hg up and phases, surprises

[Pierre-Yves David]

On 5/27/23 08:43, Uwe Brauer wrote:
>> On 5/25/23 15:49, Uwe Brauer wrote:
>
>> Your changeset are not considered as part of the "default" branch
>> while there topic is active. The "default" keyword resolve to the head
>> of the un-topiced changeset on default. This was an important update
>> in topic behavior we made a while back, otherwise when multiple people
>> work in topics, the head of branch does not means anything and keep
>> jumping from one changeset to another.
> I am not so sure (but my evolve version is a bit old),
>
> Now, all local repositories have Publish true
> and I generate a dummy (local) remote repository to which I push
>
>
> hg init local
> hg init remote
> cd local
> echo "First line" >> test.org
> hg addremove
> hg ci -m "First commit test"
> echo "Second line" >> test.org
> hg ci -m "Second commit test"
> echo "Third line" >> test.org
> hg ci -m "Third commit test"
> hg up 0
> hg branch british
> hg topic english23
> echo "Secret line" >> test.org
> hg ci -s -m "First secret"
> echo "Secret line2" >> test.org
> hg ci -m "Second secret"
> hg up default
>
>
> Now
> hg up british
>
> Gives
> abort: unknown revision 'british'!
>
> I found that bizarre: all changesets on this branch a secret but why
> can't I upgrade to them


This has nothing to do with the secret phase, this has to do with all 
these changeset having a topic, so they did not "land" on that branch yet.

With latest evolve, you get the current output from `hg branch` that 
highlight that:

   > hg branches                                             ☿ ()
   british//english23             4:90d20203a429
   default                        2:3d11032b06ce

i.e. You have a "default" branch and a "english23" topic that will 
eventually land on the "british" branch. The "british" branch does not 
exist yet, however.



-- 

Pierre-Yves David