At Thu, 14 Apr 2022 00:09:16 +0200, Tobias Nygren <tnn%NetBSD.org@localhost> wrote: Subject: Re: lang/guile30 violates basic pkgsrc mechanisms to work around an ancient bug > > One thing to consider is how the base toolchain on NetBSD is configured. > > $ /usr/bin/gcc -v 2>&1 | grep Target > Target: x86_64--netbsd > > The triplet used by pkgsrc should agree with the base system compiler. > It is conceivable that there exists software that becomes unhappy otherwise. That's an interesting point, but I don't actually see any connection. If someone could point to some existing lines of code in any package using GNU Configure that would potentially be affected.... I would also argue the base toolchain is incorrectly configured, at least by GNU Standards. :-) But that's a somewhat different topic. Also it didn't always used to be. Sadly the change which made this happen (bsd.own.mk:1.172) was incompletely or incorrectly commented (and was also made by the same person who introduced the empty value for LOWER_VENDOR in pkgsrc). (Fixing this in the toolchain may affect the finger memory for people who type the long name of toolchain commands.) In any case the point remains that software configured outside of pkgsrc, and without specifying '--host=', is going to be configured with the canonical output of "config.guess". That's always expected to work, and always has worked, and it would indeed be a serious bug if it suddenly failed to work in any way. And, Since 1997 or before "config.guess" has put the string "unknown" in the vendor part for NetBSD hosts. So I would say it is actually IN-conceivable "that there exists software that becomes unhappy" with a vendor name of "unknown" instead of "". -- Greg A. Woods <gwoods%acm.org@localhost> Kelowna, BC +1 250 762-7675 RoboHack <woods%robohack.ca@localhost> Planix, Inc. <woods%planix.com@localhost> Avoncote Farms <woods%avoncote.ca@localhost>
Attachment:
pgpSShK1rpagU.pgp
Description: OpenPGP Digital Signature