* Mattia Rizzolo (mattia@debian.org) wrote: > On Thu, Aug 31, 2017 at 10:06:52AM +0100, Chris Lamb wrote: > > > lintian: check for not wrap-and-sort formatted files > > > > Good idea. > > I don't think it is. > wrap-and-sort (or the equivalent from cme) is not widely enough adopted > as of yet. I personally try to have all my sponsee use it, and I > sometimes force a wrap-and-sort into NMUs, but it's still too rare. > > Also, wouldn't you need to pick only one of the -t, -s, -a combinations? > I see people disliking -t (trailing commas) also because it's not in > policy, and arguments for and agaist -s and -a; see the wrap-and-sort > manual if you don't understand what I'm talking about. > And that's without considering cme's configuration. You wouldn't necessarily have to proscribe one single wrap-and-sort variant, you could check a number of variations and see if any of them fit the current set of files. But yes, if policy could be more prescriptive about what the formatting should look like that would be cool. > > Alas, I fear that this would either require calling out to > > wrap-and-sort (!) and diffing the result, or essentially reimplementing > > it within Lintian itself? > > You could use libconfig-model-dpkg-perl, most probably. Therefore > elevating cme's implementation. > > Summing up: I don't think it's a practise ready to be nudged by lintian > yet. Not even at the pedantic level? -- Eric Dorland <eric@kuroneko.ca> 43CF 1228 F726 FD5B 474C E962 C256 FBD5 0022 1E93
Attachment:
signature.asc
Description: PGP signature