The following reply was made to PR toolchain/36159; it has been noted by GNATS.

From: Hiroyuki Bessho <bsh@grotto.jp>
To: gnats-bugs@NetBSD.org
Cc: tron@zhadum.org.uk
Subject: Re: toolchain/36159
Date: Wed, 18 Apr 2007 17:23:15 +0900

 > This is not the correct fix. "src/tools/compat/configure" checks for
 > the existence of "u_int32_t". Can you please check "config.log" to
 > find out what the result of that check was?
  
 As uwe@ wrote,  It's not that Cygwin doesn't have u_int32_t defined.
 
 config.log says:
 ----------------------------------------------------------------
 configure:5394: checking for uint32_t
 configure:5418: cc -c -O  conftest.c >&5
 configure:5424: $? = 0
 configure:5428: test -z 
 			 || test ! -s conftest.err
 configure:5431: $? = 0
 configure:5434: test -s conftest.o
 configure:5437: $? = 0
 configure:5448: result: yes
 configure:5525: checking for u_int32_t
 configure:5549: cc -c -O  conftest.c >&5
 configure:5555: $? = 0
 configure:5559: test -z 
 			 || test ! -s conftest.err
 configure:5562: $? = 0
 configure:5565: test -s conftest.o
 configure:5568: $? = 0
 configure:5579: result: yes
 
 ...
 
 ac_cv_type_u_char=yes
 ac_cv_type_u_int=yes
 ac_cv_type_u_int16_t=yes
 ac_cv_type_u_int32_t=yes
 ac_cv_type_u_int64_t=yes
 ac_cv_type_u_int8_t=yes
 ac_cv_type_u_long=yes
 ac_cv_type_u_quad_t=no
 ac_cv_type_u_short=yes
 ac_cv_type_uint16_t=yes
 ac_cv_type_uint32_t=yes
 ac_cv_type_uint64_t=yes
 ac_cv_type_uint8_t=yes
 ac_cv_working_alloca_h=yes
 
 ----------------------------------------------------------------
 
 and "cc -E /usr/include/stdio.h | egrep 'u_?int32'" on Cygwin gives you:
 
 typedef unsigned int __uint32_t;
 typedef __uint32_t __uint_least32_t;
 typedef unsigned long uint32_t;
 typedef __uint32_t u_int32_t;
 
 
 Sincerely,
 --
 bsh@NetBSD.org