Обсуждение: cfbot mistakenly reports that a rebase is needed
Hi, I keep an eye on my patches using cfbog [1] and I think since recently it started to mistakenly mark patches as "needs rebase" while in fact everything is OK. If we take [2] for instance, it applies fine: ``` $ git am ~/Desktop/v4-0001-pgindent-improve-formatting-of-multiline-comments.patch Applying: pgindent: improve formatting of multiline comments ``` There is a log of `git am` on cfbot [3] but I don't see any errors there: ``` === Applying patches on top of PostgreSQL commit ID b227b0bb4e032e19b3679bedac820eba3ac0d1cf === /etc/rc.d/jail: WARNING: Per-jail configuration via jail_* variables is obsolete. Please consider migrating to /etc/jail.conf. Tue Aug 12 02:21:19 UTC 2025 On branch cf/5831 nothing to commit, working tree clean === using 'git am' to apply patch ./v4-0001-pgindent-improve-formatting-of-multiline-comments.patch === Applying: pgindent: improve formatting of multiline comments ``` I've seen the same problem with another patch of mine [4]. It also applied fine but I submitted a rebased version nevertheless. Strangely enough it made cfbot happy, so I doubt it has anything to do with a WARNING from FreeBSD Jails. [1]: https://cfbot.cputube.org/ [2]: https://commitfest.postgresql.org/patch/5831/ [3]: https://cfbot.cputube.org/patch_5831.log [4]: https://postgr.es/m/CAJ7c6TNbq%3D%2Bf8YKVE6Mrwf_omfi_BARdsj2tOniS4nRVLyEBKw%40mail.gmail.com
Aleksander Alekseev <aleksander@tigerdata.com> writes:
> I keep an eye on my patches using cfbog [1] and I think since recently
> it started to mistakenly mark patches as "needs rebase" while in fact
> everything is OK. If we take [2] for instance, it applies fine:
Yeah, I have observed the same with my patches (according to [1],
two of 'em need rebase right now, though the underlying 'git am' logs
report success). In some past occurrences the gripe has gone away
without my doing anything...
regards, tom lane
[1] http://cfbot.cputube.org/tom-lane.html
On Wed, 13 Aug 2025 at 16:39, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > Aleksander Alekseev <aleksander@tigerdata.com> writes: > > I keep an eye on my patches using cfbog [1] and I think since recently > > it started to mistakenly mark patches as "needs rebase" while in fact > > everything is OK. If we take [2] for instance, it applies fine: > > Yeah, I have observed the same with my patches (according to [1], > two of 'em need rebase right now, though the underlying 'git am' logs > report success). In some past occurrences the gripe has gone away > without my doing anything... > > regards, tom lane > > [1] http://cfbot.cputube.org/tom-lane.html Took a quick look, and publicly noting my findings: Interestingly none of these patches show up as "needs rebase" in the commitfest app. So it seems that this is a cfbot webpage generation bug.
On 2025-Aug-13, Aleksander Alekseev wrote: > Hi, > > I keep an eye on my patches using cfbog [1] and I think since recently > it started to mistakenly mark patches as "needs rebase" while in fact > everything is OK. Isn't the cfbot.cputube.org web page obsolete now that we have the CI status directly in the commitfest page? -- Álvaro Herrera Breisgau, Deutschland — https://www.EnterpriseDB.com/
=?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@kurilemu.de> writes:
> Isn't the cfbot.cputube.org web page obsolete now that we have the CI
> status directly in the commitfest page?
Well, yeah, it's kind of duplicative --- but it slices the data
differently and it presents it differently. The more modern
pages don't have an easily-bookmarked summary of the state of
your own patches. (The "dashboard" page is cluttered with a ton
of other stuff, or at least it is for me.) So I still use the
cputube.org pages often, and I guess Aleksander does too.
regards, tom lane
Hi, > > Isn't the cfbot.cputube.org web page obsolete now that we have the CI > > status directly in the commitfest page? > > Well, yeah, it's kind of duplicative --- but it slices the data > differently and it presents it differently. The more modern > pages don't have an easily-bookmarked summary of the state of > your own patches. (The "dashboard" page is cluttered with a ton > of other stuff, or at least it is for me.) So I still use the > cputube.org pages often, and I guess Aleksander does too. cfbot is more convenient in my typical use cases. Let's say I want to check if any of my patches needs a rebase. I just open: https://cfbot.cputube.org/aleksander-alekseev.html ... and instantly get the answer. I can achieve the same with CF app, however the URL is going to be different every month, e.g: https://commitfest.postgresql.org/55/?text=&status=-1&targetversion=-1&author=-3&reviewer=-1&sortkey= Occasionally CI may fail not because something is wrong with my patch. In such cases the "CI status" column shows me a red icon and something like "9/10". A few extra clicks are needed for every patch to understand what happened. When using cfbot I see right away that something went wrong on Windows for almost all the patches, as an example. I could leave without cfbot but this would make my workflow less convenient.
On Thu, 14 Aug 2025 at 14:31, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > =?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@kurilemu.de> writes: > > Isn't the cfbot.cputube.org web page obsolete now that we have the CI > > status directly in the commitfest page? > > Well, yeah, it's kind of duplicative --- but it slices the data > differently and it presents it differently. The more modern > pages don't have an easily-bookmarked summary of the state of > your own patches. (The "dashboard" page is cluttered with a ton > of other stuff, or at least it is for me.) So I still use the > cputube.org pages often, and I guess Aleksander does too. FYI an equivalent page would be: https://commitfest.postgresql.org/me/?&author=-3
Hi Jelte, > FYI an equivalent page would be: > https://commitfest.postgresql.org/me/?&author=-3 That's handy. Thanks for the tip!
Aleksander Alekseev <aleksander@tigerdata.com> writes:
> Occasionally CI may fail not because something is wrong with my patch.
> In such cases the "CI status" column shows me a red icon and something
> like "9/10". A few extra clicks are needed for every patch to
> understand what happened. When using cfbot I see right away that
> something went wrong on Windows for almost all the patches, as an
> example.
Yeah, I agree with this: the "9/10" presentation is a bit too
compact, as it doesn't allow seeing patterns in the failures.
regards, tom lane