| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| policy.diff | Set Debian PolKit policy for PackageKit | Matthias Klumpp <matthias@tenstral.net> | not-needed | 2012-08-14 | ||
| 01_shutdown-daemon-on-idle.patch | [PATCH] Shutdown daemon on idle by default again (#578) This reverts commit dca1f5b2508a4632d0b9fefab771a5a9caf83a5c. Which reverted commit 0c84d71509e851db20445c747529bd7d3724f081, which reverted commit c6eb3555ec5b41e988c111d276764d55fb83bda3. Fixes #460. The memory usage of packagekitd has been observed growing well beyond half a GiB using the DNF backend. See: https://bugzilla.redhat.com/show_bug.cgi?id=1354074 https://bugzilla.redhat.com/show_bug.cgi?id=1854875 https://bugzilla.redhat.com/show_bug.cgi?id=1896964 This is not as severe on backends that are not DNF, but since many backends were originally not designed to be called by long-running processes and the PackageKit daemon serves no purpose while it is doing nothing, reinstating the timeout makes sense to free unused system resources. This is especially useful on low-spec small single-board computers and embedded systems. Originally there was surprising behavior when users mix command line dnf upgades with GUI PackageKit upgrades, and do not manually run an update check before rebooting for update. This triggered the initial default-change to never time out was addressed in a better way via PR #600, allowing us to enable the timeout by default again. |
yump <yump@users.noreply.github.com> | no | 2023-04-05 | ||
| 02_apt-better-repo-names.patch | [PATCH] apt: Display more useful pretty names for repositories Previously they all had names like "Xenial" or "Stable", which is not very useful and causes users to blindly disable repositories that they might need. |
Matthias Klumpp <matthias@tenstral.net> | no | 2023-01-09 | ||
| 03_apt-update-urgency-as-status.patch | [PATCH] apt: Set update urgency as urgency as well as status again Apparently doing anything else confuses frontends that have worked with the old code before. |
Matthias Klumpp <matthias@tenstral.net> | no | 2023-02-10 | ||
| 04_pkcon-fix-progressbar-padding.patch | [PATCH] pkcon: Widen progressbar text area | Frank Dana <ferdnyc@gmail.com> | no | 2023-01-20 | ||
| 05_fix-uninitialized-timer.patch | [PATCH] Fix use of potentially uninitialized timer ID (#599) | Gordon Messmer <gordon.messmer@gmail.com> | no | 2023-01-15 | ||
| 06_apt-fix-mimetype-search-crash.patch | [PATCH] apt: Don't crash while resolving mimetypes, modernize mime search code Previously we warned when there was no AppStream metadata at all by trying to fetch all metadata and warning if we got no result. Given the amount of data Debian and Ubuntu ship with now, this is a bad idea (as it slows down unsuccessful searches massively), and we should do this again only when AsPool has gained a more efficient is_empty() method. Therefore, this check was removed for now. In addition, we stop monitoring the pool for changes, as we will immediately close it anyway, so the additional overhead of setting up background reloading is not needed. Same applies for loading Flatpak data, which we are not concerned with in PackageKit. We also validate the package-name C string now for being NULL before converting it into a std::string, which resolves a crash. |
Matthias Klumpp <matthias@tenstral.net> | no | 2023-03-08 | ||
| 07_fix-tiny-memleak.patch | [PATCH] progress: Free the sender value on finalize Avoid leaking the sender value. |
=?UTF-8?q?Corentin=20No=C3=ABl?= <corentin.noel@collabora.com> | no | 2023-04-24 | ||
| 08_remove-references-to-packagekit.org.patch | [PATCH] Remove reference to defunct packagekit.org URL | =?UTF-8?q?Marcus=20M=C3=BCller?= <marcus@hostalia.de> | no | 2023-05-28 |