develooper Front page | perl.perl5.porters | Postings from January 2009

Re: perldelta5101

Thread Previous
From:
Sam Vilain
Date:
January 5, 2009 20:12
Subject:
Re: perldelta5101
Message ID:
4962DA20.3010102@vilain.net
Sam Vilain wrote:
> git-log perl-5.10.0..origin/blead | git-shortlog
>
> Running this gives 1,896 lines, which I've attached (hopefully will
> display inline).
>   

Of course if 'origin' isn't what you think it is then you don't get what
you expected.  And I think 'perl-5.10.0..origin/maint-5.10' is probably
more accurate.  So sorry for the large and useless post.

After I ran this on the actual history, I noticed something
interesting.  Commit 296f12bbbbaa06de9be9d09d3dcf8f4528898a49 was
detected by the fuzzy logic in my importer to be 'close enough' to being
a cross-merge from blead to maint that it was marked as such.  This is
because the logic was just looking for some integration records that
were in the right revisions, and not checking that they exhaustively
integrated all of the corresponding changes.  For the revisions I
inspected this was the right thing.

However, using:

  git diff --stat 296f12bbbb 296f12bbbb^2

You can see that marking this as a cross-merge was definitely the wrong
choice - making this a nasty wart.

You can 'correct' this in your history, using:

 echo $(git rev-parse 296f12bbbb) $(git rev-parse 296f12bbbb^1) \
   >> .git/info/grafts

Running the same shortlog command on the maint-5.10 branch mostly yields
a whole lot of useless "Integrate:" shortlogs.  So, I guess this isn't
as easy as all that.

Sam.


Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About