textproc/libxml2 I'd originally incorrectly opened bug under /databases/mariadb105-server: bug #251009 Size mismatch (when fetching from gitlab.gnome.org) or not found (when fetching from distcache.FreeBSD.org) on four of the patches causes failure to build. 0e1a49c8907645d2e155f0d89d4d9895ac5112b5.patch 50f06b3efb638efb0abd95dc62dca05ae67882c2.patch 7ffcd44d7e6c46704f8af0321d9314cd26e0e18a.patch edc7b6abb0c125eeb888748c334897f60aab0854.patch Example errors: => Attempting to fetch https://gitlab.gnome.org/GNOME/libxml2/commit/0e1a49c8907645d2e155f0d89d4d9895ac5112b5.patch fetch: https://gitlab.gnome.org/GNOME/libxml2/commit/0e1a49c8907645d2e155f0d89d4d9895ac5112b5.patch: size mismatch: expected 996, actual 994 => Attempting to fetch https://gitlab.gnome.org/GNOME/libxml2/commit/7ffcd44d7e6c46704f8af0321d9314cd26e0e18a.patch fetch: https://gitlab.gnome.org/GNOME/libxml2/commit/7ffcd44d7e6c46704f8af0321d9314cd26e0e18a.patch: size mismatch: expected 1015, actual 1013 I was able to temporarily resolve this by manually fetching the versions here (which have the correct hash) and placing them in /usr/ports/distfiles/gnome2/ https://mirror.xtom.com/freebsd-ports/distfiles/gnome2/
*** Bug 251009 has been marked as a duplicate of this bug. ***
Created attachment 219575 [details] Patch for libxml2 Update distinfo due to upstream change in GitLab instance
The cause is index line(s) being generated differently < index 301c8449..39d92182 100644 --- > index 301c84499..39d92182f 100644
Moin moin I think it's better to include the files in ${FILESDIR} when coming from gitlab. Could you update the patch accordingly? mfg Tobias
Created attachment 219723 [details] Patch for libxml2 v2 Host patches locally
Is the gnome gitlab very different from other gitlab instances (for instance, the freedesktop.org one)? We use this to pull in patches from upstream git repos quite a lot, I'm a bit surprised it doesn't work in this case. I have no objections to local patches, I'm just curious about the issue.
They most likely just updated their instance which changed how patches are generated, at least that's my guess.
Friendly ping :-)
Friendly "woah, what happened to week three and four of november" pong :D Will commit shortly...
A commit references this bug: Author: tcberner Date: Mon Nov 30 18:45:05 UTC 2020 New revision: 556670 URL: https://svnweb.freebsd.org/changeset/ports/556670 Log: textproc/libxml2: ship patches via files/ due to gitlab - gitlab has a tendency to change checksum of patches due to metadata changes - switch to including the patches in the tree directly instead of using PATCH_FILES PR: 251040 Submitted by: daniel.engberg.lists@pyret.net Reported by: David Armstrong <bink19th@pm.me> Changes: head/textproc/libxml2/Makefile head/textproc/libxml2/files/patch-CVE-2019-20388 head/textproc/libxml2/files/patch-CVE-2020-24977 head/textproc/libxml2/files/patch-CVE-2020-7595 head/textproc/libxml2/files/patch-Python-39-support
Committed, and sorry for the wait.
(In reply to Tobias C. Berner from comment #11) Was there a change in this that required the PORTREVISION bump? I'm going to skip building this locally, but I just wanted to confirm I didn't miss anything.
Hi, There's no change that needs a PORTREVISION bump it's just a compilation fix.