On Sun Aug 04 18:41:43 2013, PEVANS wrote:
Show quoted text> Ahhah. That'll be the C compiler defaulting "char" to unsigned rather
> than signed behaviour, as most others do. Surprised this hasn't come
> up before in fact.
>
> I guess the test.h file should be more specific about it. The attached
> patch should fix it.
It does fix it, thanks for being so quick!
Here's the relevant portion of the build log on an ARM builder:
+ ./Build test
t/00use.t ........................ ok
t/01constant.t ................... ok
t/02structure-numeric.t .......... ok
t/03structure-strarray.t ......... ok
t/10structure-with-tail.t ........ ok
t/11structure-args-hashref.t ..... ok
t/12structure-no-length-check.t .. ok
t/99pod.t ........................ skipped: Test::Pod 1.00 required for testing POD
All tests successful.
Files=8, Tests=33, 9 wallclock secs ( 0.12 usr 0.04 sys + 7.13 cusr 1.05 csys = 8.34 CPU)
Result: PASS
Will there be a release soon with this change included? Otherwise I can just include the patch in the Fedora package for now, if you prefer. :)