Hi,
On ven., 21 avril 2023 at 10:47, Andreas Enge <andr...@enge.fr> wrote:
>> Therefore, if many things are still missing, I suggest to apply #62967
>> [2]. It removes the annoyance you spotted out in #62954 [3]. And as I
>> explained in [4], it will be safe because ’valgrind’ appears only in the
>> testsuite of ’lz4’ and ’valgrind’ is not even run there.
>> Doing so, it will fix failures for powerpc64le and will increase the
>> coverage.
>
> CI just disappeared, so I cannot check. If r-minimal builds for powerpc64le
> on master, then it is justifiable to fix the regression (leaving open the
> possibility of an immediate revert if the patch introduces problems else-
> where, which I do not expect). Otherwise I would postpone it until after
> the merge.
Yes r-minimal builds for powerpc64le-linux on master:
https://ci.guix.gnu.org/build/1050833/details
--8<---------------cut here---------------start------------->8---
simon@pfiuh07$ guix weather -s powerpc64le-linux r-minimal
computing 1 package derivations for powerpc64le-linux...
looking for 1 store items on https://ci.guix.gnu.org...
https://ci.guix.gnu.org ☀
100.0% substitutes available (1 out of 1)
at least 49,1 MiB of nars (compressed)
52,7 MiB on disk (uncompressed)
2,529 seconds per request (2,5 seconds in total)
0,4 requests per second
at least 1 000 queued builds
aarch64-linux: 819 (81.9%)
x86_64-linux: 117 (11.7%)
powerpc64le-linux: 7 (.7%)
i686-linux: 57 (5.7%)
build rate: 189.97 builds per hour
i686-linux: 138.13 builds per hour
x86_64-linux: 27.34 builds per hour
powerpc64le-linux: 21.66 builds per hour
aarch64-linux: 5.56 builds per hour
looking for 1 store items on https://bordeaux.guix.gnu.org...
https://bordeaux.guix.gnu.org ☀
100.0% substitutes available (1 out of 1)
24,2 MiB of nars (compressed)
52,7 MiB on disk (uncompressed)
0,852 seconds per request (0,9 seconds in total)
1,2 requests per second
'https://bordeaux.guix.gnu.org/api/queue?nr=1000' returned 502 ("Bad Gateway")
--8<---------------cut here---------------end--------------->8---
Note that the patch removes the dependency between subversion and
valgrind. Therefore, as you noted, many TeXlive packages could also
build.
Well, I do not see any cons. :-) The only difference for ’lz4’ is the
path of the store item; here without and with the removal of valgrind:
--8<---------------cut here---------------start------------->8---
$ diff -r --no-dereference
/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3
/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3
diff -r --no-dereference
/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3/lib/pkgconfig/liblz4.pc
/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3/lib/pkgconfig/liblz4.pc
5,7c5,7
< prefix=/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3
< libdir=/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3/lib
< includedir=/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3/include
---
> prefix=/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3
> libdir=/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3/lib
> includedir=/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3/include
13,14c13,14
< Libs: -L/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3/lib -llz4
< Cflags: -I/gnu/store/3sabpv5k19lnifda5ihk7340xm3idxi2-lz4-1.9.3/include
---
> Libs: -L/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3/lib -llz4
> Cflags: -I/gnu/store/1j0jq03930q4c7570145vjiaa8pnajz4-lz4-1.9.3/include
--8<---------------cut here---------------end--------------->8---
Cheers,
simon
