Re: [PATCH] Error Page code upgrades

From: Adrian Chadd <adrian_at_freebsd.org>
Date: Tue, 22 Jul 2008 12:56:00 +0800

2008/7/22 Amos Jeffries <squid3_at_treenet.co.nz>:

>
> Possibly, I'm using the pre-existing load function to load each file. If
> you have a better method please let me know, this would be a good time to
> upgrade it.

But if I read that code right, you're loading in the file every time
someone wants an error, rather than loading it in once at startup and
just using that.

Modifying the error body appending stuff to be async may be nice but
is probably very overkill for this situation. Is there any way to load
in all of the error files and write some stuff to do inline character
set translation before blatting them out?

I also suggest first doing the work to convert ErrorState over to
being a class/set of classes as appropriate whilst maintaining the
current behaviour and do a release cycle with that. 3.1 is nearby; I
-could- be convinced not to complain about committing ErrorState
classify stuff if it was throught through and tested enough, but I
don't think its worth trying to work on automatic language and
character set encoding until there's been a lot more discussion and
design.

Adrian
Received on Tue Jul 22 2008 - 04:56:02 MDT

This archive was generated by hypermail 2.2.0 : Tue Jul 22 2008 - 12:00:06 MDT