Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Dag-Erling Smørgrav
Will Crawford billcrawford1...@gmail.com writes: Dag-Erling Smørgrav d...@des.no writes: Is there a way to have C::Plugin::Session (or C::P::S::Store::FastMmap) initialize the cache on first use instead of at startup? { ..., unlink_on_exit = 0 } in your config. delete the cache when you're

Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Ashley Pond V
On Wed, May 23, 2012 at 1:58 AM, Dag-Erling Smørgrav d...@des.no wrote: As an alternate solution, is there a way to have it use a different file name every time it starts? Here's a snippet to do that: conf.yml -- Plugin::Session: storage: /tmp/some-prefix-__UID__.session MyApp.pm --

Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Alexander Hartmaier
Because I've just read perlvar: $REAL_USER_ID or $UID instead of $. - Alex Am 2012-05-23 15:54, schrieb Ashley Pond V: On Wed, May 23, 2012 at 1:58 AM, Dag-Erling Smørgrav d...@des.no wrote: As an alternate solution, is there a way to have it use a different file name every time it starts?

Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Dag-Erling Smørgrav
Dag-Erling Smørgrav d...@des.no writes: Thanks, that's a neat workaround (combined with unlink_on_exit). I was wrong about that last bit - unlink_on_exit automatically defaults to a sane value (unlink the file if you created it, leave it alone if it already existed). If anything, you'd want to

Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Dag-Erling Smørgrav
Alexander Hartmaier alexander.hartma...@t-systems.at writes: Because I've just read perlvar: $REAL_USER_ID or $UID instead of $. assuming use English DES -- Dag-Erling Smørgrav - d...@des.no ___ List: Catalyst@lists.scsys.co.uk Listinfo:

Re: [Catalyst] Plugin::Session lazy start

2012-05-23 Thread Ashley Pond V
On Wed, May 23, 2012 at 7:42 AM, Alexander Hartmaier alexander.hartma...@t-systems.at wrote: Because I've just read perlvar: $REAL_USER_ID or $UID instead of $. I think this is generally excellent advice but I want to argue against it in this case. UID = sub { $ } is sufficiently semantic *and*