lynn@rave.phx.mcd.mot.com (Lynn D. Newton) (05/17/89)
I just upgraded to GNU Emacs 18.54 last week and am experiencing an annoying problem, which I wonder if anyone else has any insights on. Every few minutes my status line stops displaying the time (and mail notification). The loadst program disappears from the process table. I have (display-time) near the end of my .emacs file, and it starts fine when I come up. After it quits I can do eval-expression (display-time) and it restarts. It appears in the process table. Then it goes away again. I never had any problem with this before. I can't absolutely guarantee that I didn't do something inadvertent to change something related in the course of making the upgrade. I did a lot of work on my .emacs (248 lines -- too long to quote). Any ideas anyone? -- ================================================================= Lynn D. Newton | System Test Motorola MCD, Tempe, AZ | (Department of Heuristic Methodology) (602) 437-3739 | "The bug stops here!" lynn@rave.phx.mcd.mot.com |
wjc@ho5cad.ATT.COM (Bill Carpenter) (05/18/89)
In article <LYNN.89May16164517@rave.rave.phx.mcd.mot.com> lynn@rave.phx.mcd.mot.com (Lynn D. Newton) writes: > I just upgraded to GNU Emacs 18.54 last week and am experiencing > an annoying problem, which I wonder if anyone else has any > [loadst goes away, seems random] I would guess you have a SysV-ish UNIX (#define USG). See if your loadst is going away when you hit control-G. In GNUemacs, that's a SIGINT, and it's killing your subprocesses. This part of the 18.54 upgrade seems to be the culprit. Change it so you *do* get the setpgrp() for USG, and all will be well. (At least I have seen no problem after changing it back. I must assume it was changed in the first place to try to cure something, but I don't know what.) -- Bill Carpenter att!ho5cad!wjc or attmail!bill DON'T APPLY THIS PATCH! THIS IS THE bad-for-USG BIT FROM THE 18.54 UPGRADE. THIS IS NOT THE FIX FOR IT (do that manually). diff -rc2N dist-18.53/src/callproc.c dist-18.54/src/callproc.c *** dist-18.53/src/callproc.c Wed Feb 15 22:36:20 1989 --- dist-18.54/src/callproc.c Thu Apr 6 12:21:14 1989 *************** *** 374,383 **** close (err); ! #ifdef USG ! setpgrp (); /* No arguments but equivalent in this case */ ! #else setpgrp (pid, pid); - #endif /* USG */ setpgrp_of_tty (pid); #ifdef vipc --- 374,381 ---- close (err); ! #ifndef USG setpgrp (pid, pid); setpgrp_of_tty (pid); + #endif /* USG */ #ifdef vipc
msb@ho5cad.att.com (Mike Balenger) (05/18/89)
In article <WJC.89May17223304@ho5cad.ho5cad.ATT.COM> wjc@ho5cad.ATT.COM (Bill Carpenter) writes: > In article <LYNN.89May16164517@rave.rave.phx.mcd.mot.com> lynn@rave.phx.mcd.mot.com (Lynn D. Newton) writes: > > I just upgraded to GNU Emacs 18.54 last week and am experiencing > > an annoying problem, which I wonder if anyone else has any > > [loadst goes away, seems random] > I would guess you have a SysV-ish UNIX (#define USG). See if your > loadst is going away when you hit control-G. In GNUemacs, that's a > SIGINT, and it's killing your subprocesses. Others may have also noticed that their server, ispell, or pop-mail stuff has gone away. Same problem; same solution. -- ---------------------------------------------------------------------- <cute quote> Michael S. Balenger (201) 949-8789 <cute disclaimer> AT&T Bell Labs Room 1L-405 msb@ho5cad.att.com Crawfords Corner Road att!ho5cad!msb Holmdel, NJ 07733
jr@bbn.com (John Robinson) (05/18/89)
In article <WJC.89May17223304@ho5cad.ho5cad.ATT.COM>, wjc@ho5cad (Bill Carpenter) writes: >In article <LYNN.89May16164517@rave.rave.phx.mcd.mot.com> lynn@rave.phx.mcd.mot.com (Lynn D. Newton) writes: >> I just upgraded to GNU Emacs 18.54 last week and am experiencing >> an annoying problem, which I wonder if anyone else has any >> [loadst goes away, seems random] > >I would guess you have a SysV-ish UNIX (#define USG). See if your >loadst is going away when you hit control-G. In GNUemacs, that's a >SIGINT, and it's killing your subprocesses. This part of the 18.54 >upgrade seems to be the culprit. That's the most likely explanation. I also used to see this on a Sun (bsd). The problem was an older version of mh-inc-folder-hook which got a new mode-line string by signalling loadst. I guess if the signal from emacs crossed the timer interrupt in loadst, it could get caught with its handler down. The version of mh-inc-folder-hook which justs does surgery on the mode-line string avoids this problem. Another solution I thought of but never tried was to keep the pipe to loadst open, and just push a newline through anytime emacs wanted a mode-line update. loadst would simply wait on a gets in its loop. But surgery on the string within emacs is easier and faster. -- /jr, nee John Robinson What a waste it is to lose one's mind--or not jr@bbn.com or bbn!jr to have a mind. How true that is. -Dan Quayle
tmb@wheaties.ai.mit.edu (Thomas M. Breuel) (05/19/89)
It's not just loadst that dies; the TeX-shell also goes away when you type "^G". This is 18.54 on a System V/386 Release 3.2 system. Anybody have a fix?