Re: [bug] vtable=0x00000000 for the new FSFS when doing simultaneous commits

C.A.T.Magic wrote:
> i tried it the quick way by deleting
> /* SVN_ERR (svn_io_file_close (txn_file, pool)); */
> from svn_fs_fs__create_txn,
> and this fixes the txn collisions.
> (the file gets closed anyway, because of the pool).
> but for some strange reasons i still get

I have fixed this temporary file so that it works the way I originally
intended it to in r9660.

> svn: Unable to create new transaction.
>
> which means that apr_file_mktemp is failing -
> maybe a bug in apr for win32?

However, I am not sure why file_mktemp is still failing. I looked
through the APR code for this, and from what I can tell it can only
fail if the parent directory doesn't exist, creating the file returns
an error other than APR_EEXIST, or counting the random portion of the
string completely wraps around.

On a side note, it probably shouldn't fail on the first wrap-around,
since the start could be arbitrarily close to the wrap-around point,
but I doubt this condition is being hit here.

C.A.T.: Do you think you could figure out why apr_file_mktemp is
failing? Perhaps find the exact error code it is returning?

Thanks,
Josh

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun May 9 13:50:21 2004