On Wed, 2014-05-14 at 22:41 -0700, Russ Allbery wrote: > None of those transforms other than removing comments lose information so > far as I can tell. Here is another reason for this tag, some CSS is generated from other more expressive forms like LESS and SASS. Apparently there is no way to tell those types of generated CSS from minified CSS. So we should check for minified CSS so we also catch these cases. http://lesscss.org/ http://sass-lang.com/ > As a package maintainer, I would find this tag extremely irritating, and > would be tempted to simply automate adding overrides of it without further > consideration. To me, it feels like you are taking the desire for source > to an extreme that is going to annoy people using Lintian. I'm always > worried that this would lead to people no longer using Lintian at all > because it's annoying them with unimportant trivia. I would have thought the tag wouldn't be triggered by your packages nor triggered very often in the archive. I'm afraid it is way too late to be concerned about people ignoring lintian, people have been annoyed at lintian and adding inappropriate overrides for years. Personally I don't think lintian goes far enough and never will, which is why I'm starting to work on a wrapper for checking all of the things using as many checker tools as possible. https://bugs.debian.org/576184 https://wiki.debian.org/HowToPackageForDebian#Check_points_for_any_package https://anonscm.debian.org/gitweb/?p=collab-maint/check-all-the-things-old.git https://bitbucket.org/jwilk/maquack https://anonscm.debian.org/gitweb/?p=collab-maint/check-all-the-things.git > I also don't think the definition of preferred form of modification that > you're using here is supported by the rest of the project. I'm using "the form that upstream actually uses for modification", which as far as I can tell is what Debian has always used (cf #383465). > If you think I'm wrong, I think it would be good to raise this in > debian-devel and see if you can actually get project consensus about this. Just now this topic (CSS compression) was raised on the #debian-devel IRC channel and the consensus among the few people in the discussion was that the non-minified form should be provided too. Of course that audience is far smaller than on the debian-devel list. > In the long run, we probably need a GR on this whole general topic of > source, since it's clear that different people have *massively* different > understandings of what Debian should be doing in this area. I don't think there will ever be consensus on this topic and a GR won't help either. There will always be people who miss, ignore or are not well informed enough to detect specific missing source. Anyway, feel free to close this bug if you desire, I will write a check for suspicious-source or check-all-the-things if so. -- bye, pabs http://wiki.debian.org/PaulWise
Attachment:
signature.asc
Description: This is a digitally signed message part