Re: [groff] 1.22.4.rc4 - Final RC before official 1.22.4

[Bertrand Garrigues]

Hi Ingo,

On Sat, Dec 01 2018 at 04:44:45 PM, Ingo Schwarze <address@hidden> wrote:
> OpenBSD-current:
>
> address@hidden $ make check
> make  check-am
> make   
> make  all-am
> make  check-TESTS
> PASS: contrib/gdiffmk/tests/gdiffmk_tests.sh
> sed -e "s|address@hidden@]|/usr/ports/pobj/groff-1.22.4/groff-1.22.4.rc4|g"  
> -e "s|address@hidden@]|missing|g"  ./contrib/hdtbl/examples/test-hdtbl.sh.in 
> > contrib/hdtbl/examples/test-hdtbl.sh
> chmod +x contrib/hdtbl/examples/test-hdtbl.sh
> FAIL: contrib/hdtbl/examples/test-hdtbl.sh
[...]
> This is likely due to the fact the we configure the OpenBSD port
> as follows, in order to avoid external dependencies.  Even though
> we moved groff from our base system to ports several years ago, we
> still consider it essential system software that we do not want to
> depend on anything else, in particular not on heavier stuff like
> ghostscript or X.
>
> CONFIGURE_ARGS +=       --with-awk=awk --without-gs \
>                         --without-libiconv-prefix --without-uchardet \
>                         --without-urw-fonts-dir --without-x \
>                         pnmcrop=missing pnmcut=missing pnmtopng=missing \
>                         pnmtops=missing psselect=missing
> CONFIGURE_ENV +=        ac_cv_prog_PDFFONTS= \
>                         ac_cv_prog_PDFIMAGES= \
>                         ac_cv_prog_PDFINFO= \
>                         ac_cv_prog_XPMTOPPM=
>
> So it appears you are unconditionally testing for optional features
> in the test suite.

Yes, I will work on Colin's suggestion wo exit with 77 in order to yield
a 'skip' result if the test can't be done due to missing fonts or
programs.


> Solaris 11 / 10 / 9:
> PASS: contrib/gdiffmk/tests/gdiffmk_tests.sh
> FAIL: contrib/hdtbl/examples/test-hdtbl.sh
> FAIL: contrib/mom/examples/tests-mom.sh
> ============================================================================
> Testsuite summary for GNU Troff 1.22.4.rc4
> ============================================================================
> # TOTAL: 3
> # PASS:  1
> # SKIP:  0
> # XFAIL: 0
> # FAIL:  2
> # XPASS: 0
> # ERROR: 0
>
> The test output is long and cryptic, i can't readily make head or tail of
> it, see the attachments.

For hdtbl, looks like the 'font_x.ps' and 'font_n.ps' files are not
correctly generated:

  Checking 
/home/schwarze/groff-1.22.4.rc4.solaris11/build/contrib/hdtbl/examples/fonts_n.ps
    Error: expected 38 pages, found        1 pages
  Checking 
/home/schwarze/groff-1.22.4.rc4.solaris11/build/contrib/hdtbl/examples/fonts_x.ps
    Error: expected 38 pages, found        1 pages
 FAIL contrib/hdtbl/examples/test-hdtbl.sh (exit status: 255)

For mom examples I don't know yet.

> Anyway, three tests isn't much of a suite, i think this says more about
> whether the new release works:
>
> address@hidden $ pwd
> /co/mdocml
> address@hidden $ make regress
> [...]
> All 1199 tests OK: 491 ascii 211 man 18 utf8 25 html 219 markdown 235 lint
>
> The point being that for all the over 400 ascii tests (and also for
> the handful of utf8 tests), groff output matches the expectation.
> Admittedly, that suite isn't exercising PostScript, PDF, or HTML.

Yes 'test-hdtbl.sh' and 'tests-mom.sh' were probably a bit too hastily
written and are far from being a serious test suite, but I wanted to
have a way to automatically check if some generated 'ps' or 'pdf' files
are broken.

Regards,

Bertrand