Re: mktime() problem when building print/a2ps

[Thomas Klausner <wiz%NetBSD.org@localhost>]

On Thu, Oct 13, 2022 at 07:10:46PM -0700, Paul Goyette wrote:
> Strange problem...
> 
> Host and target are both an AMD64 NetBSD 9.99.99 build from sources
> dated 2022-09-09 at 01:58:58 UTC.  Thepackages are all being built
> in a chroot sandbox (as created with sysutils/mksandbox).  And my
> pkgsrc itself is from today 2022-10-13 at 19:29:24 UTC.
> 
> All predecessor/dependency packages have been rebuilt from today's
> source and installed in the sandbox.  When I get to print/a2ps, it
> starts to run, and fails in configure.  The last thing in the log
> file is
> 
> 	...
> 	checking limits.h usability... yes
> 	checking limits.h presence... yes
> 	checking for limits.h... yes
> 	checking for working mktime...
> 
> and it never progresses beyond this point.  It appears to be in a
> processor-bound loop making this check, and xosview reports that one
> cpu (6% of a total of 16threads) is completely busy.  top(1) reports
> that a process called ``conftest'' is running with 98.97% of one
> thread.
> 
> I don't know what conftest is, but this sure appears broken!
> 
> Any suggestions?

This takes much longer than I'd expect it to, but it finishes for me
in less than a minute.

Extracting the code for the test from the configure script and
compiling it works fine and finishes quickly (without an conftest.h).

Not sure what's wrong here.
 Thomas