[Python-apps-team] Bug#636396: mercurial: Cloning fails with: mpatch.mpatchError: patch cannot be decoded

Javi Merino cibervicho at gmail.com
Wed Sep 14 11:41:26 UTC 2011


2011/9/14 Arno Huggenfeld <ahgf80 at hotmail.com>:
>
> Thanks for the new version. Hg verify is working but I can confirm that your test fails.

Good!

> I started the tests suite with setting 0:
>
> jupiter:/# cat /proc/cpu/alignment
> User:           1042610
> System:         2
> Skipped:        0
> Half:           1808
> Word:           1040674
> DWord:          2
> Multi:          6
> User faults:    0 (ignored)
>
> @jupiter:~/mercurial-test/mercurial-1.9.2$ LANG=C make local
> ...
>
> @jupiter:~/mercurial-test/mercurial-1.9.2$ LANG=C make tests > tests.log
> ...
> I will send the results when finished but it isn't looking good.
> Or did you want me to run the test with # echo 2 > /proc/cpu/alignment ?

I wanted both versions (if possible). If the problems only happen with
alignment=0 then it's definitely an alignment problem, and we can tell
people to set alignment to 2 in order to use mercurial for the time
being.

If it's definitely an alignment problem, we can add that to the bug
upstream[0] and see what mercurial devs have to say about it.

[0] http://mercurial.selenic.com/bts/issue2877

Thanks!
Javi (Vicho)





More information about the Python-apps-team mailing list