Perl 4.036 regression test shows bug in SunOS 4.1.1 tmpfs 
Author Message
 Perl 4.036 regression test shows bug in SunOS 4.1.1 tmpfs

I have been using Perl's regression test in various settings to help confirm
that various peoples' builds of Perl are working right. I started noticing
some annoying, intermittent failures. Turns out that op/stat.t test 4 was
failing, which means that after the sequence:

        create file
        sleep 2
        make a second hard link to the file
        stat, confirm nlinks=2

there was a problem, because mtime == ctime, when ctime should be newer than
mtime. Bummer. When I moved the dir where I ran the tests from /tmp (tmpfs
in our environment) to /usr/tmp (normal local 4.2 filesystem) the problem
went away.

I'm not sure whether this eccentricity is worth programming around. I just
avoided it. If someone wanted to, an easy check is

        df . | grep ^swap

since df(1) reports:

        ; cd /tmp;df .
        Filesystem         1024-blocks  Used Available Capacity Mounted on
        swap                   51992     428    51564      1%   /tmp
        ;

-Bennett



Sun, 29 Sep 1996 22:56:32 GMT  
 
 [ 1 post ] 

 Relevant Pages 

1. Perl 4.036 on SunOS 4.1 fails make test

2. Building perl 4.036 under Sunos 4.1.3 (???)

3. Problems building Perl 4.036 on SunOS 4.1.3

4. Compiling perl 4.036 in SunOS 4.1.3 with gcc 2.4.5

5. Compiling Perl 4.036 on SunOS 4.1.3

6. Perl 4.036, gcc and SunOS/SPARC2

7. Perl 4.036 with gcc on SunOs 4.1.3

8. Compiling perl 4.036 on SunOS 5.3 (Solaris)

9. Help with installing Perl 4.036 on SunOS 5.3 (solaris 2.3)

10. Help Install - Perl 4.036 on SunOS 4.1.3

11. Compiling Perl 4.036 under SunOS 4.1.3

12. *** HELP- Oracle 7.2, Oraper2.4, SunOS 4.1.3, Perl 4.036 ***

 

 
Powered by phpBB® Forum Software