At 5:45 PM -0700 4/10/03, Michael G Schwern wrote: >Still trying to build HTML::Parser. Most if not all extension building problems on VMS with 5.8.0 were fixed by change #17688: <http://public.activestate.com/cgi-bin/perlbrowse?patch=17688&action=patch> >Comparing a Unix 5.8.0 CORE/ directory >with a VMS 5.8.0 CORE/ dir I've found the following .h files gone missing: > >dosish.h >fakesdio.h >fakethr.h >perliol.h >perlsfio.h >reentr.h >uconfig.h >unixish.h >utfebcdic.h > >Of those, MakeMaker on VMS expects to see perlsfio.h and fakethr.h. > >Normally installperl just shoves all the .h files into CORE/. VMS has >special logic to install only those header files placed into a CORE/ dir >during the build process. Rather than having to maintain that logic, >why can't VMS just shove all the .h files into CORE/ like everyone else? In principle I agree with you that there is something pretty crufty about the way we do it now. It's a fairly major surgery to the core build, though, so we'd need to proceed with caution. >The "copy some .h files into a temp CORE" logic in vms/descrip_mms.template >has this comment: > ># CORE subset for MakeMaker, so we can build Perl without sources ># Should move to VMS installperl when we get one > >I'm not sure what it means to build Perl without sources... I'm guessing that means to build a new Perl with an additional extension statically linked into it. Isn't it possible to do that without the core C code? >but AFAIK >VMS is now using installperl. Is the Big List Of .h Files To Copy To >A Temp Core logic below it obsolete? Isn't the "temp core" where installperl gets them from? -- ________________________________________ Craig A. Berry mailto:craigberry@mac.com "... getting out of a sonnet is much more difficult than getting in." Brad LeithauserThread Previous | Thread Next