On Tuesday 17 May 2011 08:29:20 Jason wrote: > On Friday 15 April 2011 08:37:07 Peter Rosin wrote: > > (Sorry to revive such an old thread) > > > > A couple of months ago Jason wrote: > > > Hi > > > > > > I noticed this a few weeks ago and I assumed I had somehow corrupted > > > my cygwin installation , so I have now wiped it , and reinstalled from > > > scratch and mpir , and I still get the same errors > > > > > > > > > /bin/sh ../libtool --tag=CC --mode=compile gcc -std=gnu99 > > > -DHAVE_CONFIG_H -I. -I../../mpn -I.. -D__GMP_WITHIN_GMP -I../.. > > > -DOPERATION_`echo fib2_ui | sed 's/_$//'` -m32 -O2 > > > -fomit-frame-pointer -mtune=nocona -march=nocona -c -o fib2_ui.lo > > > fib2_ui.c gcc -std=gnu99 -DHAVE_CONFIG_H -I. -I../../mpn -I.. > > > -D__GMP_WITHIN_GMP -I../.. -DOPERATION_fib2_ui -m32 -O2 > > > -fomit-frame-pointer -mtune=nocona -march=nocona -c fib2_ui.c -o > > > fib2_ui.o In file included from fib2_ui.c:26: > > > /usr/include/stdio.h:53: error: parse error before "fpos_t" > > > In file included from fib2_ui.c:26: > > > /usr/include/stdio.h:210: error: parse error before "fpos_t" > > > /usr/include/stdio.h:216: error: parse error before '*' token > > > /usr/include/stdio.h:379: error: parse error before "fpos_t" > > > /usr/include/stdio.h:380: error: parse error before '*' token > > > /usr/include/stdio.h:535: error: parse error before "fpos_t" > > > /usr/include/stdio.h:535: error: parse error before "fpos_t" > > > /usr/include/stdio.h:535: error: `funopen' declared as function > > > returning a function /usr/include/stdio.h:540: error: parse error > > > before "fpos_t" /usr/include/stdio.h:540: error: parse error before > > > "fpos_t" > > > /usr/include/stdio.h:540: error: `_funopen_r' declared as function > > > returning a function make[2]: *** [fib2_ui.lo] Error 1 > > > make[2]: Leaving directory `/home/root/mpir/mpir-2.1/cygwin/mpn' > > > make[1]: *** [all-recursive] Error 1 > > > make[1]: Leaving directory `/home/root/mpir/mpir-2.1/cygwin' > > > make: *** [all] Error 2 > > > > > > > > > Now stdio.h is the "first line" in the file fib2_ui.c so it appears > > > it's nothing to do with us , this also happens with mpir-1.3 and > > > mpir-2.1 . A lot of these stdio.h are used to get NULL , we can use > > > 0 , or are only needed for debugging , so we dont need them , but > > > there are a few places where we do need them ie mpn_inp.c for the > > > FILE type for input/output. > > > The simple Hello World program compiles and runs OK , so it's possible > > > that we are doing something wrong and cygwin is the first to bite > > > back. > > > > > > > > > Can someone confirm this , in case it's some aspect of my system. > > > I'll try some other lib's that have the stdio.h in , to see if they > > > have this problem. > > > > While I haven't confirm it, I have suffered from a very similar failure, > > and came to the conclusion in [1]. Hope that helps! > > > > Cheers, > > Peter > > > > [1] http://cygwin.com/ml/cygwin/2011-04/msg00167.html > > Hi > > yes your right , it happens exactly the way you say if I try to build mpir > in a subdirectory of the source of mpir , if that subdirectory is called > cygwin then it fails with the above error , but change the name to > something else and everything is OK . This clearly a problem with cygwin , > I'll forward this to the cygwin dev's > > Thanks > Jason
Actually it's probably autotools that has the problem , in MinGW64 the include paths are wrong for subprojects , ie building speed.exe in the tune subdirectory of the MPIR source tree Jason -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com. To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en.