On 3/19/14 6:26 PM, Ludovic Courtès wrote: > Hi, > > Readline 6.3’s configure fails when cross-compiling: > > --8<---------------cut here---------------start------------->8--- > [...] > > checking whether mbrtowc and mbstate_t are properly declared... yes > checking for iswlower... yes > checking for iswupper... yes > checking for towlower... yes > checking for towupper... yes > checking for iswctype... yes > checking for nl_langinfo and CODESET... yes > checking for wchar_t in wchar.h... yes > checking for wctype_t in wctype.h... yes > checking for wint_t in wctype.h... yes > checking for wcwidth broken with unicode combining characters... configure: > error: in `/tmp/nix-build-readline-6.3.drv-0/readline-6.3': > configure: error: cannot run test program while cross compiling > See `config.log' for more details > --8<---------------cut here---------------end--------------->8--- > > I fixed it in Guix by passing bash_cv_wcwidth_broken=no: > > > http://git.savannah.gnu.org/cgit/guix.git/commit/?id=f6d47eb6d5e046abc8a7a7dc0863c5661197ef0d > > A fix might be to conservatively set bash_cv_wcwidth_broken=yes in the > fourth arm of AC_TRY_RUN (or better, AC_RUN_IFELSE.)
I can't see how changing the default from `no' to `yes' is going to make a difference. There is one thing that came up a few weeks back. There is a typo in the definition of the wcwidth check: http://lists.gnu.org/archive/html/bug-readline/2014-02/msg00010.html That might make a difference, though I'm not sure how. Chet -- ``The lyf so short, the craft so long to lerne.'' - Chaucer ``Ars longa, vita brevis'' - Hippocrates Chet Ramey, ITS, CWRU [email protected] http://cnswww.cns.cwru.edu/~chet/ _______________________________________________ Bug-readline mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-readline
