Bug 70375

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020723

Description of problem:
The glibc 2.2.5 shipped with Red Hat 7.3 doesn't seem to be based on the
official glibc 2.2.5: An md5 sum of the glibc-2.2.5 tar-ball in the source RPM
for glibc-2.2.5-37 doesn't match the official glibc-2.2.5.tar.bz2. This is
strange enough by itself, I think.

What's worse: The mktime() function's behaviour has changed due to
http://bugs.gnu.org/cgi-bin/gnatsweb.pl?debug=&database=default&cmd=view+audit-trail&cmd=view&pr=2738
This results in odd, hard-to-trace bugs in different software packages.
PostgreSQL is affected (and the PostgreSQL developers have stated that they will
_not_ try to adjust to the new mktime() behaviour), as are PHP applications
developed where I work.

The glibc developers seem to have changed POSIX interpretation. That may be
correct, but the changed behaviour isn't part of the official glibc 2.2.5. I
believe that it's a serious mistake to incorporate non-official changes like
this into a point-release of Red Hat. Incorporating it into Red Hat 8.0 might be
OK (although it will probably still be cursed at by many developers).

Please release a new set of glibc packages without the unofficial, changed
mktime() behaviour. Otherwise, Red Hat 7.3 will not be compatible with previous
Red Hat releases, and it will not be compatible with other Linux distributions.

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. Example: With php's mktime() function, this will yield different results on
Red Hat 7.3 versus older Red Hat releases:
2. print mktime(1,1,1,1,1,1969);
	

Actual Results:  On Red Hat 7.3: -1

Expected Results:  On older Red Hat releases:
-31535939

Additional info: