Follow-up Comment #3, bug #68676 (group groff): Hi Deri,
At 2026-09-09T16:25:31-0400, Deri James wrote:
> Follow-up Comment #2, bug #68676 (group groff):
>
> By accident I put a comment in bug #68667 instead of here!
Thanks for giving it another shot. :)
> The test script which smoke-tested pdf output failed because it was
> constrained to a particular version of gropdf. I offered an alternative
> test:-
>
> printf ".nf\n.kern 0\nHello my big beautiful World"|test-groff -Tpdf
> -P-f3 | grep -Faq '0.000 Tw [ (Hello my big beautiful World)] TJ';
> echo $?
[...]
> The grep pattern is for a complete line in the pdf which will not be
> affected by possible changes to kern pairs in the font (.kern 0) nor
> any filling adjustments (.nf). The position of this line in the file
> may well change across gropdf versions but there is a low probability
> this line will change.
And that `grep` won't care which line in the file matches, so it should
be robust to that.
> The -P-f3 ensyres that the object streams are not compressed binary
> data.
Understood.
> The change to this test means that it is currently just testing if
> troff inckudes the line "T pdf".
True, but substantial complexity hides beneath that simple test.
If the "tmac path" is misconfigured such that GNU troff can't find the
device description file, or the font description file for "TR", or if no
Perl interpreter is installed, the test will fail.
> I agree that specific driver tests belong in their own tests
> directory.
Would you like to add the foregoing "big beautiful world" test to
src/driver/gropdf/tests? ;-)
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?68676>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
