A stupid problem with a nested repo

[Victor Sudakov]

Yuya Nishihara wrote:

[dd]

> > > > Anyway, "hg commit" in the parent repo committed the .hgsubstate file,
> > > > and the problem seems gone for now. 
> > > > 
> > > > Is this normal? Am I expected to commit the .hgsubstate file?
> > > 
> > > Yes, that's correct.
> > > 
> > > Until you committed the .hgsubstate, the main repo was tied to some older
> > > revision of the "namedb" repo. 
> > 
> > If that is the case, why didn't "hg status" in the parent repo show .hgsubstate
> > as "Modified"?
> 
> It's somewhat special. The .hgsubstate  file isn't "modified" until commit.
> The main repository is say dirty because subrepository contents don't match
> with the tracked state.
> 
> > > That's why "hg status -S" in the main repo showed modifications from that revision.
> > 
> > This is not my first nested repo, I don't ever remember that .hgsubstate
> > needed any special treatment. Is it documented?
> 
> "hg help subrepo" will provide some notion about .hgsubstate file. I don't
> know if it's well documented.

Well, it says 

" updating subrepos is a manual process. Simply check out target
  subrepo at the desired revision, test in the top-level repo, then commit
  in the parent repository to record the new combination. "

Thanks for showing me the right direction.

> 
> In general, you need to commit the main repo after committing changes to its
> subrepository.

I've mostly used "commit -S" so far, that is probably why I've never run
into this problem before. So, live and learn.

-- 
Victor Sudakov,  VAS4-RIPE, VAS47-RIPN
2:5005/49 at fidonet http://vas.tomsk.ru/