[Rubygems-developers] Misc Test Errors Fixed -- One odd error remaining

Mauricio Fernández batsman.geo at yahoo.com
Mon Aug 9 09:12:20 EDT 2004


On Mon, Aug 09, 2004 at 08:10:36AM -0400, Jim Weirich wrote:
>   1) Failure:
> test_extract_entry_works(TC_TarReader::TC_TarInput)
>     [./test/test_package.rb:540:in `test_extract_entry_works'
>      ./test/test_package.rb:529:in `each_with_index'
>      ./test/test_package.rb:529:in `each_entry'
[...]
> <16768> expected but was
> <17792>.

It's clearer in octal: 040600 vs 042600.  I'm reviewing SuSv2 to try
to find the meaning of the bits over 11, but the fact that S_ISGID is
being set spuriously might be FS-dependent (package.rb is only doing a
normal chmod). This is probably a non-issue since higher bits are rarely
(never) used.

In other words, it is very likely that the FIXME was right:
    502             #FIXME: are these modes system dependent?
	503             @entry_modes = [0100612, 0100636, 040600]
The assertion should probably be
 assert_equal(@entry_modes[i], File.stat(name).mode & 07777)
with @entry_modes restricted to meaningful values.

> Both systems are running Debian Linux.  The actual failure occurs in an 
> assert that is not run on a windows box which leads me to believe that 
> the results are system dependent.  But both of my systems are Debian 
> linux.  Since this occurs in the bowls of the new tar/gzip code, I'm 
> hoping someone with more knowledge about that area can take a look.  If 
> you can't reproduce the problem, feel free to ask me to try anything.

Which filesystems are you using? If they're different, that might
explain the chmod oddity. Note that the unit tests ran OK at least on
FreeBSD and Gentoo.

-- 
Running Debian GNU/Linux Sid (unstable)
batsman dot geo at yahoo dot com



More information about the Rubygems-developers mailing list