On Jan 19, 2005, at 12:19 PM, Philip Martin wrote:
> fitz@tigris.org writes:
>
>> Author: fitz
>> Date: Mon Jan 17 23:09:32 2005
>> New Revision: 12773
>
>> --- branches/locking/subversion/libsvn_fs_fs/lock.c (original)
>> +++ branches/locking/subversion/libsvn_fs_fs/lock.c Mon Jan 17
>> 23:09:32 2005
>> @@ -756,6 +756,7 @@
>> svn_lock_t *new_lock;
>> svn_fs_root_t *root;
>> svn_revnum_t youngest;
>> + apr_pool_t *subpool = svn_pool_create (pool);
>>
>> SVN_ERR (svn_fs_fs__check_fs (fs));
>>
>> @@ -795,6 +796,8 @@
>> path);
>> }
>>
>> + SVN_ERR (svn_fs_fs__get_write_lock (fs, subpool));
>> +
>> /* Is the path already locked?
>>
>> Note that this next function call will automatically ignore any
>> @@ -826,6 +829,9 @@
>> SVN_ERR (write_lock_to_file (fs, new_lock, pool));
>> *lock_p = new_lock;
>>
>> + /* Destroy our subpool and release the lock. */
>> + svn_pool_destroy (subpool);
>> +
>
> At present you are using the subpool to manage the lock; that's fine
> but how about using the subpool to manage more resources? Obviously it
> should not be used to allocate any memory that gets returned from the
> function, but it could get passed instead of pool for anything that
> only requires transient allocations.
*nod*
I should have put a TODO in the code to do this instead of in my
personal notes file. I'll hopefully get to that today (Friday) as well
as Lundblad's suggested doc tweak and token check.
Thanks,
-Fitz
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Jan 21 08:44:53 2005