On 11/03/2016 10:03 AM, Brian Dolbec wrote: > On Thu, 3 Nov 2016 09:10:14 -0700 > Zac Medico <zmed...@gentoo.org> wrote: > >> On 11/02/2016 11:17 PM, Brian Dolbec wrote: >>> On Wed, 2 Nov 2016 21:57:09 -0700 >>> Zac Medico <zmed...@gentoo.org> wrote: >>> >>>> On 11/02/2016 05:17 PM, Zac Medico wrote: >>>>> On 11/02/2016 02:09 PM, Mike Gilbert wrote: >>>>>> The -q option allows git-update-index to succeed even if there >>>>>> are locally modified files present. >>>>>> >>>>>> X-Gentoo-Bug-URL: https://bugs.gentoo.org/552814#c58 >>>>>> --- >>>>>> pym/portage/sync/modules/git/git.py | 5 +---- >>>>>> 1 file changed, 1 insertion(+), 4 deletions(-) >>>>>> >>>>>> diff --git a/pym/portage/sync/modules/git/git.py >>>>>> b/pym/portage/sync/modules/git/git.py index dc94ec9..3db6dc1 >>>>>> 100644 --- a/pym/portage/sync/modules/git/git.py >>>>>> +++ b/pym/portage/sync/modules/git/git.py >>>>>> @@ -118,10 +118,7 @@ class GitSync(NewBase): >>>>>> if exitcode == os.EX_OK and self.repo.sync_depth >>>>>> is not None: # update-index --refresh is needed on some >>>>>> filesystems # (e.g. with overlayfs on squashfs) >>>>>> - update_index_cmd = [self.bin_command, >>>>>> 'update-index'] >>>>>> - if quiet: # -q needs to go first >>>>>> - update_index_cmd.append('-q') >>>>>> - update_index_cmd.append('--refresh') >>>>>> + update_index_cmd = [self.bin_command, >>>>>> 'update-index', '-q', '--refresh'] >>>>>> exitcode = >>>>>> subprocess.call(update_index_cmd, >>>>>> cwd=portage._unicode_encode(self.repo.location)) >>>>> >>>>> LGTM. >>>>> >>>>> I just checked the man page, and this behavior is documented >>>>> there: >>>>> >>>>> -q >>>>> Quiet. If --refresh finds that the index needs an update, the >>>>> default behavior is to error out. This option makes git >>>>> update-index continue anyway. >>>>> >>>> >>>> Actually we should probably add --unmerged, as suggested numerous >>>> times by Martin Väth. >>> >>> I suspect we won't hear the end of our breakage of a perfectly >>> working git module until maybe we add enough git-sync-options >>> configurable in repos.conf to account for every possible file >>> system and every possible repo useage/abuse-age... >>> >>> >>> Maybe we should just ship several versions of a git module... one >>> for every use case ;) (kidding) >> >> Well the Funtoo people are disproportionately affected by anything we >> do to this module, so we can expect to hear them complain loudly when >> it breaks. We shouldn't let that stop us from adding important new >> features though. I think Martin Väth is on the right track in >> suggesting that we use `git update-index -q --unmerged` like git >> status does. > > yes, But I have also told them from the start, with this plug-in > system, they can easily add their own custom git module that fits their > exact needs.
Sure I suppose they could. However, I'm pretty sure that our generic module will suit their needs as soon as we've worked out the bugs. -- Thanks, Zac
