[fold: different for branches and bookmarks]

[Pierre-Yves David]

On 12/19/2016 10:25 PM, Arne Babenhauserheide wrote:
>
> Uwe Brauer writes:
>
>>    > Uwe Brauer writes:
>>
>>
>>    > This sounds like a typical case for rebasing — or more elegantly using
>>    > the evolve extension.
>>
>>    > Upstream requires a single commit (one patch). So you’d ideally have
>>    > just that.
>>
>>    > With the evolve extension you’d simply issue
>>
>>    >   $ hg fold -r master -r feature
>>
>>
>> I just found out that fold behaves differently when using branches
>> instead of bookmarks. Using bookmarks, I obtain the same graph as you
>> but not for branches.
>
> That’s strange. I included evolve-testers in as recipient. This sounds
> like something to check.

The call `hg fold -r master -r feature` seems wrong. It will fold 
'(master+feature)::. + .::(master+feature)' Most certainly not what you 
want as master will be included in the resulting folded changesets.

You want `hg fold --exact 'only(master, feature)'

>> For branches I do:
>>
>>
>> hg init
>> echo feature > new.el
>> hg add new.el
>> hg commit -m feature1
>> echo offical2 >> new.el
>> hg commit -m feature2
>> echo feature3 >> new.el
>> hg commit -m feature3
>> hg branch feature
>> echo feature > feature.el
>> hg add feature.el
>> hg commit -m feature
>> echo feature2 >> feature.el
>> hg commit -m feature2
>> hg fold -r default -r feature
>>
>> Which gives the following graph
>>
>> @  changeset:   2:da5fe919132e
>> |  tag:         tip
>> |  user:        Uwe Brauer <oub at mat.ucm.es>
>> |  date:        Mon Dec 19 09:15:19 2016 +0000
>> |  summary:     feature3
>> |
>> o  changeset:   1:fe3d513ead50
>> |  user:        Uwe Brauer <oub at mat.ucm.es>
>> |  date:        Mon Dec 19 09:15:19 2016 +0000
>> |  summary:     feature2
>> |
>> o  changeset:   0:a6d03a55d8f1
>>    user:        Uwe Brauer <oub at mat.ucm.es>
>>    date:        Mon Dec 19 09:15:18 2016 +0000
>>    summary:     feature1
>>
>> While using bookmarks would give:
>>   changeset:   5:e06ba3130622
>> |  bookmark:    master
>> |  tag:         tip
>> |  parent:      1:80ab55999696
>> |  user:        Uwe Brauer <oub at mat.ucm.es>
>> |  date:        Mon Dec 19 09:17:22 2016 +0000
>> |  summary:     feature3
>> |
>> | x  changeset:   4:3ff9dbef4b64
>> | |  bookmark:    feature
>> | |  user:        Uwe Brauer <oub at mat.ucm.es>
>> | |  date:        Mon Dec 19 09:17:17 2016 +0000
>> | |  summary:     feature2
>> | |
>> | x  changeset:   3:6f10b3f58a52
>> | |  user:        Uwe Brauer <oub at mat.ucm.es>
>> | |  date:        Mon Dec 19 09:17:17 2016 +0000
>> | |  summary:     feature
>> | |
>> | x  changeset:   2:71f8b88b0831
>> |/   user:        Uwe Brauer <oub at mat.ucm.es>
>> |    date:        Mon Dec 19 09:17:16 2016 +0000
>> |    summary:     feature3
>> |
>> o  changeset:   1:80ab55999696
>> |  user:        Uwe Brauer <oub at mat.ucm.es>
>> |  date:        Mon Dec 19 09:17:16 2016 +0000
>>
>>
>> With the script
>>
>> hg init
>> hg bookmark master
>> echo feature > new.el
>> hg add new.el
>> hg commit -m feature1
>> echo offical2 >> new.el
>> hg commit -m feature2
>> echo feature3 >> new.el
>> hg commit -m feature3
>> hg bookmark feature
>> echo feature > feature.el
>> hg add feature.el
>> hg commit -m feature
>> echo feature2 >> feature.el
>> hg commit -m feature2
>> hg fold -r master -r feature

The issue here is that 'hg fold' seems buggy with bookmark and "forgot" 
to move the bookmark. This just seems like a bug.

-- 
Pierre-Yves David