perl-modperl mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tyler MacDonald <>
Subject Re: registry agony continues...
Date Fri, 04 Nov 2005 23:04:07 GMT
Tyler MacDonald <> wrote:
> > Can you try this on another system?  That would tell you whether or not
> > there is truly anything wrong with the code, or if you should be looking
> > at reinstalling this system.
> 	I'm going to test this extensively this weekend if I get the time,
> including testing it on another system. Unfortuantely the other system is
> also perl 5.6.7, httpd 2.0.55, linux 2.6, so if I get the same problems more
> testing will be neccessary...
> 	The problem is with compiling code after the apache server has
> already started (ModPerl::Registry, maybe eval qq{} as well, not sure); if I
> compile everything during the startup phase, everything's happy.

	OK, I haven't tested this *extensively* yet, but I have found a
solution that seems to work; If I add:

	PerlSwitches -C0 -Mbytes

	To my httpd.conf, everything seems to be happy. This forces perl to
look at all data it is handed in byte context, rather than fiddling around
in unicode.

	As for why the problem was intermittent, I'm guessing that some .pm
files had unicode/UTF8 characters in them, which was causing the perl
interpreter to shift itself into unicode mode. As far as I understood, this
shift was supposed to be scope-specific, so my entire interpreter shouldn't
have gotten borked, but there you have it.

	Oddly enough, setting my system's locale to "C" wasn't enough to fix
this. These inconsistencies lead me to believe that perl's unicode support
is dubious at best and that I should try to stay as far away from it as I


View raw message