Control: severity -1 wishlist Hi!
On Mon, 2014-05-26 at 20:25:17 +0200, Andreas Metzler wrote: > Package: dpkg-dev > Version: 1.17.9 > Severity: normal > ------------------- > (SID)ametzler@argenau:/tmp/TASN/dpkg-bug/testing/libtasn1-3.6$ find -name .pc > (SID)ametzler@argenau:/tmp/TASN/dpkg-bug/testing/libtasn1-3.6$ cat > debian/patches/series > 20_off-by-one-error.patch > (SID)ametzler@argenau:/tmp/TASN/dpkg-bug/testing/libtasn1-3.6$ patch -p1 > --dry-run < debian/patches/20_off-by-one-error.patch > checking file lib/decoding.c > Reversed (or previously applied) patch detected! Assume -R? [n] > Apply anyway? [n] > Skipping patch. > 1 out of 1 hunk ignored > (SID)ametzler@argenau:/tmp/TASN/dpkg-bug/testing/libtasn1-3.6$ dpkg-source > --before-build `pwd` ; echo $? > 0 > ------------------- > > I would have expected dpkg-source to exit with an error instead. Well, this is documented behavior, please see the dpkg-source man page, section “Format: 3.0 (quilt)”, option --before-build. This seems to be just fallout from the UI changes to make life easier for people not using or having quilt installed and similar. I agree though, that this is quite suboptimal behavior, and I guess I'll be changing it to check all applied patches, and failing whenever one of the patches will fail to apply, which might impose a slight performance degradation, particularly in packages with either tons of patches, or huge patches. But I don't think there's any other sane way to improve the heuristic here, and handle legitimate cases like yours. Thanks, Guillem -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
