Re: porting Apache::File issues

2001-11-28 Thread Stas Bekman
>>compatibility. The only issue is with tmpfile. Where should it live? >> > > copy-n-paste the 1.x code for now. in the future, would > like to see APR::File->mktemp (apr_file_mktemp) and a perlio layer defined > in terms of apr_file_t to use it. ok BTW, from perl571delta.pod: Anonymous tem

Re: perlrequire/perlmodule are running twice and slow

2001-11-28 Thread Daniel Jacobowitz
On Tue, Nov 27, 2001 at 09:27:18AM +0800, Stas Bekman wrote: > Daniel Jacobowitz wrote: > > >On Tue, Nov 27, 2001 at 02:02:57AM +0800, Stas Bekman wrote: > > > >>I've a bunch of them, but I think that the one in /lib/ that's used: > >> > >>% urpmf libthread_db > >>glibc:/lib/libthread_db-1.0.so >

Re: porting Apache::File issues

2001-11-28 Thread Doug MacEachern
On Fri, 23 Nov 2001, Stas Bekman wrote: > I think the following addition to compat.pm covers Apache::File looks good, +1 > compatibility. The only issue is with tmpfile. Where should it live? copy-n-paste the 1.x code for now. in the future, would like to see APR::File->mktemp (apr_file_mkte

Re: un-acceptable function prototypes in Wrap/XS?

2001-11-28 Thread Doug MacEachern
On Thu, 29 Nov 2001, Stas Bekman wrote: > do we have such a function? I couldn't find it. I've started to grab > strings in loop over av_len: > > av_argv = (AV*)SvRV(ST(1)); > (const char *)SvPV_nolen(AvARRAY(av_argv)[i])) > > but I have a gut feeling that it's not very effective way to go.

Re: source_scan doesn't pick up new MPXS_ method :(

2001-11-28 Thread Doug MacEachern
On Wed, 28 Nov 2001, Stas Bekman wrote: > that's not the problem. The problem is that the xs_generate has a bug > and it won't create Wrap/Apache/Subprocess/Apache__SubProcess.h if there i sure hope it wouldn't create that file! > are only MPXS_ functions in the map file. I've added a dummy

Re: un-acceptable function prototypes in Wrap/XS?

2001-11-28 Thread Stas Bekman
Doug MacEachern wrote: > On Mon, 26 Nov 2001, Stas Bekman wrote: > > >>I've learned that WRAP/XS cannot handle 'char **' prototypes. It just >>steps over the function. >> > > we shouldn't be using char ** prototypes for anything. yes, yes, I've moved on since then, it was my initial thought

Re: un-acceptable function prototypes in Wrap/XS?

2001-11-28 Thread Doug MacEachern
On Mon, 26 Nov 2001, Stas Bekman wrote: > I've learned that WRAP/XS cannot handle 'char **' prototypes. It just > steps over the function. we shouldn't be using char ** prototypes for anything. > I still wasn't able to user the same workaround used for > modperl_bucket_read. Any solutions fo