[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1002295: Uploaded the files



Control: tags -1 moreinfo

Hi Scorpion,

I wonder if you actually checking things, with the hints given by Bastian.

- the watch file is still broken [1]
  - upstream seems not to do releases, where did you get the 0.2.0 version
    from? Maybe it would be a good idea to ask upstream to do releases and tag them.
  - if upstream does not do releases, your version scheme needs to be
    considering that, eg. by the scheme proposed by uscan(1) -- see below [1].
- https://mentors.debian.net/package/multi-timer/ suggests to me that you did
  NOT run lintian. Many sponsors will not upload with lintian warnings. [2]
- For initial uploads, d/changelog only says "Initial Upload (Closes #$YOURITP),
  there should no other entries.
- Consider sending your patches upstream. We should avoid carrying patches in Debian
  hard, so there need to be a good reason for the patch and deviation of upstream.
  I don't see that reasons, or they are not  documented in the patch header, which
  is not following the dep3 format.
- "mt" is proably not a good choice for the script name, as this is too short and pollutes
  the namespace. Please consider to rename it. (best, take that to upstream.)
  Suggeston: multi-time, like the manpage. I guess this would be more appropiate.

- d/mt.pdf creates issues:
  - pdf as documentation are not the best choice.
  - in this case, it seems to be a print out of firefox, of upstream's homepage.
    This is an issue, as the pdf would need to be created at build time.
    - which you cannot download on build time
    - and is 404ing.

  I suggest to drop the pdf -- its not that useful anyway and the manpage
  should contain all the bits, not refering to the pdf. (Yes, you need to fix
  the manpage -- dont forget to send it upstream afterwards.)

- d/alarm_001.wav
  Why do you divert from upstream on that? As said, patches needs a good
  reason…

- d/README
  - probably should be named README.md?
  - should sent upstream to be included in the repository.
  - this notes are because you are shipping it in the debian directory:
    - it has install instructions for Debian. Those are not needed when
      people install a package and should be covered by d/control already.
    - users are not supposed to change binaries shipped by a package. So
      the note
        "The default command is "loginctl lock-session", changed it if
         you want/need to use something else for locking the screen like
          i3lock etc.  It is at line 548."
      is not approiate and needs to be handled by code, not by asking people to edit.
      - IIUIC, the section "Interface to Sysmonitor Indicator Option"
        tells people to create a FrankenDebian [3]. This is not appropiate.
        Your option is to package https://github.com/fossfreedom/indicator-sysmonitor
        (preferred) or just not mentioning it.
        (Telling people to follow bad practices is not the option you have.)

- d/control
  - package description needs some work to convey what is the package
    about better...
  - Why does it need a versioned dependency on bash?
    - bash is essential, and the version >4.6 is already fullfilled in
      old-stable.
    ( - I guess it is not needed)

- lots of trailing whitespaces in d/* .. you might find wrap-and-sort(1)
  userful.

- The orig.tar should NOT have the .git repository directory.
  A proper d/watch and using uscan to create the tarball will take care of that ;-)

Disclaimer: I do not intend to sponsor this package.

There are issues with the packaging; please also note that you will need
to familarize with the Debian BTS [4], this is a required skill for a
package maintainer. I'm tagging this bug "moreinfo" again, the usage of
this tag in the scope of sponsorship-requests is documented in [5].
When you have fixed _all_ things, remove the tag by interacting with the
BTS... This way you show us that you learned how to use the BTS.

[1] uscan --verbose --no-download
uscan info: uscan (version 2.22.2) See uscan(1) for help
uscan info: Scan watch files in .
uscan info: Check debian/watch and debian/changelog in .
uscan info: package="multi-timer" version="0.2.0-1" (as seen in debian/changelog)
uscan info: package="multi-timer" version="0.2.0" (no epoch/revision)
uscan info: ./debian/changelog sets package="multi-timer" version="0.2.0"
uscan info: Process watch file at: debian/watch
    package = multi-timer
    version = 0.2.0
    pkg_dir = .
uscan info: opts: mode=git
uscan info: line: https://github.com/pippim/multi-timer.git HEAD
uscan info: Parsing mode=git
uscan info: line: https://github.com/pippim/multi-timer.git HEAD
uscan info: Last orig.tar.* tarball version (from debian/changelog): 0.2.0
uscan info: Last orig.tar.* tarball version (dversionmangled): 0.2.0
uscan info: Looking at $base = https://github.com/pippim/multi-timer.git with
    $filepattern = HEAD found
    $newfile     = HEAD
    $newversion  = 0.0~git20220219.d86a3b6
    $lastversion = 0.2.0
uscan info: Upstream URL(+tag) to download is identified as    https://github.com/pippim/multi-timer.git HEAD
uscan info: Filename (filenamemangled) for downloaded file: multi-timer-0.0~git20220219.d86a3b6.tar.xz
uscan info: Newest version of multi-timer on remote site is 0.0~git20220219.d86a3b6, local version is 0.2.0
uscan info:  => Only older package available from:
             => https://github.com/pippim/multi-timer.git HEAD
uscan info: Removing git repo (../multi-timer-temporary.377801.git)
uscan info: Scan finished

[2] https://nthykier.wordpress.com/2012/02/23/some-sponsors-are-evil-and-pedantic/
[3] https://wiki.debian.org/DontBreakDebian#Don.27t_make_a_FrankenDebian
[4] https://www.debian.org/Bugs/
[5] https://wiki.debian.org/Mentors/BTS#Tags


Reply to: