0.15.3 test failures

David Relson relson at osagesoftware.com
Mon Sep 15 00:12:00 CEST 2003


On Sun, 14 Sep 2003 23:51:52 +0200
Matthias Andree <matthias.andree at gmx.de> wrote:

> David Relson <relson at osagesoftware.com> writes:
> 
> > Looks like the right value in the wrong order.  I'll fix the test by
> > adding a call to "sort" for 0.15.4.
> 
> Make sure to use "env LC_COLLATE=C sort" to immunize things against
> environment difficulties. sort is sensitive...

Looking at Clint's reported result and the test script, it appears that
file name globbing is giving non-sorted results, a problem I've not
encountered on my box.  It's not clear to me that there _should_ be a
need for us to sort the output to compensate.  Is there a standard that
pertains to this?

At present, tests/t.frame has the following:

	LANG=
	LC_COLLATE=
	LC_ALL=
	unset LANG || :
	unset LC_COLLATE || :
	unset LC_ALL || :

Are you suggesting we change that?




More information about the bogofilter-dev mailing list