tag:blogger.com,1999:blog-17885055.post112947619653117498..comments2010-08-01T12:04:39.723+02:00Julio Merinohttp://www.blogger.com/profile/08233346614736256024noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-17885055.post-1129718422742446122005-10-19T12:40:00.000+02:002005-10-19T12:40:00.000+02:00[Originally posted at 2005-03-14 11:16 pm UTC]<BR/><BR/>well for the variant part, I suppose you have to run autoconf to find out but it seems bizarre to me to run a compilation phase to test for the existence of routines when a .pc file type approach can front-load that during install of the library.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-17885055.post-1129718396774287662005-10-19T12:39:00.001+02:002005-10-19T12:39:00.001+02:00[Originally posted at 2005-03-14 08:00 pm UTC]<BR/><BR/>About the openssl thing, you are right (I've suffered it too). I think I'll post something about that.<BR/><BR/>But what do you mean with "running autoconf to find mostly static values"? When you run autoconf, all you need is the pkgconfig.m4 file to be able to use the PKG_CHECK_MODULES macro. Everything else about libraries is searched for while running the configure script, and that data is not "static". (I think I misunderstood you.)Julio M. Merino Vidalhttp://www.blogger.com/profile/08233346614736256024noreply@blogger.comtag:blogger.com,1999:blog-17885055.post-1129718370341941842005-10-19T12:39:00.000+02:002005-10-19T12:39:00.000+02:00[Originally posted at 2005-03-14 03:58 am UTC]<BR/><BR/>The mixed state, some do & some don't is bad. YMMV but I found openssl expectations of a .pc in some pkg (gaim-encryption) bit me.<BR/><BR/>I've often felt the cost of re-re-re-re-re-running autoconf to find mostly<BR/>static values is insane, so I too am temprementally accepting of .pc files to<BR/>encode what a given package believed was valid machine state. But I also see why some people dislike this: its better to ask, than assume.<BR/><BR/>-GeorgeAnonymousnoreply@blogger.com