On 20 February 2014 00:17, Karl Williamson <pub...@khwilliamson.com> wrote: > On 02/19/2014 05:07 PM, Steve Hay wrote: >>> >>> Ah... I now have run/locale.t passing after doing a clean and fresh >>> >build. I only rebuilt my existing build before, trusting that the >>> >Makefile would rebuild the necessary, but it seems not :-/ So this >>> >could be a false alarm. >>> > >>> >I've just set a full test suite running and will report back in >>> >15-20min (whatever it takes) to confirm if all is well. >> >> Sorry - that was slower than I expected! >> >> t/run/locale.t does indeed pass all tests now, but now I have >> lib/locale.t failing 6 tests (otherwise everything else has passed): > > > A patch to remove those tests is about to be pushed. (I forgot that they > are only valid in a UTF-8 locale, and that isn't necessarily at the time.)
Thanks, all tests are passing in blead now (baca455461522d62967cae39a044a49583cb6aae). > > But it is really strange that things didn't work after the && fix. You > shouldn't have been able to run a perl until the syntax error in the .c was > fixed and perl was compiled. I changed the "and" to "&&" myself and built that to see what happened, and then simply rebuilt that build when you applied your fixes. Something seems to have gone wrong in that rebuild process, presumably a gremlin in the win32/Makefile.