My Hg Workbench is sometimes very slow. I am under windows7, using last version of Tortoise Hg. I have several large repositories being tracked by Hg locally For example, I have one repository that
- has been converted from CVS using cvs2hg (from svn2hg)
- has more than 7000 changesets
- is more than 1GB heavy
- uses extensions largefiles, hgk, mercurial_keyring
- has several binary files being tracked as large files
- has a few binary files not tracked as large files( including one file of 300 MB)
With this repo, it can take
- from 10 seconds, up to 30 seconds to open the repository
- from 3 seconds, up to 10 seconds to change a changeset
and sometimes, doing a filter in the changeset list or doing a diff with local from an old revision can take several minutes.
Note 1 : This is the case for all the team (approx 10 people) using this repository, not only on my PC
Note 2 : in the past, it was even much much longer but that was due to the collumn "Changes" being displayed. Now, I removed it and it is not as bad but far from good anyway.
Note 3 : this repos has been cloned from a RhodeCode server but the problem is in local configuration, not during push or pull operation, so I think it is not related.
Note 4 : after cvs2hg conversion, I had a problem of End Of Line (I had been unable to make EOL conversion to work properly with the convertor, I don't know why). So after the conversion, I had to apply a EOL conversion on all my text files, to commit and to push. That made quite a big changeset which affected everyline of every text file. I am not sure this could be the reason for my problem because as far as I remember, it was already slow before this conversion, but I prefered to notice it anyway.
In any cases, this slowness comes from the workbench itself : when I browse the repository on the rhodecode server, it is not slow (or far to be so slow).
The main problem is that I have been pushing for several monthes to switch from cvs to mercurial and as it is slow, I am afraid my boss wants to revert back to cvs ... that is why I set the priority to Critical.
1/ Do you think my pb comes from a bug (or a lack of optimzation) inside the workbench or is it the result of something I did (in which case there would be no issue)
2/ Is there a way I could have a profile log or even a general log that could give me information about what is taking so long ?
Tell me if you need any other information, or if I can help in any way to make things move forward on this issue
Thanks in advance