From: Eric B. <win...@gm...> - 2005-05-26 15:09:17
|
building from cvs on mac os x 10.4, and it builds ok and runs make check fine (up until the ansi tests--behind a proxy so it can't download the tests). however, when I start, clisp will display the banner, and immediately say "Bye." and exit. there is a problem compiling with readline, so I used the following config: ./configure --without-dynamic-ffi --without-readline --with-debug -- eblood |
From: Sam S. <sd...@gn...> - 2005-05-26 15:32:01
|
> * Eric Blood <jvaxljbbfgre@tznvy.pbz> [2005-05-26 09:09:07 -0600]: > > building from cvs on mac os x 10.4, and it builds ok and runs make > check fine (up until the ansi tests--behind a proxy so it can't > download the tests). however, when I start, clisp will display the > banner, and immediately say "Bye." and exit. > > there is a problem compiling with readline, so I used the following config: > > ./configure --without-dynamic-ffi --without-readline --with-debug It has been reported as a poll() problem. $ ./configure --without-dynamic-ffi --without-readline --with-debug build-dir $ cd build-dir $ vi unixconf.h comment out "#define HAVE_POLL 1" proceed to make. -- Sam Steingold (http://www.podval.org/~sds) running w2k <http://www.memri.org/> <http://www.camera.org> <http://www.dhimmi.com/> <http://www.honestreporting.com> <http://www.mideasttruth.com/> An elephant is a mouse with an operating system. |
From: Eric B. <win...@gm...> - 2005-05-26 19:23:51
|
On 5/26/05, Sam Steingold <sd...@gn...> wrote: > It has been reported as a poll() problem... Thanks. That did the trick. --=20 eblood |
From: Bruno H. <br...@cl...> - 2005-05-31 19:16:36
|
Eric Blood wrote: > > It has been reported as a poll() problem... > > Thanks. That did the trick. Thanks for the confirmation. poll() on MacOS X 10.4 really seems broken, see http://lists.apple.com/archives/darwin-dev/2005/May/msg00220.html http://comments.gmane.org/gmane.os.apple.fink.gnome/1193 http://cvs.sourceforge.net/viewcvs.py/fink/dists/10.4-transitional/unstable/main/finkinfo/gnome/glib2.info?r1=1.3&r2=1.4 I could just disable clisp's use of poll() on Darwin altogether, but it's better to have an autoconf test. Can some kind soul please run the following test program on MacOS X 10.4.0: $ gcc polltest.c -o polltest $ ./polltest ============================= polltest.c ============================= #include <stdio.h> #include <errno.h> #include <fcntl.h> #include <poll.h> int main () { struct pollfd ufd; int result; { ufd.fd = 0; ufd.events = POLLIN; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } { ufd.fd = 1; ufd.events = POLLOUT; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } { ufd.fd = open("/dev/tty", O_RDONLY); ufd.events = POLLIN; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } { ufd.fd = open("/dev/tty", O_WRONLY); ufd.events = POLLOUT; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } { ufd.fd = open("/dev/null", O_RDONLY); ufd.events = POLLIN; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } { ufd.fd = open("/dev/null", O_WRONLY); ufd.events = POLLOUT; ufd.revents = 0; errno = 0; result = poll(&ufd,1,0); printf("fd = %d: %d/%d %d -> %d\n", ufd.fd, result, errno, ufd.events, ufd.revents); } return 0; } ==================================================================== Bruno |
From: Pascal C. <pc...@p-...> - 2005-05-31 20:22:23
|
On 31 May 2005, at 21:10, Bruno Haible wrote: > Eric Blood wrote: > >>> It has been reported as a poll() problem... >>> >> >> Thanks. That did the trick. >> > > Thanks for the confirmation. poll() on MacOS X 10.4 really seems > broken, see > http://lists.apple.com/archives/darwin-dev/2005/May/msg00220.html > http://comments.gmane.org/gmane.os.apple.fink.gnome/1193 > http://cvs.sourceforge.net/viewcvs.py/fink/dists/10.4-transitional/ > unstable/main/finkinfo/gnome/glib2.info?r1=1.3&r2=1.4 > > I could just disable clisp's use of poll() on Darwin altogether, > but it's > better to have an autoconf test. Can some kind soul please run the > following test program on MacOS X 10.4.0: > > $ gcc polltest.c -o polltest > $ ./polltest fd = 0: 1/0 1 -> 32 fd = 1: 1/0 4 -> 32 fd = 3: 1/0 1 -> 32 fd = 4: 1/0 4 -> 32 fd = 5: 1/0 1 -> 32 fd = 6: 1/0 4 -> 32 Pascal -- 2nd European Lisp and Scheme Workshop July 26 - Glasgow, Scotland - co-located with ECOOP 2005 http://lisp-ecoop05.bknr.net/ |
From: Bruno H. <br...@cl...> - 2005-06-01 11:36:04
|
Pascal Costanza wrote: > > $ gcc polltest.c -o polltest > > $ ./polltest > > fd = 0: 1/0 1 -> 32 > fd = 1: 1/0 4 -> 32 > fd = 3: 1/0 1 -> 32 > fd = 4: 1/0 4 -> 32 > fd = 5: 1/0 1 -> 32 > fd = 6: 1/0 4 -> 32 Thanks for the result. The expected result (as found on Linux and Solaris) is fd = 0: 0/0 1 -> 0 fd = 1: 1/0 4 -> 4 fd = 3: 0/0 1 -> 0 fd = 4: 1/0 4 -> 4 fd = 5: 1/0 1 -> 1 fd = 6: 1/0 4 -> 4 I've added an autoconf test against this to clisp and will also propagate it to gnulib. You can now try to see whether clisp from CVS builds out of the box. Bruno |
From: Pascal C. <pc...@p-...> - 2005-06-01 12:41:07
|
On 1 Jun 2005, at 13:35, Bruno Haible wrote: > Pascal Costanza wrote: > >>> $ gcc polltest.c -o polltest >>> $ ./polltest >>> >> >> fd = 0: 1/0 1 -> 32 >> fd = 1: 1/0 4 -> 32 >> fd = 3: 1/0 1 -> 32 >> fd = 4: 1/0 4 -> 32 >> fd = 5: 1/0 1 -> 32 >> fd = 6: 1/0 4 -> 32 >> > > Thanks for the result. The expected result (as found on Linux and > Solaris) is > > fd = 0: 0/0 1 -> 0 > fd = 1: 1/0 4 -> 4 > fd = 3: 0/0 1 -> 0 > fd = 4: 1/0 4 -> 4 > fd = 5: 1/0 1 -> 1 > fd = 6: 1/0 4 -> 4 > > I've added an autoconf test against this to clisp and will also > propagate it > to gnulib. > > You can now try to see whether clisp from CVS builds out of the box. No, it doesn't, but seemingly for unrelated reasons: spvw.d: In function 'getSP': spvw.d:642: warning: function returns address of local variable In file included from spvw.d:666: spvw_language.d: In function 'init_language': spvw_language.d:235: error: label at end of compound statement make: *** [spvw.o] Error 1 Pascal -- 2nd European Lisp and Scheme Workshop July 26 - Glasgow, Scotland - co-located with ECOOP 2005 http://lisp-ecoop05.bknr.net/ |
From: Bruno H. <br...@cl...> - 2005-06-01 12:55:20
|
Pascal Costanza wrote: > spvw_language.d: In function 'init_language': > spvw_language.d:235: error: label at end of compound statement > make: *** [spvw.o] Error 1 Oops. Fixed. And now? Bruno |
From: Pascal C. <pc...@p-...> - 2005-06-01 13:54:58
|
On 1 Jun 2005, at 14:54, Bruno Haible wrote: > Pascal Costanza wrote: > >> spvw_language.d: In function 'init_language': >> spvw_language.d:235: error: label at end of compound statement >> make: *** [spvw.o] Error 1 >> > > Oops. Fixed. And now? Yes, the current cvs version builds a working clisp out of the box now. Thanks a lot! Pascal -- 2nd European Lisp and Scheme Workshop July 26 - Glasgow, Scotland - co-located with ECOOP 2005 http://lisp-ecoop05.bknr.net/ |