Had this problem, yes. Solved it, no. Accepted it, yes.
I have wrapped evals around all the calls to find_or_create in my
code and catch that error. I still have issues with the flash (and
session) with find_or_create. THis may just be coincidence but in
the error emails I generate, Internet Explorer is responsible for all
of these errors. Go figure.
On Sep 25, 2007, at 11:22 PM, Tobias Kremer wrote:
Quoting Ian Docherty <[EMAIL PROTECTED]>:
Did I not see something on here, or on DBIx-class list some time
ago to
the effect that in find_or_create there is a short window where
having
found that a record does not exist it then creates it. I presume that
two threads could each detect that a record does not exist and
then one
of them would create this error.
Hmm .. That would explain the other issue I'm facing: An
update_or_create
command that once in a while fails with a duplicate entry error.
The two
threads you're mentioning would have to come from the same user at the
same time though to cause this problem assuming that no two users have
identical session keys ...
Reversing the order (try a create first, catch the error, then do
a find
if the create fails) was ISTR the proposed answer.
But this may not help you if the problem is in the
Session/Store/DBIC/Delegate plugin.
Right :( Is this a problem that only affects the MySQL backend
because the
other storage types (memcached, mmap, etc) do not check the uniqueness
of the session key?
Has nobody else met this problem yet?
--Tobias
_______________________________________________
List: Catalyst@lists.rawmode.org
Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst
Searchable archive: http://www.mail-archive.com/
catalyst@lists.rawmode.org/
Dev site: http://dev.catalyst.perl.org/
_______________________________________________
List: Catalyst@lists.rawmode.org
Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst
Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/
Dev site: http://dev.catalyst.perl.org/
