John Marino <mfl-commissioner%marino.st@localhost> writes: > On 7/22/2012 05:20, David A. Holland wrote: >> >> Modified Files: >> pkgsrc/net/coda: Makefile distinfo >> pkgsrc/net/coda/patches: patch-coda-src_resolution_rename.cc >> >> Log Message: >> Improve previous patch: blindly setting MAXNAMLEN to 255 is dangerous, >> as it is e.g. 511 on NetBSD. Instead, get it from NAME_MAX, which is >> POSIX. PKGREVISION -> 3 as a precaution. >> > > I don't have any comment to the patch itself as it is an > improvement. However, any system other than dragonfly that doesn't > have MAXNAMLEN defined is still going to fail to build* -- the patch > to config.h.in is very gross, but it saved having to patch about half > a dozen files. To really do this in a generic way, we probably need > to remove the config.h.in patch and duplicate the resolution_rename.cc > fix to every file that mentions MAXNAMLEN. Maybe a better way is to > create one header to defined MAXNAMLEN (if necessary) and add that > header to every file that needs it. I think doing major surgery on programs with non-dead upstreams should be the approach of last resort, and goes beyond what pkgsrc should be doing. > John > > * Honestly, any system not specifically supported by coda will > probably fail to build even if MAXNAMLEN is fixed -- there are other > platform specific code that will probably cause a build failure if the > right conditional branch isn't presented. Indeed. In particular coda needs a kernel module. If that isn't known to work -- and all the MAXNAMELEN issues are trivial compared to the kernel module -- all of this is just noise.
Attachment:
pgpNmHwfUKmtn.pgp
Description: PGP signature