Greg Hudson wrote:
> On Wed, 2006-03-01 at 00:16 +0100, Branko Čibej wrote:
>
>>> Here's my brilliant idea: set bdb_cache to NULL in a pool cleanup for
>>> bdb_cache_pool. In the svn_fs_destroy handler, check if bdb_cache is
>>> NULL and don't touch the underlying environment handle if so.
>>>
>
>
>> Yes, I was thinking along those lines. Except that it means we can't
>> allocate the environment handles from pools, because then their lifetime
>> will by definition be shorter than the cache's.
>>
>
> Eh? I don't understand.
>
> apr_terminate invoked
> Environment handle pools are cleaned up
> Cache pool is cleaned up --> bdb_cache set to NULL
> FS objects are cleaned up, but don't touch environment handles
>
> What's the problem?
>
As I said in an earlier post, I'd have to check if the cache is there
every time the environment handle is used, not just in the svn_fs_t
cleanup code. Can be done, but it's messy.
Anyway, thanks for the suggestions, everyone. I'll think about them a
bit and decide what to do.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 1 01:36:17 2006