I’ve been running mercurial to interface with our subversion repo, today, when pulling a large changeset (hg pull --rebase), I get an out of memory error.
The changeset is pretty big: roughly 5.7 million lines, all but 10 in one csv file (2.8 million insertions, 2.8 deletions), still I’d like to stick with mercurial and not have to downgrade back to svn.
Is there a way I can increase the amount of memory Mercurial can use, or some other way to pull the commit?
I’m running the Mercurial version 2.0.1 that came bundled with a slightly old version of TortoiseHg.
In principle Mercurial uses as much memory as it needs to, so it sounds like your machine doesn’t have enough memory to hold all data structures needed for SVN-Mercurial conversion.
What I would suggest you try is, copy your repo on an USB stick, and then pull from SVN on a machine with more internal memory. Then copy (or pull) the repository back to your machine.