[patch] make new parts of testsuite work via make check

[Rob Landley]

On Tuesday 20 September 2005 12:48, Bernhard Fischer wrote:
> On Tue, Sep 20, 2005 at 09:10:29AM -0500, Rob Landley wrote:
> >On Tuesday 20 September 2005 02:36, Bernhard Fischer wrote:
> >> >Looking at the patch...
> >> >
> >> >Okay, I'd really rather _not_ introduce unnecessary environmental
> >> > dependencies on running these tests.  When it's run outside of make,
> >> > $srcdir is not set.
> >>
> >> Hm? My copy of trunks runtest reads:
> >> $ grep srcdir= runtest
> >> [ -n "$srcdir" ] || srcdir=$(pwd)
> >
> >Yeah, but if you run the test directly then srcdir has not been set.
>
> If srcdir isn't set then just fall back to ./
> so in *.test:
> . ${srcdir:-.}/testing.sh
> (instead of the . ${srcdir}/testing.sh i had in the propsed patch).
>
> Would that do it?

As long as you don't mind the "links" directory getting created who knows 
where and potentially not getting cleaned up by make clean, then yeah.

However, putting that in each .tests file is a bit overkill if you can just 
insert srcdir at the start of $PATH.

> With your recent change, i now get:
> $ make check 2>&1 >&1 | egrep -v "(FAIL|PASS|UNTESTED)"
> bindir=obj.ia32/busybox.oorig
> srcdir=obj.ia32/busybox.oorig/testsuite \
> obj.ia32/busybox.oorig/testsuite/runtest
> obj.ia32/busybox.oorig/testsuite/runtest: line 106:
> ./busybox.tests: No such file or directory

Yeah, the current directory isn't want it expects it to be.

Oh, I think I understand what you're trying to do.

I just checked in a patch.  Let me know what's still broken. :)

Okay, I checked in that patch _yesterday_.  (Had this email composing window 
open on my desktop since yesterday, never hit "send"...)

> The following tests fail if you're running on a terminal or console with
> COLUMNS not equal to 80:
> busybox --help busybox
> busybox (about the same as the one above)
> busybox --help (ditto)
> busybox-suffix
> busybox-suffix --help

Hmmm.  Dunno why.  The help text is captured by calling busybox, and then we 
make sure that various different invocations give us the same helptext...?

I'll try to remember to take a look...

Rob