[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: m4-1.4.11 strtod test failure

From: Eric Blake
Subject: Re: m4-1.4.11 strtod test failure
Date: Thu, 03 Apr 2008 19:20:22 -0600
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv: Gecko/20080213 Thunderbird/ Mnenhy/

Hash: SHA1

According to Bob Friesenhahn on 4/3/2008 10:15 AM:
| Under Solaris 10U4 x86, m4's strtod test blows up.  Hopefully the code
| does not now rely on Linux glibc's strtod irritatingly "wrong" behavior
| related to supposed "hex" values.  I see that the lines which fail
| include a comment which lists all of the known operating systems in the
| universe except for Linux. ;-)

Thanks for the report.  Yes, that was a risk I took in changing strtod
right before the release.  And no, Linux is not perfect (although it comes
closest) - most released versions of glibc have at least one known bug
that wasn't fixed in CVS until last week.  Cygwin came a close second at
having the fewest strtod failures.

| With these failures is m4 still ok for most uses or should I avoid
| installing it?

m4 1.4.11 is probably perfectly fine to use - strtod is used in only one
place, when parsing floating point arguments corresponding to %a, %e, %f,
and %g format specifiers in the `format' builtin.  Not very common in the
first place (for that matter, format is a GNU extension that has not been
picked up by BSD m4).  What the failure probably means is that while your
system strtod had problems, the gnulib replacement also has problems, and
it comes down to a judgment call of which bugs are worse.

For that matter, can you please post the relevant portion of config.log,
where we were investigating whether strtod works?  Maybe the problem is
that strtod passes all the tests at configure time, meaning that strtod.m4
is not strict enough at filtering out bugs.  Since all of your failures
are with hex-floats, and gnulib's version handles those, I'm guessing that
configure guessed (wrongly) that your strtod is working.  Therefore, if
you want to reconfigure with 'gl_cv_func_strtod_works=no ./configure', you
will probably see an improvement in testsuite performance.

| PASS: test-strstr
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:738: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:739: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:748: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:749: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:758: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:759: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:768: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:769: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:778: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:779: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:788: assertion failed
| /home/bfriesen/src/gnu/m4-1.4.11/tests/test-strtod.c:789: assertion failed
| FAIL: test-strtod

Please also run 'make -k' so that you can run the rest of the M4 testsuite.

- --
Don't work too hard, make some time for fun as well!

Eric Blake             address@hidden
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


reply via email to

[Prev in Thread] Current Thread [Next in Thread]