Re: can libsvn_fs handle parallel writes to the same txn?

Carlos Alberto Costa Beppler writes:
> I don´t know the internals of a commit operation, but is it composed
> only of a write operation? The are many tests "reads" that could be
> done in paralel, leaving only the "write" part do be serialized. Am I
> wrong?
>

I think you're correct in that we couldn't and shouldn't disallow
multiple entities having the same transaction open for reading.

I think we should error out if multiple processes/threads try to modify
a transaction simultaneously, because
a) it isn't supported
and
b) if we instead choose to block on this, clients could get stuck
waiting for other write operations for long, without knowing
why.

Better to just bail with a specific error code.

I had a design in mind when I was going to sleep yesterday.
I'll try to resurrect it tonight...
Thanks,
//Peter

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue May 23 08:57:15 2006