This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Add git mirror for binutils?
- From: Jim Meyering <jim at meyering dot net>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: Jie Zhang <jie dot zhang at analog dot com>, Binutils <binutils at sourceware dot org>
- Date: Mon, 05 Oct 2009 22:01:02 +0200
- Subject: Re: Add git mirror for binutils?
- References: <6dc9ffc80909291125m2e5b3d97q39e934bb6b72706f@mail.gmail.com> <4AC2FBBD.50002@analog.com> <874oqkygdm.fsf@meyering.net> <6dc9ffc80909300549s15529c0dy39d83e3c889c4c39@mail.gmail.com> <87y6ntlydz.fsf@meyering.net> <6dc9ffc80910051234l52b1726at8945e13d657e6a35@mail.gmail.com>
H.J. Lu wrote:
> On Sat, Oct 3, 2009 at 12:39 AM, Jim Meyering <jim@meyering.net> wrote:
>> H.J. Lu wrote:
>>> On Wed, Sep 30, 2009 at 1:40 AM, Jim Meyering <jim@meyering.net> wrote:
>>>> H.J. (or whoever can say, with authority),
>>>>
>>>> Do you want to keep all of the existing branches?
>>>> Are you interested in making any meta-data (authorship or log) changes,
>>>> e.g., if git commits should reflect an author different from the
>>>> cvs committer?
>>>>
>>>
>>> I only speak for myself. I only care about trunk and I'd like to keep all
>>> commit history so that git blame works correctly.
>>
>> Here's a first cut:
>>
>> Â Âgit clone git://sources.redhat.com/git/binutils
>>
>> Determining which subdirectories of src/ to include
>> and ensuring that the mirroring process does the right
>> thing with those, and only those, was a little tricky.
>> Please ensure that all of the required parts (and no more)
>> are included in a git clone.
>>
>> Also, while all branches *should* be included, I have made
>> no attempt to compare the contents of any git/cvs branches.
>> If you care about a particular branch, please ensure at least
>> that the same files are on the head of each.
>
> It works for me on Linux/x86-64. But it hasn't been updated.
> Can you fix that?
Sure.
Take this with a grain of salt.
I've just run the mirror script manually and watched thousands
of warnings like the following scroll by:
WARNING: Invalid PatchSet 14884, Tag binutils_latest_snapshot:
gas/config/obj-bout.c:1.16=after, gas/config/tc-m68k.c:1.69=before. Treated as 'before'
WARNING: Invalid PatchSet 14884, Tag binutils_latest_snapshot:
gas/config/obj-bout.c:1.16=after, gas/config/tc-m68k.h:1.22=before. Treated as 'before'
WARNING: Invalid PatchSet 14884, Tag binutils_latest_snapshot:
gas/config/obj-bout.c:1.16=after, gas/config/tc-maxq.c:1.7=before. Treated as 'before'
...
WARNING: Invalid PatchSet 15593, Tag binutils_latest_snapshot:
cpu/ms1.cpu:1.3=after, gas/testsuite/gas/mt/badinsn.s:1.1=before. Treated as 'before'
...
WARNING: Invalid PatchSet 18977, Tag binutils_latest_snapshot:
binutils/acinclude.m4:1.7=after, ld/Makefile.in:1.255=before. Treated as 'before'
WARNING: Invalid PatchSet 18977, Tag binutils_latest_snapshot:
binutils/acinclude.m4:1.7=after, ld/aclocal.m4:1.17=before. Treated as 'before'
WARNING: Invalid PatchSet 12622, Tag binutils_latest_snapshot:
bfd/elfarm-nabi.c:1.27=after, ChangeLog:1.531=before. Treated as 'before'
WARNING: Invalid PatchSet 12622, Tag binutils_latest_snapshot:
bfd/elfarm-nabi.c:1.27=after, configure:1.156=before. Treated as 'before'
WARNING: Invalid PatchSet 12622, Tag binutils_latest_snapshot:
bfd/elfarm-nabi.c:1.27=after, configure.in:1.246=before. Treated as 'before'
None of those should matter, since it looks like
they're all from the very early days, and hence predate
the mass conversion, and are ignored by git-cvsimport as
being "too early" to be useful.
I've set the mirroring cron job to run every 30 minutes,
starting at 20 and 50 after each hour.