On II.19.2004, I configured, build, tested and installed autoconf-2.59
without any apparent problems. The make check said that 175 tests were
successful and 8 tests were skipped.
When I try to upgrade to autoconf-2.61, the make check now gives
problems. But first the 'configure' step (I took the defaults) and the
'make' step appear to execute successfully. The 'make check' ran
through the first 91 tests with either an 'ok' or (very occasionally)
that the test was skipped. However test 92 AC_CONFIG_FILES, HEADERS,
LINKS, and COMMANDS in the testing config.status section just hangs the
computer and Ctrl-C does not stop it, so finally I had to give
Ctrl-Alt-Delete and reboot. After some work, I discovered the section
in ./tests/torture.at which initiated this test. I used EMACS to delete
lines 130-209 of this file and re-ran make check which now continues
(with the original numbering above 91 one less) through test 221 with
again either an 'ok' or the test (again, very occasionally) was
skipped. Unfortunately, test 220 (in the revised numbering) AC_PROG_SED
also causes a hang and again I had to reboot.
I am on SED-4.1.5 which is the latest version of SED that I could find
on the GNU archive. I also am using GCC-4.1.1 but I don't think it was
used in the build step.
One other point: The build directory is a subdirectory of /Aux/Install
which is on partition /dev/hdb4 while directory / is on partition
/dev/hda4. Also, I am running the configure, build and make check as
root. I mention these things since I build quite a bit of LINUX
software and very occasionally there have been problems in the past. I
configured, built and tested GCC-4.1.1 (and then installed it) without
any problems that I could see. The build took 5 hours and 36 minutes
and the testing took 5 hours and 41 minutes. Also, when I tested
SED-4.1.5, I got a message that all 75 tests were passed.
Recently I had a problem with gettext-0.61.1 and I contacted Bruno
Haible with attachments of what had happened as well as other files
which I thought might be useful. He responded very quickly with
requests for more information which I sent back and he was able to
diagnose what was happening and he had prepared a fix for the next
version of gettext. Meanwhile, I, based on comments he made, was able
to figure out what I needed to do to get the error in the build to go
away, and successfully tested and installed it. I mention this since I
wondered if perhaps I should have sent this e-mail to
[EMAIL PROTECTED] Also, if any files from the autoconf-2.61 build
directory would be useful, I would be happy to send them.
With this kind of problem in the testing, I have been very reluctant to
go ahead and install autoconf-2.61 although I note that the configure
file in GLIBC-2.5 was generated with 2.61.
Any help you could give me would be appreciated.
Tom Cundiff
