Advocacy vs. git

[Martin Geisler]

Hans Meine <meine at informatik.uni-hamburg.de> writes:

> On Tuesday 28 April 2009 01:33:17 Martin Geisler wrote:
>> The histedit extension allows you to quickly decide on the fate of a
>> bunch of changesets by editing a text file in your editor. You make some
>> commits
>>
>>   [a] --- [b] --- [c] --- [d]
>>
>> and then decide to rework changesets b, c, and d:
>>
>>   hg histedit b
>>
>> This drops you into your editor with a file that looks like this:
>>
>>   pick 655c435efe92 b
>>   pick c47d17a27e19 c
>>   pick b04d79f9e2dd d
>
> Thanks.  That's very useful as a start, and Augie's README was very 
> informative, too.  I'll try it out now.
>
>> See also the tests here:
>>
>>   http://bitbucket.org/durin42/histedit/src/tip/tests/
>
> Actually, I *did* look at them (in the absence of any docs), but it's
> hard to understand them without knowing the context (stuff being shown
> in an editor etc.).

True... I had just seen so many blog posts praising 'git rebase -i' that
I forgot that not everybody was familiar with the concept.

> Looks like histedit could become really useful.
>
> It would be great if there was a way to keep track of "shared"
> changesets, i.e. ones that I were already pushed or cloned, such that
> editing them should be prohibited (or warned about). AFAICS it's not
> safely possible (in particular, anyone with a local login can clone my
> world-readable repo without me knowing it), but when everyone
> pushes/pulls to/from a shared central repo, it might be feasible (and
> useful).

It would be possible to keep track of what is pulled into the repository
and what is pushed out. But if the repository is readable by others
tracking can never be 100% accurate.

But what if I pull in some changesets and delete the original repository
because I don't need it any more -- the extension will then abort or
warn even though it is safe for me to edit the changesets. So the false
positive rate might be too big.

-- 
Martin Geisler

VIFF (Virtual Ideal Functionality Framework) brings easy and efficient
SMPC (Secure Multiparty Computation) to Python. See: http://viff.dk/.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
Url : http://selenic.com/pipermail/mercurial/attachments/20090428/ae427408/attachment.pgp