.. in manifests not found / only partial clone

[Benito van der Zander]

Hi,
despite having set the extended +a attribute on all .i files, I again 
lost some history data :(

Two files in revision 3673.

$ hg verify
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
  components/pascal/data/tests/xpath3_tests.pas at 3673: 2e97f004def3 in 
manifests not found
  components/pascal/data/tests/xquery3_tests.pas at 3673: 5b6e569914b9 in 
manifests not found
4699 files, 3677 changesets, 14665 total revisions
2 integrity errors encountered!
(first damaged changeset appears to be 3673)


I tried to restore it like the RepositoryCorruption wiki says with

$ hg clone -r 3672 ~/hg fixed
adding changesets
adding manifests
adding file changes
added 3661 changesets with 14625 changes to 4698 files

but this fails with "rev points to unexpected changeset ... " later, 
because the clone is incomplete.

$ hg log -l 1
changeset:   3661:28da7a51c9ab
...

11 revisions are missing.

How can this happen?



Does clone not clone all branches?

Because 11 revisions is just the number of revisions in other branches 
than default.



In fact, when I try to clone the branches directly it fails:

$ hg branches
default                     3676:0fb3610282a4
videlibri-sk                3217:51428989e7dc
$ hg clone -b videlibri-sk . /tmp/wtf
abort: unknown branch 'videlibri-sk'!
$ hg clone -r 3217:51428989e7dc . /tmp/wtf
abort: unknown revision '3217:51428989e7dc'!

Yet the data is there:

$ hg diff -r 3216:3217
diff -r aef0b49cf497 -r 51428989e7dc 
programs/internet/VideLibri/android/android/src/de/benibela/videlibri/BridgeAndroid.java
....




(By the way it is very strange, that the example in the wiki has the 
broken revision number 7633
  while the broken revision in my repository has the number 3673. Does 
Mercurial not work with these digits?)


How can I fix this mess?



Best,
Benito