On Tue, 16 Apr 2002, Stas Bekman wrote:
> I see that you've solved it, Doug. But before we move on, can we please
> make it easier in the future to spot such a thing in the C code? What
> should be changed to get the C code trace ala Carp::confess? Can the
> taint checker figure out that the problem happens in the C code and not
> Perl and call something different than Perl_croak? I guess not. I'd
> prefer to have it segfault at the taint problem place (it dies in any
> case) so with debugger I could easily pinpoint the problematic code.
if $SIG{__DIE__} = \&Carp::confess doesn't work, then the problem is an
issue with perl. Perl_croak calls PL_diehook (aka $SIG{__DIE__}), if it
isn't printing the expected info, there isn't anything modperl can do
about it. even if Carp::confess did work, it wouldn't have helped to
solve this problem.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
