This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [BZ#13982] [PATCH 2/3] Split ar_SD into ar_SD and ar_SS
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Chris Leonard <cjlhomeaddress at gmail dot com>
- Cc: libc-alpha <libc-alpha at sourceware dot org>
- Date: Tue, 8 Oct 2013 00:11:50 +0000
- Subject: Re: [BZ#13982] [PATCH 2/3] Split ar_SD into ar_SD and ar_SS
- Authentication-results: sourceware.org; auth=none
- References: <CAHdAatbRWAFtKB61N53gfWnYaBMJ7kqV3X0Xr+FB_6krD5TCwQ at mail dot gmail dot com> <Pine dot LNX dot 4 dot 64 dot 1310071636580 dot 31470 at digraph dot polyomino dot org dot uk> <CAHdAatY8PFd_JJJNbprMOXei7HtgungtMCC-S55qR9A+9t+8ZQ at mail dot gmail dot com>
On Mon, 7 Oct 2013, Chris Leonard wrote:
> I believe you are overstating the extent of problems caused by "my
> recent commits". There was one particularly series of commits that I
> broke into three to update the iso-3166 data (and the related locale
> splits). Admittedly there were several typos that have been caught
> and corrected.
And such typos should be caught by testing before the patch is submitted
or committed. This is not the first time recently a patch you've
committed has caused problems that would have been caught by standard
testing. <https://sourceware.org/ml/libc-alpha/2013-09/msg00146.html> and
<https://sourceware.org/ml/libc-alpha/2013-09/msg00725.html> are previous
examples where I set out to test a localedef patch and found in the
process that noise had been introduced in the baseline state of "make
localedata/install-locales", making it harder than necessary to verify the
localedef patch wasn't responsible for any problems.
Of course, if your commits draw attention to themselves by causing
regressions, that also draws attention to ChangeLog omissions or git log
entries that don't accurately reflect the patch contents, as those make it
harder than necessary to identify the patch responsible for the
regression, even if they might not otherwise matter much.
> Typically my commits do not extend beyond the localedata/locales
> directory, only rarely straying farther afield. I do test locales as
> I work on them and before I commit them. I do not have a build system
> set up, so I do not know if I can perform the requested "make" test or
> not.
Making a mistake once is fine. But after one of your patches first caused
such a problem, that should have been enough warning that you needed to
set up a build environment and test subsequent commits.
> I am contributing to glibc in order to address the chronic neglect of
> localedata tickets, I am proud of the work I've done and find your
> unwelcoming tone to be a rather curious way to treat a newcomer that
> is only trying to address an area that has been neglected for so long.
You're a committer, with all the responsibilities that implies, so not a
newcomer.
--
Joseph S. Myers
joseph@codesourcery.com