Bug 245571
| Summary: | editors/emacs does not give a tty to sub-shells | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Bertrand Petit <bsdpr> | ||||
| Component: | Individual Port(s) | Assignee: | emacs (Nobody) <emacs> | ||||
| Status: | Closed Overcome By Events | ||||||
| Severity: | Affects Only Me | CC: | ashish, bsdpr, jrm | ||||
| Priority: | --- | Flags: | bugzilla:
maintainer-feedback?
(emacs) |
||||
| Version: | Latest | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
|
Description
Bertrand Petit
2020-04-12 15:04:45 UTC
(In reply to Bertrand Petit from comment #0) I'm not able to reproduce it. This is what I did: 1. M-x shell RET 2. hostname!abbe:~ % lsof -n -p $$ lsof -n -p $$ COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME zsh 44814 abbe cwd VDIR 4148061892,1756814850 181 4 /usr/home/abbe zsh 44814 abbe rtd VDIR 2317175353,554032710 25 4 / zsh 44814 abbe 0u VCHR 0,116 0t6 116 /dev/pts/7 zsh 44814 abbe 1u VCHR 0,116 0t6 116 /dev/pts/7 zsh 44814 abbe 2u VCHR 0,116 0t6 116 /dev/pts/7 zsh 44814 abbe 10u VCHR 0,116 0t520 116 /dev/pts/7 zsh 44814 abbe 12r VREG 1313969477,1458822731 182144 959 /usr/local/share/zsh/5.6.2/functions/Completion.zwc zsh 44814 abbe 13r VREG 1313969477,1458822731 194784 2055 /usr/local/share/zsh/5.6.2/functions/Misc.zwc zsh 44814 abbe 14r VREG 1313969477,1458822731 269528 1014 /usr/local/share/zsh/5.6.2/functions/Completion/Base.zwc Could you describe in exact steps how to reproduce your problem ? Thanks! I use an emacs configured as shown bellow:
cd /usr/ports/editors/emacs ; make showconfig
===> The following configuration options are available for emacs-26.3_3,3:
ACL=off: Access control list support
DBUS=off: D-Bus IPC system support
FCITXPATCH=off: Patch to pass C-Space to Fcitx input method
GCONF=off: GConf configuration backend support
GNUTLS=on: SSL/TLS support via GnuTLS
GSETTINGS=off: GSettings API for application settings
LCMS2=off: Little CMS 2.x support
M17N=off: M17N support for text-shaping [implies XFT]
MAILUTILS=off: Mail framework via GNU Mailutils
MODULES=on: Dynamic modules support
OTF=on: Opentype fonts [implies XFT]
SCROLLBARS=off: Toolkit scroll-bars
SOURCES=on: Install sources
THREADS=on: Threading support
XFT=on: Xft font library support
XIM=off: X Input Method
XML=on: XML format or parser support
XWIDGETS=off: Experimental GTK widgets [implies GTK3]
====> Graphics support
CAIRO=off: Experimental Cairo library support [implies PNG, XFT]
GIF=on: GIF image format support
JPEG=on: JPEG image format support
MAGICK=off: ImageMagick image processing library support
PNG=on: PNG image format support
SVG=on: Scalable vector graphics support [implies PNG]
TIFF=on: TIFF image format support
XPM=on: XPM pixmap image format support
====> File notification support: you can only select none or one of them
GFILE=off: File notification via gfile
KQUEUE=on: File notification via kqueue
====> Sound (audio) support: you can only select none or one of them
ALSA=off: ALSA audio architecture support
OSS=on: Open Sound System support
====> X11 toolkit: you can only select none or one of them
GTK2=off: GTK+ 2 GUI toolkit support
GTK3=off: GTK+ 3 GUI toolkit support
MOTIF=off: Motif widget library support
XAW=on: Athena widgets
XAW3D=off: Athena3D widgets
===> Use 'make config' to modify these settings
Then for testing I run it as:
emacs -q -f shell
In the shell buffer, I get this:
bash: cannot set terminal process group (-1): Inappropriate ioctl for device
bash: no job control in this shell
[/tmp]$ lsof -n -p $$
lsof: WARNING: device cache mismatch: /dev/random
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
bash 9443 elrond cwd VDIR 1999562137,1302847080 88 3 /tmp
bash 9443 elrond rtd VDIR 1953227420,2260328126 40 4 /
bash 9443 elrond 0u VCHR 0,176 0t194 176 /dev/pts/3
bash 9443 elrond 1u VCHR 0,176 0t194 176 /dev/pts/3
bash 9443 elrond 2u VCHR 0,176 0t194 176 /dev/pts/3
bash 9443 elrond 255u VCHR 0,176 0t194 176 /dev/pts/3
lsof: WARNING: /home/elrond/.lsof_memo3 was updated.
[/tmp]$ stty
speed 9600 baud;
lflags: -echo echoe echoke echoctl
oflags: -onlcr tab0
cflags: cs8 -parenb
erase kill
<undef> <undef>
(emacs-version)
"GNU Emacs 26.3 (build 1, amd64-portbld-freebsd12.1, X toolkit)"
(In reply to Bertrand Petit from comment #2) I'm not sure where that message from bash comes from. I tried it in a pristine bash shell (no ~/.bash* files), and am not able to reproduce it: [abbe@host ~]$ yes >/dev/null C-c C-c [abbe@host ~]$ yes >/dev/null & [1] 68485 [abbe@host ~]$ [abbe@host ~]$ [abbe@host ~]$ yes >/dev/null& [2] 68486 [abbe@host ~]$ [abbe@host ~]$ [abbe@host ~]$kill %1 [abbe@host ~]$ kill %2 [1]- Terminated yes > /dev/null [abbe@host ~]$ [2]+ Terminated yes > /dev/null [abbe@host ~]$ yes | grep no & [1] 68504 [abbe@host ~]$ [abbe@host ~]$ [abbe@host ~]$ kill %1 [abbe@host ~]$ [1]+ Terminated yes | grep no [abbe@host ~]$ -UUU:**--F1 *shell* Bot L174767 (Shell:run) ---------- I can perform job control operations as you can see. Could you elaborate exactly the problem you are facing, like not able to kill child processes, or something ? Thanks! (In reply to Ashish SHUKLA from comment #3) The problem is not related to bash or any other shell, /bin/sh is affected as well. To illustrate this further here is the output of a (attached) test command, first when executed by a shell in shell-mode, then directly run by comint, and eventually from an xterm == using /bin/sh in shell-mode == $ ./ttytest /dev/pts/0 is a tty tcsetpgrp fail, Inappropriate ioctl for device == using comint == M-x comint-run ttytest /dev/pts/8 is a tty setpgid fail, Operation not permitted Process ttytest exited abnormally with code 1 == using a /bin/sh running under an xterm == $ ./ttytest /dev/pts/1 is a tty no error I currently have no more ideas to explore this issue further. Created attachment 213629 [details]
test that attempts to become the foreground process
The same emacs version running on an older 12.1-STABLE version does not exhibit the incorrect behavior. The bug could lie in the kernel or any library (base or ports) on which emacs depends. Affected systems run svn rev 359586 (amd64). The unaffected system run svn rev 354128 (armv7). Hello Bertrand, Thanks for your report. I also cannot reproduce on 12.1 and I do not have a 12-STABLE system available to test on. If you have the time and motivation you could help us out by testing the following. - We are preparing for the upcoming 27.1 release. Could you test our WIP update? https://reviews.freebsd.org/D23966 - Could you try editors/emacs-devel? - Could you bisect commits from the STABLE branch to try and identify which change caused the problem for you? Regards, Joseph (In reply to Joseph Mingrone from comment #7) I did install emacs-27.0.91,3 after applying the diff you provided against the ports tree. I used the same 12-1 STABLE revision as before. Aside from incompatibilities in my .emacs, that version fully restored job control in shell mode when it is running either /bin/sh or /usr/local/bin/bash. Likewise, from the same patched ports tree, I installed emacs-devel-28.0.50.20200416,2. This version also provide a properly configured tty to shell mode sub-processes. In both cases no other port was upgraded since the original report. Since I may be the only one affected it is surely not worth examining this issue further. Thanks. We'll close this then. Of course, please re-open if necessary. |
