Sign me up Login

Details about package keyd

Name: keyd
Uploader: Richard Hansen <rhansen@rhansen.org> (Debian QA page)
Description: keyd - keyboard key remapping daemon
keyd-application-mapper - keyboard key remapping daemon - application-specific remapper

Package uploads

Upload #2

Information

Version: 2.4.3-1
Uploaded: 2024-01-24 06:40
Source package: keyd_2.4.3-1.dsc
Distribution: unstable
Section: utils
Priority: optional
Homepage: https://github.com/rvaiya/keyd
Vcs-Browser: https://salsa.debian.org/rhansen/keyd
Vcs-Git: https://salsa.debian.org/rhansen/keyd.git
Closes bugs: #1060023

Changelog

 keyd (2.4.3-1) unstable; urgency=medium
 .
   * Initial release. (Closes: #1060023)

QA information

Comments

  1. very interested… waiting for наб. wonder if using this keyboard input as on ios/macos would be possible. press a long then be able to enter ä…
    Gürkan Myczko at Feb. 12, 2024, 5:03 a.m.

Upload #1

Information

Version: 2.4.3-1
Uploaded: 2024-01-07 02:40
Source package: keyd_2.4.3-1.dsc
Distribution: unstable
Section: utils
Priority: optional
Homepage: https://github.com/rvaiya/keyd
Vcs-Browser: https://salsa.debian.org/rhansen/keyd
Vcs-Git: https://salsa.debian.org/rhansen/keyd.git
Closes bugs: #1060023

Changelog

 keyd (2.4.3-1) unstable; urgency=medium
 .
   * Initial release. (Closes: #1060023)

QA information

Comments

  1. Also using this on my Chromebook (took me a while to find this because you didn't link this page from your ITP), so comments:
    
    1. debian/patches/0003-Remove-generated-files-from-the-repository.patch is a full megabyte, because it includes the entirety of data/keyd-application-mapper.1.gz, data/keyd.1.gz, data/keyd.compose, which are, well, binary gzips; IMO you should repack this all away by adding
      Files-Excluded: data/*.gz
       data/keyd.compose
    to d/copyright instead of shipping it and the patch.
    
    2. in the build log I see
      for f in docs/*.scdoc; do \
              target=${f%%.scdoc}.1.gz; \
              target=data/${target##*/}; \
              scdoc < "$f" | gzip > "$target"; \
      done
    but https://www.debian.org/doc/debian-policy/ch-docs.html#manual-pages says this should be gzip -9.
    
    3. scripts/generate_xcompose also generates src/unicode.c. repack it away as well since you're already deleting data/keyd.compose
    
    4. scripts/generate_xcompose spends well over a minute spinning with one core 100%ed on my fast system and strace shows nothing but brk() during this time, and the files it makes are distributed by upstream. IMO this looks like scripts/generate_xcompose used by upstream to update the unicode version it supports, and downstream shouldn't. the source is all there, so no DFSG issues – IMO don't regen it
    
    5. you aren't patching away the groupadd from the Makefile!!!
      mkdir -p /home/nabijaczleweli/uwu/keyd.deb/debian/tmp/usr/share/doc/keyd/examples/
      groupadd keyd
      groupadd: Permission denied.
      groupadd: cannot lock /etc/group; try again later.
      make[1]: [Makefile:64: install] Error 10 (ignored)
    r-r-r: no doesn't mean the package /will not/ be built as root, only that it /may/ be (https://www.debian.org/doc/debian-policy/ch-controlfields.html#s-f-rules-requires-root) – 
    
    6. you aren't actually installing the keyd group, so the resulting packages are broken
    Needs work наб at Jan. 10, 2024, 5:29 p.m.
  2. Thank you for the very helpful review наб!
    
    I uploaded a new version of the package that I believe addresses all of your comments.  Responses to your individual points:
    
    1. I decided to keep the generated files as you suggested in point #4.
    2. I removed the compression so that dh_compress can take care of it.  This also avoids overwriting the data/*.1.gz files.
    3. I decided to keep the generated files.
    4. Done.
    5. Done.
    6. Done. I'm not 100% sure I did it properly.  I'm using addgroup in postinst to create the group, and I'm NOT removing the group when the package is removed or purged.
    Ready Richard Hansen at Jan. 24, 2024, 9:44 a.m.