Hi Felix, Il 23/12/19 10:58, Felix Lechner ha scritto: > Hi s3v, > > This message only deals with your supplemental question about geany-plugins. > > On Sun, Dec 22, 2019 at 3:21 AM s3v <c0llap...@yahoo.it> wrote: >> >> duplicate-short-description was not emitted for geany-plugins despite there >> are >> two binary packages with the same short description. > > That tag is emitted locally: > > $ dget > https://deb.debian.org/debian/pool/main/g/geany-plugins/geany-plugins_1.36+dfsg-1.dsc > $ frontend/lintian geany-plugins_1.36+dfsg-1.dsc > I: geany-plugins source: debian-watch-uses-insecure-uri > http://plugins.geany.org/geany-plugins/geany-plugins-(.*).tar.gz > I: geany-plugins source: duplicate-short-description > geany-plugin-git-changebar geany-plugin-keyrecord > I: geany-plugins source: out-of-date-standards-version 4.1.2 (released > 2017-11-30) (current is 4.4.1) > I: geany-plugins source: public-upstream-key-not-minimal > upstream/signing-key.asc has 125 extra signature(s) for keyid > B507ACD04BA283C9 > I: geany-plugins source: public-upstream-key-not-minimal > upstream/signing-key.asc has 16 extra signature(s) for keyid > FBD5225B588752A1 > I: geany-plugins source: testsuite-autopkgtest-missing > I: geany-plugins source: unused-file-paragraph-in-dep5-copyright > paragraph at line 80 > I: geany-plugins source: unused-override > file-contains-fixme-placeholder debian/control:124 FIXME > I: geany-plugins source: wildcard-matches-nothing-in-dep5-copyright > geanylatex/* (paragraph at line 75) > I: geany-plugins source: wildcard-matches-nothing-in-dep5-copyright > geanylatex/src/geanylatex.c (paragraph at line 80) > P: geany-plugins source: insecure-copyright-format-uri > http://www.debian.org/doc/packaging-manuals/copyright-format/1.0/ > P: geany-plugins source: package-uses-old-debhelper-compat-version 9 > P: geany-plugins source: rules-requires-root-missing > >> I don't see geany-plugins in >> https://lintian.debian.org/tags/duplicate-short-description.html > > The web page states at the top that not all occurrences were shown. > Perhaps that message could be clearer: > > * only 1024 instances are listed on this page.
Ouch! I read the message box on the top of the web page but I thought the alphabetically ordered list was truncated to the first (but complete) 1024 items. My bad :| Thank you for explanation and for your time. > >> Could I open new bug report? > > In general, it is probably better to open a separate report. The bug > number issue is not something you or I can do anything about; perhaps > they can encode it in base64 to make it shorter. Also, bugs regularly > are no bugs at all, so that is not offensive. For complex secondary > issues, we will duplicate and retitle your report. Either way is fine. I will do :) >> duplicate-short-description tag has been emitted for a lot of Haskell >> packages >> but short descriptions seem to be different. > > Lintian checks d/control in the source package. All of the entries look like: > > Description: ${haskell:ShortDescription}${haskell:ShortBlurb} > ${haskell:LongDescription} > . > ${haskell:Blurb} > >> I admit I've not checked all Haskell packages but all short descriptions in >> control file are in the form >> >> Description: ${haskell:ShortDescription}${haskell:ShortBlurb} >> >> maybe ${haskell:ShortBlurb} is not properly handled? > > You are right. The variable is not handled at all. The current Lintian > check does not work for Haskell packages. > > Checking the source packages, as we currently do, has the advantage > that no built packages need be present. Lintian can run separately on > *.dsc, *.deb and *.changes, and we prefer to run on source packages > whenever possible. That does not work when variables are used, and > should perhaps be changed for all of Lintian. > > A hybrid option would be to check the installation packages second, if > present, and suppress any tags not actually warranted. > > This is a great bug and will require some thinking. > I think most immediate solution could be substituting Haskell variables by Lintian itself before emitting tags. ${haskell:ShortDescription} and ${haskell:LongDescription} are already in d/control, ${haskell:ShortBlurb} and ${haskell:Blurb} are always the same. A most general solution would be preferable, I know. But, from my experience of translator in DDTP, Haskell packages are only ones using variables in their descriptions. I'm able to open new bug report if other variables will be encountered or Haskell packages change their variable values. duplicate-short-description is a "wishlist" tag and I think Lintian's Team efforts may be focused elsewhere if handling variables feature is too hard to be implemented. Please note that Debian Policy have not suggestions for variables in package descriptions, so those variables would be in any form or place. My 2 cents :) Kind regards and thanks again for your great work on Debian packages.