[PATCH 3 of 6 V3] histedit: replace various nodes lists with replacement graph (and issue3582)

[Sean Farley]

On Wed, Oct 10, 2012 at 10:54 AM, Augie Fackler <raf at durin42.com> wrote:
> (responding on the list for posterity)
> On Oct 10, 2012, at 3:59 AM, pierre-yves.david at logilab.fr wrote:
>
>> @@ -140,30 +139,29 @@
>>> pick d36c0562f908 1 c
>>> pick 0efacef7cb48 3 f
>>> pick ae467701c500 2 d
>>> EOF
>>   $ hg histedit 1 --commands commands.txt --verbose | grep histedit
>> -  histedit: Should update metadata for the following changes:
>> -  histedit: moving bookmarks five from 0efacef7cb48 to 1be9c35b4cb2
>> +  histedit: moving bookmarks three from ae467701c500 to 1be9c35b4cb2
>>   histedit: moving bookmarks four from ae467701c500 to 1be9c35b4cb2
>> -  histedit: moving bookmarks three from ae467701c500 to 1be9c35b4cb2
>> +  histedit: moving bookmarks five from 0efacef7cb48 to 7c044e3e33a9
>>   saved backup bundle to $TESTTMP/r/.hg/strip-backup/ae467701c500-backup.hg (glob)
>>
>> We expect 'five' to stay at tip, since the tipmost bookmark is most
>> likely the useful signal.
>>
>>   $ hg log --graph
>>   @  changeset:   3:1be9c35b4cb2
>> -  |  bookmark:    five
>>   |  bookmark:    four
>>   |  bookmark:    three
>>   |  tag:         tip
>>   |  user:        test
>>   |  date:        Thu Jan 01 00:00:00 1970 +0000
>>   |  summary:     d
>>   |
>>   o  changeset:   2:7c044e3e33a9
>> +  |  bookmark:    five
>>   |  user:        test
>>   |  date:        Thu Jan 01 00:00:00 1970 +0000
>>   |  summary:     f
>>   |
>>   o  changeset:   1:d36c0562f908
>
>
> This change in test output isn't right: the workflow I use with histedit (which I should blog about) depends on tipmost bookmarks staying tipmost. I could be swayed that it's a defect for inactive bookmarks (I think), as that wouldn't break my particular workflow.

This is something I've been meaning to bring up and ask Pierre-Yves
about: with the new obsolete features in mercurial, we'll have the
ability to leave old changesets around and thus not move inactive
bookmarks. This has been brought up on the issue tracker before but
dismissed by Matt:

http://bz.selenic.com/show_bug.cgi?id=3586

It seems that with Pierre-Yves' new stuff that we'll be able to
implement such behavior: never move inactive bookmarks, ever.
Thoughts?