control: tags -1 -moreinfo >>>>> "Adam" == Adam D Barratt <a...@adam-barratt.org.uk> writes:
Adam> Control: tags -1 + moreinfo >> freeradius (2.2.5+dfsg-0.2) unstable; urgency=high >> >> * Disable OpenSSL version check; Debian will maintain ABI >> stability or change the soname, Closes: #765871 * Non-Maintainer >> Upload Adam> The package also appears to remove the "git-ids" patch from Adam> the series file; was that intentional? Yeah. So, the reason the git-ids patch is there is because upstream uses .gitattributes to enable $Id$ in sources. git dpm doesn't really deal well with this situation, and git-buildpackage always generates a set of sources that dpkg-source can't deal with. The difference are differences in the blob hashes recorded in the checked out files. I manually included a patch to update the blob hashes so that the sources would actually build. After the 2.5+dfsg-0.1 upload I realized that I could just as easily reverse-apply this patch, reduce upstream changes, and remove this build artifact. So I did that with this upload. The difference is that we ship with the upstream blob hashes for modified files. The effect is that for modified config files, it looks like the sysadmin modified them rather than the build process. I think that's semi-more-correct if you're tracking which upstream version things came from, but I can make reasonable arguments for either direction. I think having a patch that does nothing but update blob hashes in files was kind of silly. sorry for not calling this out. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org