Interim maintainer tree

Junio is on vaction for the next week. In his absence Peff and I are trying to keep up with current patches in my fork:

  git://repo.or.cz/git/spearce.git
  http://repo.or.cz/r/git/spearce.git

Right now the tree matches Junio's last push, I'll try to pick up the patches since then and push later today.

On Fri, Sep 25, 2009 at 09:05:04AM -0700, Shawn O. Pearce wrote:

I've collected a few more topics in my repo:

  git://github.com/peff/git.git

Rather than cook my own, slightly different 'next', I'm just publishing the tips of the topics themselves. Which means they are subject to rewinding until you merge them into 'next', at which point I will leave them stable.

What's in there now is:

  as/parseopt-ambiguous
    Fix for "git branch --no-merge" with my test added. Should be ready
    for next.

  jk/reflog-date
    My reflog fix. Should be ready for next.

  mr/gitweb-snapshot
    v5, replacing what Junio had in pu. Should probably stay in pu until
    we hear from Jakub.

  mr/instaweb
    Looked OK to me, but I have no capacity to actually test it.
    Unlikely to break anything else, though, so probably ready to cook
    in next.

  mv/lock-error
    My tweaked version, which got the OK from Miklos. Ready for next, I
    think.

  np/clone-smaller
    Nico's fix to stop copying unnecessary objects during clone. Patch
    looks sane, I confirmed that it fixes the issue, and it doesn't seem
    to break normal clones. Ready to cook in next, I think.

-Peff

"Shawn O. Pearce" <spearce@spearce.org> writes:

Thanks, both.

I've fetched, but haven't fully examined "log ..spearce/*" nor "log ..peff/*" yet.

I noticed that some topics in 'pu' have been rebased (not complaining, but just making sure I am not hallucinating).

Do you have preferences/suggestions as to how to proceed? Should I just take the tips over at this point, or do you have some more patches you were polishing but haven't pushed out that I should wait for?

On Sat, Oct 03, 2009 at 11:06:30PM -0700, Junio C Hamano wrote:

For my stuff, I think you can just take over my tips. I was just collecting and polishing topics, while Shawn was doing integration. For topics I have, please use my versions rather than applying from the list; many of them have extra fixes squashed in, acked-by's collected from the list, etc.

Here's a brief status report on what's in my tree:

  proposed-master
    One-off patches that I think should go straight to master.

  ef/mscv-noreturn
    This is the latest round and I think should be ready for at least
    'next' (maybe even 'master' as it is really about the build and not
    about functionality).

  ef/msys-imap
    This is from an RFC which has generated some comments. He should be
    posting another round soon. 'pu' at best.

  fc/mutt-alias
    Latest round that addressed comments. Ready for 'next' if not
    'master'.

  jn/gitweb-patch
    After some comments with Jakub, I think the code is right but he
    promised a re-roll with more in the commit message.

  mr/gitweb-snapshot
    This is probably the pu rewind you saw. He posted a v5 of his
    series. I didn't look at it closely, but Jakub ack'd it.

  tf/doc-pt-br
    Minor translation update, ack'd by somebody who can read it. :)
    Ready for 'master'.

There are a few patches floating on the list that I haven't picked up or looked too closely at yet. Just so we don't drop anything, they are:

  - curl http auth tweak:
    http://article.gmane.org/gmane.comp.version-control.git/129455
    The author said he didn't really test it, and I haven't set up http
    auth to test it with, but probably somebody should do so before
    applying. :)

  - a new rev-cache from Nick

  - silence gcc warning:
    http://article.gmane.org/gmane.comp.version-control.git/129485
    The warning is overly cautious, I think, but it is a dubious enough
    construct that it is probably worth fixing.

  - enable openssl on msvc
    http://article.gmane.org/gmane.comp.version-control.msysgit/7238
    This goes on top of ef/msys-imap, but I think that will be getting a
    re-roll.

-Peff

Jeff King <peff@peff.net> writes:

Many topics Shawn merged to his 'pu' were identical to what you have, except for mr/gitweb-snapshot. The patch-ids in the series are the same but you have Jakub's Ack added, so I'll pick yours up and rebuild.

I briefly looked at them after looking at what Shawn queued to his 'master'; they all looked sane.

I've queued all of these to 'pu' for tonight as I haven't still fully recovered yet from jetlag, but I agree with the above assessment.

Heh ;-)

What Shawn had in his 'pu' had one patch removed from what I had earlier (perhaps my 'tip list' was faulty). I'll pick the new one up later.

I cannot reach gmane now, but if this is about -Wextra, I'd rather not touch it before the release. "comparison between signed and unsigned" tends to be excessive and IMNSHO it is crazy to use -Wextra and -Werror together.

Ok.

Thanks.

On Sun, Oct 04, 2009 at 02:54:08AM -0700, Junio C Hamano wrote:

Actually, I mislabeled this. It is about a glibc run-time complaint not a gcc warning, which is more worrisome. It is triggered by an argv list where we pass (argv - 1) to something expecting only to index it starting from '1'. I'm not sure yet if it is glibc being picky or if we violate that assumption somewhere. I'll post a followup in the thread.

-Peff

Junio C Hamano <gitster@pobox.com> wrote:

Sorry, I proved to be a very poor interim maintainer this cycle.
I just didn't have the time to keep up with integration each day,
and then dropped the ball on Thursday and Friday and failed to pull
anything from Peff or the list.  Peff, thanks for keeping up a bit
and at least having some tips for Junio to pick up from.
 

Yes, some topics in 'pu' got rebased. Wednesday I ran RB after preparing a new master, and then rebuilt pu around it. Unfortunately I somehow lost a patch on Nick's rev-list series, not sure how I managed to do that, but I guess I did, sorry.

Were you no longer in the habit of running RB when you rebuilt pu?
 

On Mon, Oct 05, 2009 at 07:56:15AM -0700, Shawn O. Pearce wrote:

It was an interesting experience this time, having two people doing it. I think the communication overhead may have overwhelmed any help we were giving each other. Perhaps if there had been more, larger series, it might have made sense to assign somebody to track a topic and its iterations, and then finally sign off on it being ready. But being -rc period, things seem to have settled down, and most patches were one-off bugfixes all over the place.

-Peff