Data corruption local rep, bitbucket repo has more heads than local one?

[Uwe Brauer]

Hi

I have a strange problem.
In my local repo

 hg heads gives me 
changeset:   92:77560cdc346d
tag:         tip
parent:      87:2700940f341d
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Tue May 23 16:14:53 2017 +0000
summary:     Correct H_0 en P4/H6

changeset:   72:8e632d9a64f4
branch:      Uwe
parent:      70:db25b273eba4
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Fri May 19 10:29:53 2017 +0000
summary:     Actualize matlab file for cal IC

changeset:   38:727faccf9762
branch:      linda
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Sat May 06 10:26:57 2017 +0000
summary:     Finish merge hoja3_corr en hoja3

And the graph looks ok
@  changeset:   92:77560cdc346d
|  tag:         tip
|  parent:      87:2700940f341d
|  user:        Uwe Brauer  <oub at mat.ucm.es>
|  date:        Tue May 23 16:14:53 2017 +0000
|  summary:     Correct H_0 en P4/H6
|
o  changeset:   87:2700940f341d
|  user:        Uwe Brauer  <oub at mat.ucm.es>
|  date:        Tue May 23 10:15:29 2017 +0000
|  summary:     H6: Repair region of acceptation of E1/E3/E4, correct small error with Sres in E4
|

That repo I push regularly to bitbucket, but somehow a data corruption
occurred.

When I clone the bitbucket repo to another directory and run
hg heads and hg log -G
the outcome is different

changeset:   80:77560cdc346d
tag:         tip
parent:      77:2700940f341d
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Tue May 23 16:14:53 2017 +0000
summary:     Correct H_0 en P4/H6

changeset:   79:fc056d395e04
parent:      77:2700940f341d
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Tue May 23 16:14:53 2017 +0000
summary:     Correct H_0 en P4/H6

changeset:   78:921b9b8667b1
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Tue May 23 16:14:53 2017 +0000
summary:     Correct H_0 en P4/H6

changeset:   62:8e632d9a64f4
branch:      Uwe
parent:      60:db25b273eba4
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Fri May 19 10:29:53 2017 +0000
summary:     Actualize matlab file for cal IC

changeset:   54:1f3c65063543
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Wed May 17 12:14:30 2017 +0000
summary:     Correct E20/21/H4

changeset:   52:a97d4ee3feb5
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Wed May 17 08:58:55 2017 +0000
summary:     Acclara 10

changeset:   41:aaf89a0885f5
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Mon May 08 15:10:08 2017 +0000
summary:     Quitar ejercico duplicado de estudio dental

changeset:   36:727faccf9762
branch:      linda
user:        Uwe Brauer  <oub at mat.ucm.es>
date:        Sat May 06 10:26:57 2017 +0000
summary:     Finish merge hoja3_corr en hoja3

And hg log -G
@  changeset:   80:77560cdc346d
|  tag:         tip
|  parent:      77:2700940f341d
|  user:        Uwe Brauer  <oub at mat.ucm.es>
|  date:        Tue May 23 16:14:53 2017 +0000
|  summary:     Correct H_0 en P4/H6
|
| o  changeset:   79:fc056d395e04
|/   parent:      77:2700940f341d
|    user:        Uwe Brauer  <oub at mat.ucm.es>
|    date:        Tue May 23 16:14:53 2017 +0000
|    summary:     Correct H_0 en P4/H6
|
| o  changeset:   78:921b9b8667b1
|/   user:        Uwe Brauer  <oub at mat.ucm.es>
|    date:        Tue May 23 16:14:53 2017 +0000
|    summary:     Correct H_0 en P4/H6
|
o  changeset:   77:2700940f341d
|  user:        Uwe Brauer  <oub at mat.ucm.es>
|  date:        Tue May 23 10:15:29 2017 +0000
|  summary:     H6: Repair region of acceptation of E1/E3/E4, correct small error with Sres in E4
|
o  changeset:   76:637df8ba9677
|  user:        Uwe Brauer  <oub at mat.ucm.es>
|  date:        Tue May 23 09:43:41 2017 +0000
|  summary:     Modify tables

Which reflects the new heads.

So I am puzzled. Which repo is corrupted the local one, or the bitbucket
one.

What would be the best solution?
Delete the bitbucket one and recreate it, pushing the local one
without the heads.
Delete the local one clone bitbucket and merge/rebase the heads?

Thanks

Uwe Brauer