I just saw today that there are "x86_64-mingw32" targets, so it should
actually test for "x86_64-*" for size_t of size 8 and replace
"*mingw32*" with "*86-*". I only have access to an i686 to test this.
Thank you
William A. Rowe, Jr. wrote:
Will your proposed changes interfere with mingw x86_64 support?
Thanks for your proposed patch
Yours,
Bill
Dzonatas wrote:
Is this acceptable to default the cache variables to simplify the needed
hints?
This changes the "mingw" case in configure.in:
--- ../configure.in 2006-09-08 16:41:45.000000000 +0000
+++ configure.in 2007-03-23 23:28:49.000000000 +0000
@@ -411,6 +411,17 @@
eolstr="\\n"
proc_mutex_is_global=1
OBJECTS_PLATFORM='$(OBJECTS_win32)'
+ if test "$cross_compiling" = yes; then
+ ac_cv_file__dev_zero=no
+ ac_cv_func_setpgrp_void=no
+ apr_cv_tcp_nodelay_with_cork=no
+ case $host in
+ *mingw32*)
+ ac_cv_sizeof_ssize_t=4
+ ac_cv_sizeof_size_t=4
+ ;;
+ esac
+ fi
;;
*cygwin*)
OSDIR="unix"
--
