Re: [Pan-users] Pan 0.99999999999999

[Duncan]

Rhialto posted on Thu, 16 Jun 2011 16:33:03 +0200 as excerpted:

> On Thu 16 Jun 2011 at 16:19:22 +0200, Rhialto wrote:
>> Any ideas about what could have happened?
> 
> Hmmm it looks like K. Haley's fixes have been rebased. This commit:

I (gmane) seems to be missing your original post as you quoted above, 
message-id address@hidden according to the refs, so I 
don't know exactly what you're referencing, but it's threaded under Ron 
Johnson's post mentioning the git branch with the binary uploads, so 
presumably... you're talking about Heinrich Mueller's branch.  It's 
possible he rebased.

But as a recent thread on LKML (as covered by H-Online's kernel log) 
pointed out, it could also be possible that a fix was committed, then 
another tree merged in reverting that commit due to choosing the wrong 
side of the merge when the conflict came up, then finally, the fix-commit 
again appearing with a re-sync with the original tree and a second 
merge.  (The thread on LKML was in the context of what this does to 
bisect, making certain issues trace to a totally unrelated commit due to 
bisect picking the wrong parent in a merge to bisect down in a complex 
scenario like the above... there's some changes most likely coming to git-
bisect as a result of all that.  I can probably find the H-Online 
coverage and thus the thread again, if someone's interested.  Both 
watching the understanding unfold as the thread progressed and seeing 
Linus's eventual analysis of the whole thing was... fascinating.)

Meanwhile, I run khaley's testing branch here and once saw a duplicated 
commit on it as well, but figured it's the testing branch, so a screwup 
or rebase once in awhile isn't unexpected.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman