Re: [PATCH] Identify LWLocks in tracepoints

Поиск
Список
Период
Сортировка
От Andres Freund
Тема Re: [PATCH] Identify LWLocks in tracepoints
Дата
Msg-id 20210510160822.ce4f4xvzga6pwnm4@alap3.anarazel.de
обсуждение исходный текст
Ответ на Re: [PATCH] Identify LWLocks in tracepoints  (Tom Lane <tgl@sss.pgh.pa.us>)
Ответы Re: [PATCH] Identify LWLocks in tracepoints  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-hackers
On 2021-05-09 19:51:13 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2021-05-08 13:13:47 -0400, Tom Lane wrote:
> >> (I wonder why flaviventris and serinus are still using an "experimental"
> >> compiler version that is now behind mainstream.)
> 
> > The upgrade script didn't install the newer version it because it had to
> > remove some conflicting packages... Should be fixed for runs starting
> > now.
> 
> Looks like that didn't work ...

Looks like it did, but turned out to have some unintended side-effects
:(.

The snapshot builds are now new:
https://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=flaviventris&dt=2021-05-10%2015%3A43%3A56&stg=configure
configure:3966: ccache /usr/lib/gcc-snapshot/bin/gcc --version >&5
gcc (Debian 20210421-1) 11.0.1 20210421 (prerelease) [gcc-11 revision
fbb7739892e:d13ce34bd01:3756d99dab6a268d0d8a17583980a86f23f0595a]

But the aforementioned dependencies that needed to remove broke the
installed old versions of gcc/clang.

I started to build the old versions of llvm manually, but that then hits
the issue that at least 3.9 doesn't build with halfway modern versions
of gcc/clang. So I gotta do it stepwise (i.e. go backwards, build llvm
n-2 with n-1), will take a bit of time.

Greetings,

Andres Freund



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: Why do we have perl and sed versions of Gen_dummy_probes?
Следующее
От: Tom Lane
Дата:
Сообщение: Re: [PATCH] Identify LWLocks in tracepoints