>>>>> 1. It allows to add new translations or update existing ones >>>>> without having to change/revalidate the (fat) qasmixer package. >>>>> >>>>> 2. The l10n package can be used for all architectures. >>>>> This should save some precious bytes on the package servers. >>>> >>>> It is possible to use this structure: >>>> package qasmixer, depends on: >>>> package qasmizer-data >>>> The -data package may then contain all stuff that works on all >>>> architectures, including all the localized strings in the .desktop file. >>> >>> Of course, you'll get the same lintian warning. >> >> Yes, right, the warning would remain. :/ >> >> Here's the more details current upstream scenario. >> >> qasmixer_X.Y.tar.bz2 <- Two number version >> qasmixer-l10n_X.Y.Z.tar.bz2 <- Three number version >> >> X and Y always match, but Z grows with every new l10n submit. >> >> What I can do is to create either >> >> 1. two packages, qasmixer_X.Y and qasmixer-l10n_X.Y.Z >> - Has the lintian warning issue >> >> 2. one package qasmixer_X.Y.Z with both tarballs merged >> - Most package data would be duplicated with every l10n submit >> - Binary would be recompiled on every l10n submit > > I suppose I was being a bit glib. If you have a corner case where the > lintian warning can be safely ignored and it is not something that > should be fix in lintian, I think you should just override the lintian > warning.
I see linitian supports warning overrides. http://lintian.debian.org/manual/section-2.4.html Then I'd like to stick with 1 which reflects the upstream situation better. Thanks for the feedback so far! -- Sebastian -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4eb11bcc.5080...@gmx.de