Geoffrey Young wrote:

What I meant is this:

        if (count != 1) {
            status = OK;
        }
        else {
            SV* status_sv = POPs;
            if (status_sv == &PL_sv_undef) {
                /* ModPerl::Util::exit returns undef */
                status = OK;
            }
            else {
                status = (IV)SvIV(status_sv);


I don't think you don't need that cast.

Sure, we don't need that case. It was a leftover from IVX.


ok, so the difference between the approaches is that mine is explicit, giving us some useful debug messages for coerced strings in case we have future problems.

I tend to suggest to keep things simple, we can always expand it in the future if we really encounter this problem. I don't think we ever had such a problem with mp1.


I just tested that the undef problem disappears (I can easily reproduce it) and the rest works just as before.


well, better than before, since the pv/iv issue is removed :)

;)


it doesn't matter to me whether we do it your way or my way, so long as the string issue is fixed. I happen to prefer more explicit debug messages myself.


I'll try again at getting a test reproduce this problem.


should be just a matter of setting PerlSwitches -w in the test suite (though who knows what that will start doing :)

Yup, it reveals a few more issues with undef. I'm looking at it.


I'll commit that change a bit later.

Also how about moving 'PerlSwitches -wT' into Apache::Test's autogenerator? Do you think any module developer will want not to have these turned on by default? I believe most will simply forget to add these, causing more problems.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:[EMAIL PROTECTED] http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com


--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to