New GNUstep make puts ffcall libraries in Libraries/lib
Hi, I just started to try to use the new GNUstep make 2.0 to build GNUstep under Windows again. I use the default filesystem layout (gnustep) and I ./configure'd make with a prefix. I am using MinGW. The only thing I changed from a working chaing of scripted commands to build GNUstep base is to replace GNUstep make with 2.0. Now, what happens is that ./configure of base complains about not having ffcall installed. Prior to configuring base I actually did make install ffcall, but this seems to have put the libraries in System/lib instead of System/Library/Libraries as it did previously. I expect this to be the reason configure of base fails. Is putting the libraries in System/lib the expected behaviour? Should I add this directory to LD_LIBRAY_PATH? Or did I miss some step and should make have put ffcall's binaries in System/Library/Libraries anyway? Regards, Tom Koelman ___ Gnustep-dev mailing list Gnustep-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: New GNUstep make puts ffcall libraries in Libraries/lib
Nicola Pero [EMAIL PROTECTED] writes: Presumably you configure it using ./configure --prefix=/usr/GNUstep/System and you notice that that would install the lib into /usr/GNUstep/System/Library/Libraries with gnustep-make v1 and now it installs the lib into /usr/GNUstep/System/lib. What is likely happening is that we no longer install the config.site file that was telling ffcall's configure to install libs into System/Library/Libraries instead of System/libs. (reason that we don't install config.site is that we are no longer sure about layouts, we don't want to risk overwriting another config.site etc) I personally would install ffcall into /usr/local/lib, but if you really want it installed in gnustep's directories, you probably want to do ./configure --libdir=/usr/GNUstep/System/Library/Libraries --includedir=/usr/GNUstep/System/Library/Headers that should work both with gnustep-make v1 and v2. Thanks, Nicola! Just for the record, this indeed solved my build troubles. I'll upload a new Windows installer soon. Tom ___ Gnustep-dev mailing list Gnustep-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: GNUSTEP_CONFIG_FILE value ('/etc/GNUstep/GNUstep.conf') is not an absolute path
Sheldon Gill [EMAIL PROTECTED] writes: I use the following incantations: In the directory with source of GNUstep make: cd /gnustep-source/make ./configure --prefix=$GNUSTEP_INSTALL_DIR --with-config-file=$GNUSTEP_INSTALL_DIR/GNUstep.conf make install [...] Build ffcall [...] # Build and install the base library itself cd /gnustep-source/base ./configure --disable-xml --with-config-file=$GNUSTEP_INSTALL_DIR/GNUstep.conf make install messages=yes warn=no Now, if I run the defaults.exe program or some other GNUstep using program I get the following: 2006-03-06 14:57:21.708 defaults.exe[3680] GNUSTEP_CONFIG_FILE value ('/etc/GNUstep/GNUstep.conf') is not an absolute path. Please rebuild GNUstep-base specifying a valid path to the config file. What do I do wrong? As far as I can tell I already build GNUstep-base specifying a valid path to the config file. If you have specified a file and it's been picked up correctly you will find your definition in: core/base/Source/config.h I suggest you check that first. My config.h does not contain a definition of GNUSTEP_CONFIG_FILE. It does contain this: /* Disable GNUSTEP_CONFIG_FILE environment variable */ #define OPTION_NO_ENVIRONMENT 1 Are you using the *release* of base 1.11.2 or a later CVS version? Yes, I am using the release. Is there maybe some documentation about the way this configuration file configuration works? Regards, Tom Koelman ___ Gnustep-dev mailing list Gnustep-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnustep-dev
GNUSTEP_CONFIG_FILE value ('/etc/GNUstep/GNUstep.conf') is not an absolute path
Hi, I am fixing the Windows Installer with regard to the GNUstep.conf file. I am building base 1.11.2 and gui 0.10.2. I use the following incantations: In the directory with source of GNUstep make: cd /gnustep-source/make ./configure --prefix=$GNUSTEP_INSTALL_DIR --with-config-file=$GNUSTEP_INSTALL_DIR/GNUstep.conf make install [...] Build ffcall [...] # Build and install the base library itself cd /gnustep-source/base ./configure --disable-xml --with-config-file=$GNUSTEP_INSTALL_DIR/GNUstep.conf make install messages=yes warn=no Now, if I run the defaults.exe program or some other GNUstep using program I get the following: 2006-03-06 14:57:21.708 defaults.exe[3680] GNUSTEP_CONFIG_FILE value ('/etc/GNUstep/GNUstep.conf') is not an absolute path. Please rebuild GNUstep-base specifying a valid path to the config file. What do I do wrong? As far as I can tell I already build GNUstep-base specifying a valid path to the config file. Regards, Tom Koelman ___ Gnustep-dev mailing list Gnustep-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnustep-dev
Re: More Windows stuff ... example gui apps work for me now
Marko Riedel [EMAIL PROTECTED] writes: Then I get the following error while building gnustep-base: GSFFCallInvocation.m: In function `GSInvocationCallback': GSFFCallInvocation.m:1003: error: union has no member named `_longlong' GSFFCallInvocation.m:1004: error: union has no member named `_ulonglong' GSFFCallInvocation.m:1003: warning: value computed is not used GSFFCallInvocation.m:1004: warning: value computed is not used make[2]: *** [shared_obj/GSFFCallInvocation.o] Error 1 make[1]: *** [libgnustep-base.all.library.variables] Error 2 make: *** [internal-all] Error 2 Does anyone know what is happening here? I can reproduce this. What actaully goes wrong is the test for checking whether floats are returned in integer registers in ./configure of ffcall. When I compile the test snippet[1] there manually, gcc 3.4.2 warns me that $ gcc conftest.c conftest.c: In function `main': conftest.c:5: warning: function called through a non-compatible type conftest.c:5: note: if this code is reached, the program will abort and indeed the test aborts. However, it depends on how dr watson on your machine is configured whether you actually notice. On my machine it silently writes a drwatson.log and configure assumes the answer to the question are floats returned in integer registers? is no. On another machine an error dialog appears and conftest.exe is locked while that hasn't been done yet. Nicola, do you have any idea whether this test is needed, and whether gcc or the test is at fault? Regards, Tom Koelman ---[1] - float x = (float)1.2; float y = (float)1.3; float fun () { return x*y; } int main() { int val = (* (int (*) ()) fun) (); exit (!(val == 0x3FC7AE15 || val == 0x15AEC73F)); } ___ Gnustep-dev mailing list Gnustep-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnustep-dev