of info files on my st now (including gcc docs, gulam doc). If you can't find the info files look for .texinfo files. They are the source for info (texinfo-format-buffer converts them). > Also, I am running emacs from gulam, and "M-x suspend-emacs" or C-z just > cause emacs to bomb off (2 bombs). I end up in gulam again and everything > seems ok, but when I type exit, I go back to the desktop, not back into > emacs. Any hints? (Sure would be nice if this worked... gnuemacs takes awhile > to get under way, seems that executing loadup.el is the biggest bottleneck) > I have tried this feature with my shell (it is Master 5.6) and everything worked fine. I am calling the shell via _shell_p (at $4F6) with an empty string as parameter for gulam if it's magic number can be found. Before this the string "Program stopped!!!" is written to stderr (handle 2). The call is done in user mode of the 68000 and the available stack at this point is about 8 kB. If meta-flag is non-nil the conterm variable is configured to return the Shift-key status when doing tos-i/o. Emacs does no unusual things beside that. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!! I would like to hear about experiences of users that have other shells !!! !!! which use the _shell_p vector as entry to the commandline-interpreter. !!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Those shells are called with "-i" as single argument to the routine pointed to by _shell_p. Have you tried '(call-process "ls" nil t nil)' from the scratch-buffer (or after ESC-ESC) ??? > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > % Andy Cassino % > % uucp: hplabs!hplsla!andyc domain: andyc%hplsla@hplabs.hp.com % > % Hewlett-Packard Lake Stevens Instrument Division % > % 8600 Soper Hill Road Everett, WA 98205-1298 % > % (206) 335-2211 % > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - Edgar
andyc@hplsla.HP.COM (Andy Cassino) (08/04/89)
roeder@sbsvax.UUCP (Edgar Roeder) writes, with regard to suspending gnuemacs: | |I have tried this feature with my shell (it is Master 5.6) and everything |worked fine. I am calling the shell via _shell_p (at $4F6) with an empty string |as parameter for gulam if it's magic number can be found. Before this the string |"Program stopped!!!" is written to stderr (handle 2). The call is done in user |mode of the 68000 and the available stack at this point is about 8 kB. If |meta-flag is non-nil the conterm variable is configured to return the Shift-key |status when doing tos-i/o. Emacs does no unusual things beside that. | |!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! |!!! I would like to hear about experiences of users that have other shells !!! |!!! which use the _shell_p vector as entry to the commandline-interpreter. !!! |!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! | |Those shells are called with "-i" as single argument to the routine pointed |to by _shell_p. | |Have you tried '(call-process "ls" nil t nil)' from the scratch-buffer |(or after ESC-ESC) ??? | Well, the details of the implementation are above my head, I haven't been programming the ST very long! Is there something specific I need to set in the shell environment to get this to play? I do get the "Program stopped!!!" message, BTW. I tried C-z from Mark Williams MSH and got a different result. I got "Program stopped!!!" in the mini-buffer, but emacs did not stop. No bombs in MSH, though! Hmm, I need to try call-process. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Andy Cassino % % uucp: hplabs!hplsla!andyc domain: andyc%hplsla@hplabs.hp.com % % Hewlett-Packard Lake Stevens Instrument Division % % 8600 Soper Hill Road Everett, WA 98205-1298 % % (206) 335-2211 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%