bzr move duplicates content of binary files until repacked

[Leonardo Montecchi]

Hello,
I have tried to add to a bzr reository a directory containing 10-20 binary files of about 2 MB each (photos).
Whenever trying to move (i.e. rename) the files it seems that the whole content of the files is duplicated in the repository, instead of just storing the new name.

Steps to reproduce:

bzr init
--copy into the repository a binary file--
cp /home/user/Pictures/test.jpg
bzr add test.jpg
bzr commit
--check the size of the repository (it should be about the size of the binary file)--
du -sh .bzr/
--rename the file--
bzr mv test.jpg bug.jpg
bzr commit
--check the size of the repository--
du -sh .bzr/

You will notice that the size of the repository has doubled.
Is this an intended behavior?

[Leonardo Montecchi (lmontecchi)]

Bazaar (bzr) 2.2.1
  Python interpreter: /usr/bin/python 2.6.6
  Python standard library: /usr/lib/python2.6
  Platform: Linux-2.6.35-23-generic-x86_64-with-Ubuntu-10.10-maverick
  bzrlib: /usr/lib/python2.6/dist-packages/bzrlib
  Bazaar configuration: /home/montex/.bazaar
  Bazaar log file: /home/montex/.bzr.log

[Max Bowsher (maxb)]

Following on from the previous commands, try:

bzr pack --clean-obsolete-packs
du -sh .bzr/

The duplicated storage is, in my experience, reclaimed.

Granted, ideally it would never be duplicated even temporarily, but this is better than it being duplicated permanently.

[John A Meinel (jameinel)]

We record a new entry in the per-file graph for the object, because it "changed" somehow (in this case, a new parent directory). At the moment this requires writing a text. Commit always stores unpacked texts (without delta compression against existing content).

When autopack/'bzr pack' decides to combine the contents, it will see that they are identical and then the duplication will be removed.