Re: [brlcad-devel] [brlcad-commits] SF.net SVN: brlcad:[49943] brlcad/trunk/regress/vdeck-comgeom-g.sh

[Tom B. <tom...@gm...>]

On Mon, Apr 9, 2012 at 13:11, Christopher Sean Morrison <br...@ma...> wrote:
>
> On Apr 7, 2012, at 12:52 PM, tbr...@us... wrote:
>
>> Added Paths:
>> -----------
>>    brlcad/trunk/regress/vdeck-comgeom-g.sh
>
> That's quite a mouth-full... how about just "comgeom.sh"
> since both tools pertain to import/export of the comgeom format? :)

I agree.  I admit I choked on the name.  I really would like to see
vdeck turned into g-comgeom (and maybe the pair renamed or aliased to
g-gift <=> gift-g for consistency and clarity).

I'll change it.

> Impressive seemingly exhaustive test too..

Hm, not really, the two tests should at least signal any incompatible
changes in vdeck and gross errors in comgeom-g.  I just assume that
the current vdeck conversion case is correct for now and use it for
the baseline for future revisions (it looks good when reconverted to
.g, and I spot checked a few solid params), and rely on the existing
non-zero exit tests in comgeom-g to tell about recognized badness.

> Glad to see the update to standard I/O redirects [instead] of relying
> on Expect since that should make deployment testing easier.

I'm glad I found an example in red because I don't think I've found a
book or tutorial that shows how to do it.  My first cut used Expect
but then I didn't find any other use of Expect in regress and knew the
same idiom had to be used for mged stuff.

> All around, Nice!

Thanks.  It helps to have an itch!

Best regards,

-Tom