[Catalyst] Enabling debug mode with fastcgi..

Toby Corkindale toby.corkindale at strategicdata.com.au
Wed Jan 20 05:38:04 GMT 2010


On 19/01/10 21:45, Tomas Doran wrote:
>
> On 19 Jan 2010, at 00:26, Toby Corkindale wrote:
>> As I said, I've tried the -e (same as -keeperr) options in combination
>> with the *_DEBUG environment variables.
>
> That should (and does) work here..

Hmm. That's interesting to know.. I swear it isn't working on the 
machines I was testing this on the other day. Many things are old on 
them, so maybe it's something in a fastcgi library or other.
I'll investigate some more.

>> (The front-end apache servers aren't logging any catalyst debug output
>> anywhere; possibly this is a config error there, or possibly
>> intentional on the part of sysadmins - however either way it's not
>> something I'm about to change.)
>
> Yeah, that's got to be config of some sort, Catalyst does log down the
> fcgi socket by default..
>
>> For what it's worth, or anyone reading this in the archives in the
>> future - this is still on Catalyst Runtime 5.8.16.
>> (.17 was busted, .18 doesn't work out of the box in our environment
>> due to some of the changes and so isn't
>> considered production ready by us just yet.)
>
> I don't see a bug report for this anywhere? Did I miss it?

"make catalyst_par" doesn't produce a working PAR, as the 
Catalyst::ScriptRunner modules is not included inside the PAR.
Introduced in 5.8.17. Reported 15 Dec 2009.

Temporary fix is to add to Makefile.PL:
catalyst_par_classes('Catalyst::ScriptRunner');
catalyst_par_classes('Catalyst::Script::FastCGI');
..etc..

but of course if you do that, the package will fail to build on older 
pre-version-17 setups, and I didn't really want to get into the 
Makefile.PL complexity of optionally including them based on revisions 
of modules..

Cheers,
Toby



More information about the Catalyst mailing list