New version of kile out. Reproducible: Always Steps to Reproduce:
Created attachment 42134 [details] Kile 1.7.1 Ebuild This is basically the 1.7beta2 ebuild without the $MY_P stuff.
committed :)
From version 1.7, kile will adopt a system similar to K3b's for i18n strings (i.e.: there is a separate package). I predated K3b's ebuild (which had a mistake that will be reported in a separate bugreport, BTW), and made a version of kile 1.7.1 that compiles on my machine.
Created attachment 42404 [details] Ebuild for Kile 1.7.1, with i18n support
Created attachment 42602 [details] Improved ebuild This ebuild implements a suggestion from bug 68522; the -w 10000 option in the last fmt call ensures that if someone wants all possible languages, the output will not be cut at 75 characters (e.g. university or multinational-company machines may actually want this for their foreign users, so this is not only a theoretical case). Currently the LANGS variable is only 48 characters long, so this is just "good practice" to avoid problems in the future, when more translations will be available.
Thanks for pointing to the i18n tarball, Frederico. Same for the |fmt issue. It wasn't an issue while testing, because I had an extra echo in the src_compile, which "stringified" it, but I removed it later. Regarding "need-kde" in your own ebuilds: Please make sure to call it after R/DEPEND. This fixes not all issues with it, but the most problematic one.
Thanks for the suggestion Carsten, I wasn't aware of the placement issue of need-kde. I confirm spotting the ebuild in portage, I think the bug can be closed if there are no other comments.