Re[2]: [lmi] 'build' and 'host' in wx configury [Was: To configure wx-2.6.2]

[Vadim Zeitlin]

On Wed, 11 Feb 2009 02:56:35 +0000 Greg Chicares <address@hidden> wrote:

GC> On 2009-02-07 13:52Z, Vadim Zeitlin wrote:
GC> > On Sat, 07 Feb 2009 11:24:11 +0000 Greg Chicares <address@hidden> wrote:
GC> [...]
GC> > GC> > Options to 'configure' are given here
GC> > GC> >   http://cvs.savannah.gnu.org/viewvc/lmi/lmi/install_wx.make
GC> > GC> > and seem generally unremarkable.
GC> > 
GC> >  Why use --enable-commondlg which is on by default there though?
GC> 
GC> I don't remember. Perhaps it wasn't on by default in the distant past?

 I'm quite sure it never was. In any case, it is safe to remove it now.

GC> > And I'd consider switching to the official mingw32-cross-compiler as
GC> > soon as it is released.
GC> 
GC> That's probably best. We'll have to test it first, of course.

 Yes, certainly. The only problem with cross-compiling is that it
disables the run-time tests in configure which is a pity because it could
work in this particular case as mingw32 executables can run under Cygwin. 
Unfortunately I don't think we can do anything about this, this problem
should really be solved at autoconf level somehow (e.g. by allowing to
specify that we can execute the cross-compiled binaries in some way).


GC> [0] "how 'configure' output changes with a new '--build' setting"
GC> 
GC> Here's the result of diffing two build logs
GC>   diff -U0 log-20080911T1708Z log-20090207T0117Z
GC> snipped down the the wx configury part. I snipped a few lines that
GC> differed only in their line-endings.
GC> 
GC> -make[1]: Entering directory `/opt/lmi/wx-scratch/wxWidgets-2.8.7/gcc344'
GC> +make[1]: Entering directory `/opt/lmi/wx-scratch/wxWidgets-2.8.9/gcc344'
GC> -../configure --disable-gif --disable-apple_ieee --without-libjpeg 
--without-libtiff --without-regex --enable-monolithic --enable-shared 
--disable-threads --enable-stl --enable-std_iostreams
GC> --enable-commondlg --disable-compat24 --disable-vararg_macros 
--prefix=/opt/lmi/local VENDOR='344' CPPFLAGS='-DNO_GCC_PRAGMA' 
CC='/MinGW_/bin/gcc' CXX='/MinGW_/bin/g++'  && make && make install
GC> +../configure --build=i686-pc-cygwin --host=i686-pc-mingw32 --disable-gif 
--disable-apple_ieee --without-libjpeg --without-libtiff --without-regex 
--enable-monolithic --enable-shared --disable-threads
GC> --enable-stl --enable-std_iostreams --enable-commondlg --disable-compat24 
--disable-vararg_macros --prefix=/opt/lmi/local VENDOR='344' 
CPPFLAGS='-DNO_GCC_PRAGMA' CC='/MinGW_/bin/gcc'
GC> CXX='/MinGW_/bin/g++'  && make && make install
GC> @@ -4332,2 +2616,2 @@
GC> -checking host system type... i686-pc-cygwin
GC> -checking target system type... i686-pc-cygwin
GC> +checking host system type... i686-pc-mingw32
GC> +checking target system type... i686-pc-mingw32
GC> @@ -4554 +2838 @@
GC> -checking for gcc... /MinGW_/bin/gcc
GC> +checking for i686-pc-mingw32-gcc... /MinGW_/bin/gcc
GC> @@ -4557 +2841 @@
GC> -checking whether we are cross compiling... no
GC> +checking whether we are cross compiling... yes
GC> @@ -4566,0 +2851 @@
GC> +checking for i686-pc-mingw32-g++... /MinGW_/bin/g++
GC> @@ -4569,0 +2855 @@
GC> +checking for i686-pc-mingw32-ranlib... no
GC> @@ -4570,0 +2857 @@
GC> +checking for i686-pc-mingw32-ar... no
GC> @@ -4572,0 +2860 @@
GC> +checking for i686-pc-mingw32-strip... no
GC> @@ -4573,0 +2862 @@
GC> +checking for i686-pc-mingw32-windres... no
GC> @@ -4574,0 +2864 @@
GC> +checking for i686-pc-mingw32-dlltool... no

 These should go away if RANLIB, AR &c are specified on the command line.

GC> @@ -4631,0 +2922,5 @@
GC> +checking for unordered_map... no
GC> +checking for tr1/unordered_map... no
GC> +checking for hash_map... (cached) no
GC> +checking for ext/hash_map... (cached) yes
GC> +checking for GNU hash_map and hash_set... (cached) yes

 Strange, I don't understand where does this one come from.

GC> @@ -4637,2 +2932,2 @@
GC> -checking for zlib.h >= 1.1.4... no
GC> -checking for zlib.h... (cached) no
GC> +checking for zlib.h >= 1.1.4...
GC> +checking for zlib.h... no
GC> @@ -4640,2 +2935,2 @@
GC> -checking for png.h > 0.90... no
GC> -checking for png.h... (cached) no
GC> +checking for png.h > 0.90...
GC> +checking for png.h... no

 Missing "no"s here are strange too but shouldn't matter.

GC> @@ -4667 +2962,3 @@
GC> -checking if snprintf supports positional arguments... yes
GC> +checking if snprintf supports positional arguments... configure: WARNING: 
Assuming Unix98 printf() is not available,
GC> +define HAVE_UNIX98_PRINTF as 1 in setup.h if it is available.
GC> +no

 This one is probably ok as I think mingw32 printf() -- being a Microsoft
version -- doesn't support Unix98 features. But it would be still better if
we could run the test instead of just assuming that it doesn't work.

GC> @@ -4688 +2985 @@
GC> -checking for ranlib... (cached) ranlib
GC> +checking for i686-pc-mingw32-ranlib... ranlib
GC> @@ -4692,2 +2989,3 @@
GC> -checking for ar... (cached) ar
GC> -checking for strip... (cached) strip
GC> +checking for i686-pc-mingw32-ar... ar
GC> +checking for i686-pc-mingw32-strip... strip
GC> +checking for i686-pc-mingw32-nm... no
GC> @@ -4694,0 +2993 @@
GC> +checking for command to install directories... /usr/bin/install -c -d
GC> @@ -4697 +2996,2 @@
GC> -checking for windres... (cached) windres
GC> +checking for i686-pc-mingw32-windres... windres
GC> +checking if the compiler supports precompiled headers... yes

 The last one is probably because we disable PCH under Cygwin (because we
assume that the compiler is the Cygwin one, which has known bugs in its PCH
support) so it also means that you should see significant build time
decrease in the second case...

 Regards,
VZ