romangg added a comment.
In D24068#534383 <https://phabricator.kde.org/D24068#534383>, @broulik wrote: > > The CHANGELOG keyword is cumbersome to use because the summary line most often already tells sufficiently what the commit is about. So it would only be a repetition of what is already written. Besides why should we want to have a disparity between changelog and commit log anyway? > > I use the CHANGELOG tag quite often. The difference being that the summary is perhaps something like "Avoid null pointer dereference" with CHANGELOG being a more friendly line for the changelog "Fix crash when doing xyz". I went quickly through KScreen, libkscreen, KWin, KWayland, plasma-desktop and plasma-workspace repositories to check if you're an outlier or if it is more often used than I expected. Results: | Repo | Commits w. `CHANGELOG` | Last usage | Kai Uwe share | | KScreen | 2 | May 2016 | 0% | | libkscreen | 2 | Sep 2017 | 0% | | KWin | 2 | Mar 2017 | 100% | | KWayland | 1 | Oct 2016 | 0% | | plasma-desktop | 26 | Jun 2019 | 92% | | plasma-workspace | 41 | Aug 2019 | 88% | | So I admire your expertness on using the available tools to their fullest potential in contrast to everybody else but the numbers also show that the concept hasn't really caught on. I think a more integrated approach in the commit workflow with the prefix can improve the changelogs also quite much. Maybe this will then also encourage people to use the CHANGELOG keyword more often. Since of course with a common prefix it would still be possible to use the keyword. REPOSITORY R104 KScreen REVISION DETAIL https://phabricator.kde.org/D24068 To: romangg, #plasma Cc: broulik, ngraham, plasma-devel, LeGast00n, The-Feren-OS-Dev, jraleigh, fbampaloukas, GB_2, ragreen, ZrenBot, alexeymin, himcesjf, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, mart
