Theo E. Schlossnagle wrote:
Josh Chamas wrote:
This is when an old $Response object is no longer valid. This is
normally
caused by my closure or bad scoping issues.
I assumed as much. However, it seems that:
sub AppendToLog { shift->{asp}->Log(...); }
should bde
sub AppendToLog {
my $self = shift;
if($self && $self->{asp}) {
$self->{asp}->Log(...);
}
}
in Response.pm. Why? My $Response variable still exists and as I don't
call Log directly it seems that it shouldn't return such a hard error
(die) when Apache::ASP makes that call internally.
No, $Response is no longer valid, and has been destroyed
during a prior request phase. It has only been cached incorrectly,
and this error results. Any use of $Response at this point would
be incorrect, so it does not make sense to spare the error with
a test for $Response->{asp}
Set these things to help catch these problems:
PerlSetVar UseStrict 1
PerlWarn On
I can do the later... But not the first. While we run both on almost
all our code bases, this particular code base is a legacy one (written
by another company) with over 200k lines of client .pm and .asp files.
(over 100k of custom perl modules alone). So, retroactively adapting
all legacy code to use strict is infeasible.
If you call Apache::ASP->Load() on all the files, you can find the
errors with UseStrict at compile time, so you do not get errors
at runtime. The process of converting over to UseStrict will catch
other errors besides the one that started this thread so would be
good for the code base to be converted to using.
Regards,
Josh
________________________________________________________________
Josh Chamas, Founder phone:925-552-0128
Chamas Enterprises Inc. http://www.chamas.com
NodeWorks Link Checking http://www.nodeworks.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
