This one time, at band camp, Claudio Moratti said: > Yes, k9copy (like the new qVamps) are frontend to vamps... > I have had to work on k9copy package, because, from the last version, some > things are changes (now it provides a library...)! > > I put k9copy package on http://www.knio.it/debian/k9copy (and after the > holidays qvamps)... > > I'd like some feedback about k9copy package, because this is my first > welldone > (I hope) package providing a library...
Some feedback: debian/rules: You conditionally copy in config.{sub,guess} but unconditionally rm them. This is not idempotent and will lead to failures to build a second time. debian/NEWS: cron (3.0pl1-74) unstable; urgency=low ... ?? This looks like a cut and paste error. debian/libk9copy0.files: usr/lib/libk9copy.la should go in the -dev package debian/dirs: You install /usr/lib/k9copy/ in k9copy, but the libraries are installed into /usr/lib/. Probably the best thing is to just drop that directory from the dirs file. It's only a single library, so you don't really need a subdirectory. If you really want to install into a subdirectory, then you should use k9copy.dirs and libk9copy0.dirs and so forth so that you don't have /usr/lib in the k9copy package, and then use libk9copy.install to install the files in the subdirectory. debian/menu: I'm not sure what to say here. There seems to be very little coherence in Debian proper about where things are dumped in the menu hierarchy, so I am not going to say your choice is wrong - it certainly makes a kind of sense. However, I note that at least k3b and xcdroast put their menu entry in Apps/Tools. It seems like it would make sense to have k9copy nearby those, although I think you could argue they are in the wrong place. This may not be worth fixing right now, as there is an evolving thread on another list about reforming the menu policy, allowing for finer grained submenus to make some snse of the whole mess. I am not saying you have to change anything, just mentioning it. General: Lots of compiler warnings. Not a problem in and of itself, but I always worry about underlying coding practices when I see lots of 'comparison between signed and unsigned' and 'takes type int but arg 4 is long int' type warnings. For the future, you may want to do a code audit, but I don't see it as a sticking point for now. Hope that's helpful, -- ----------------------------------------------------------------- | ,''`. Stephen Gran | | : :' : [EMAIL PROTECTED] | | `. `' Debian user, admin, and developer | | `- http://www.debian.org | -----------------------------------------------------------------
signature.asc
Description: Digital signature