perl-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Philippe M. Chiasson" <>
Subject Re: Compat implementations of method_register and server_root_relative in tests ?
Date Mon, 09 Aug 2004 21:04:08 GMT

Philippe M. Chiasson wrote:
> Stas Bekman wrote:
>>Philippe M. Chiasson wrote:
>>>from todo/release :
>>>* the following methods/functions are using compat implementations in
>>> tests and should use the real 2.0 API: method_register,
>>> server_root_relative
>>>As far as I can see, this is not the case anymore, so should we remove that
>>>entry ? Or am I missing something ?
>>I'm not sure about method_register, since it didn't exist in 1.0. I 
>>think both has to do with the pool argument. Since method_register is 
>>now $s->method_register, there is no problem with it indeed.
>>As for server_root_relative, there is still an issue with it. Since we 
>>have the pool problem with this method, I've documented it as deprecated
>>and inefficient (since it copies the returned string, just in case). 
>>I've suggested to use:
>>   File::Spec->catfile(Apache::ServerUtil::server_root, @_);
>>as a better solution. One thing I've missed, and Geoff (as usual) has 
>>corrected me, is that it doesn't cover the case when the argument $_[0] 
>>is already a full path instead of relative. This is an undocumented 
>>feature as far as Apache docs go.
>>So back to the compat issue. At the moment lib/Apache/ has:
>>sub server_root_relative {
>>     my $class = shift;
>>     File::Spec->catfile(Apache::ServerUtil::server_root, @_);
>>which doesn't cover that special case. I suppose before doing catfile we 
>>need to check whether $_[0] is already absolute, in which case we should 
>>not concat server_root in.
> So would this patch resolve this open issue ?

Forgot to include test with the patch:

Philippe M. Chiasson m/gozer\@(apache|cpan|ectoplasm)\.org/ GPG KeyID : 88C3A5A5     F9BF E0C2 480E 7680 1AE5 3631 CB32 A107 88C3A5A5

View raw message