Aaron wrote: > I've dug around on SF.net, but can't find any old autogen > tarballs. I'm specifically looking for 5.9.2.
If you know the release id number, you can get at it by editing the URL for the current release, replacing the ID. I mark old releases as "hidden" so as not to have a cluttered download page. You can also go to ftp://ftp.gnu.org/gnu/autogen and pull whatever you want. > Side note: is there an ETA for 5.9.5? RSN, now that I'm back from my vacation. On Tue, May 6, 2008 at 8:33 AM, Alain Guibert <[EMAIL PROTECTED]> wrote: > Hello, > > On my old Linux box (Intel Pentium 200 MMX, Debian Bo 1.3.1, > kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33, GNU ld cygnus-2.7.1, > libiconv 1.12, pth 2.0.7, gettext 0.17, i586-pc-linux-gnulibc1, > autoconf 2.62, fileutils 3.16, Make 3.75, bash 2.0.0(1)), gmp 4.2.2, > and guile 1.8.4, while building AutoGen 5.9.4, make fails: > > | checking for inttypes.h... no > | checking for stdint.h... no If neither of these are available, you may not have a development system installed. This is a development tool, so development sources should probably be installed. Nevertheless, I am supposed to have guards around including optional headers, though there may be a little oversight somewhere: > | [...] > | gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../autoopts > -DPKGDATADIR=\"/usr/local/share/autogen\" -g -O2 -c libopts.c > -Wp,-MD,.deps/libopts_la-libopts.TPlo -fPIC -DPIC -o > .libs/libopts_la-libopts.o [...] > Commenting-out the #include <stdint.h> line seems sufficient to go > further. However later make segfaults: > > | make[4]: Entering directory `/tmp/autogen-5.9.4/getdefs' > | top_builddir=.. top_srcdir=.. PATH=`cd ../columns >/dev/null && pwd`:$PATH > ; export top_builddir top_srcdir PATH ; /tmp/autogen-5.9.4/agen5/autogen > -L../autoopts ./opts.def > | /bin/sh: mktemp: command not found > | AutoGen aborting on signal 11 (Segmentation fault) in state DONE > | processing template ä§&@ä§&@pts/options.tpl > | on line 302 > | for function Text (18) My usages of "mktemp" are supposed to all have a fallback of using "$$" in some temporary name. I'll check into it as soon as I can get to it. The seg fault should not be a consequence of the mktemp failure. > | make[4]: *** [opts.h] Error 134 > | make[4]: Leaving directory `/tmp/autogen-5.9.4/getdefs' > | make[3]: *** [/tmp/autogen-5.9.4/getdefs/getdefs] Error 2 > | make[3]: Leaving directory `/tmp/autogen-5.9.4/agen5' > | make[2]: *** [all-recursive] Error 1 > | make[2]: Leaving directory `/tmp/autogen-5.9.4/agen5' > | make[1]: *** [all-recursive] Error 1 > | make[1]: Leaving directory `/tmp/autogen-5.9.4' > | make: *** [all] Error 2 > > Note the strange characters in template name. Also strange is that gdb > on the produced core file says sig 6: autogen traps all the signals it can in order to do diagnostics. A template might contain, for example: (sprintf "%s" (compute-a-bad-string-address)) triggering a seg fault. That gets trapped and the trap handler emits the error message, indicating which signal and where we were in processing. I think that when in state "DONE", the template name may be invalid. For next release, I'll look at that code and be sure that either the name is valid or the "processing template" line is not emitted. (I vaguely remember from several years ago that it is supposed to be valid.) > Filename corruption made me think to stop compiler optimizations, > preseting "CFLAGS=-O0 ./configure": Same segfault at the same stage. I do not see this for myself (of course). What is your platform triplet (config.guess)? Would you be willing to build autogen with CFLAGS='-g -O0' and LDFLAGS=-static, then run this ``../agen5/autogen -L../autoopts ./opts.def'' under gdb? GDB should capture the segfault and yield an accurate stack trace before siglongjmp() gets called. Thank you!! Regards, Bruce ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ Autogen-users mailing list Autogen-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/autogen-users
