[non topological heads]

[Matt Harbison]

On Fri, 02 Jun 2017 08:38:57 -0400, Uwe Brauer <oub at mat.ucm.es> wrote:

>
>    > On Thu, 01 Jun 2017 06:56:47 -0400, Uwe Brauer <oub at mat.ucm.es>  
> wrote:
>
>    > I've got to ask a stupid question, since IDK what the corruption is  
> or
>    > where it is.
>
>
>    > Can you clone from BB, and then from inside that local
>    > clone, pull from the corrupted repo?
>
> You mean the corrupted repo which I did not repair with convert? I did
> not try this out.
>
> Ok when I do that (in the BB clone) I obtain
>
> hg pull ../Orginal
> pulling from ../Orginal
> searching for changes
> adding changesets
> adding manifests
> adding file changes
> remote: abort: unexpected error: empty or missing revlog for  
> Lista/Examen-Geo2/anova.csv
> transaction abort!
> rollback completed
> abort: pull failed on remote
>
>    > Even if you can't pull everything, maybe you can specify a revision
>    > and pull in pieces at a time?
>
>
> So I just pull the first new commit.
>
> That is the last clean commit in BB is 7cdd8585c0ac
>
> The first new one in the corrupt repo  is add711370282
>
> So I run
>
> hg pull -r  add711370282 ../Original
>
> abort: data/Lista/Examen-Geo2/anova.csv.i at 3107306eb740: no match found!
> warning: outgoing hook exited with status 255
> adding changesets
> adding manifests
> adding file changes
> added 1 changesets with 3 changes to 3 files
> 9 new obsolescence markers
> (run 'hg update' to get a working copy)
> abort: data/Lista/Lista-Ej/lista.org.i at f68105cb677d: no match found!
> warning: changegroup.diffstat hook exited with status 255
>
> Hg log -G
>  changeset:   519:add711370282
> |  branch:      year-2016-17
> |  tag:         tip
> |  user:        Uwe Brauer  <oub at mat.ucm.es>
> |  date:        Tue May 16 17:43:22 2017 +0000
> |  summary:     Small changes in Hoja6/8
>
> But
> hg verify
> Lista/Lista-Ej/lista.org at 519: manifest refers to unknown revision  
> f68105cb677d
>  README.org at 519: manifest refers to unknown revision d721d07b2857
> 1168 files, 520 changesets, 1777 total revisions
> 2 integrity errors encountered!
> (first damaged changeset appears to be 519)

Ouch.  It looks like the transaction abort is missing from the output of  
the second one.

> Hg strip -r 519
>
> abort: data/Lista/Lista-Ej/lista.org.i at f68105cb677d: no node!
>
> Now this is *nice*. I'd say this is even a bug. That should not have
> happened either the pulling should have been aborted or the strip should
> have been worked.

Yeah, once it is corrupted, I can see that strip may not work.

Is this public code?  i.e. can you make the initial BB repo and the  
converted one available somewhere, along with a zipped up copy of the  
corrupt repo?  Looking at it will be easier than going rounds via email.

>
>    > I don't understand. `hg help glossary` says a topological head is a
>    > commit with no children. How can something be a head but not a
>    > topological head?
>
>
>  Now I am confused
>
> Why is there a difference between
>
> hg heads
>
> And
>
> hg heads -t

See Mischa's response and/or `hg help heads`.

>
> Thanks again. I presume I have to life with the 'repaired version' via
> convert and forget the BB repo, or better say regenerate it on BB
>
> Uwe
>
> _______________________________________________
> Mercurial mailing list
> Mercurial at mercurial-scm.org
> https://www.mercurial-scm.org/mailman/listinfo/mercurial