[PATCH util-macros] XORG_WITH_GROFF: add HAVE_GROFF_HTML Automake conditional

[Gaetan Nadon]

On Sun, 2010-06-20 at 11:30 -0700, Dan Nicholson wrote:

> On Sat, Jun 19, 2010 at 2:13 PM, Gaetan Nadon <memsize at videotron.ca> wrote:
> > On Sat, 2010-06-19 at 08:50 -0700, Dan Nicholson wrote:
> >
> > On Fri, Jun 18, 2010 at 6:57 PM, Gaetan Nadon <memsize at videotron.ca> wrote:
> >> Groff uses grohtml to generate html output format. This program, in turn,
> >> uses a number of pnm* commands from the netpbm package, psselect
> >> from the psutils package and the ghostscript package.
> >>
> >> These are independently installed, so they could be missing.
> >> A check is made to ensure those dependencies are installed.
> >> If not, the makefile can use the conditional to supress the target.
> >>
> >> Signed-off-by: Gaetan Nadon <memsize at videotron.ca>
> >> ---
> >>  xorg-macros.m4.in |   18 ++++++++++++++++++
> >>  1 files changed, 18 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/xorg-macros.m4.in b/xorg-macros.m4.in
> >> index a8e12ed..a10f80b 100644
> >> --- a/xorg-macros.m4.in
> >> +++ b/xorg-macros.m4.in
> >> @@ -552,6 +552,12 @@ AM_CONDITIONAL([HAVE_DOXYGEN], [test "$have_doxygen"
> >> = yes])
> >>  # --with-groff:         'yes' user instructs the module to use groff
> >>  #               'no' user instructs the module not to use groff
> >>  #
> >> +# Added in version 1.9.0:
> >> +# HAVE_GROFF_HTML: groff has dependencies to output HTML format:
> >> +#                 pnmcut pnmcrop pnmtopng pnmtops from the netpbm
> >> package.
> >> +#                 psselect from the psutils package.
> >> +#                 the ghostcript package.
> >> +#
> >>  # If the user sets the value of GROFF, AC_PATH_PROG skips testing the
> >> path.
> >>  #
> >>  # OS and distros often splits groff in a basic and full package, the
> >> former
> >> @@ -591,6 +597,7 @@ elif test "x$use_groff" = x"no" ; then
> >>  else
> >>    AC_MSG_ERROR([--with-groff expects 'yes' or 'no'])
> >>  fi
> >> +
> >>  # We have groff, test for the presence of the macro packages
> >>  if test "x$have_groff" = x"yes"; then
> >>     AC_MSG_CHECKING([for ${GROFF} -ms macros])
> >> @@ -608,9 +615,20 @@ if test "x$have_groff" = x"yes"; then
> >>     fi
> >>     AC_MSG_RESULT([$groff_mm_works])
> >>  fi
> >> +
> >> +# We have groff, test for HTML dependencies
> >> +if test "x$have_groff" = x"yes"; then
> >> +   AC_PATH_PROGS(GS_PATH, [gs gswin32c])
> >> +   AC_PATH_PROG(PNMTOPNG_PATH, [pnmtopng])
> >> +   AC_PATH_PROG(PSSELECT_PATH, [psselect])
> >> +fi
> >
> > Thanks for looking into this, but I don't think I agree with this
> > approach.
> >
> > That's ok. I vaguely remember drawing a line there when writing the original
> > macro.
> > I am somewhat spoiled by my distro where all the stuff I need is already
> > installed.
> > When I dig  more I find that groff depends on netbpm, psutils and
> > ghostscript.
> > These are separately installed packages.
> >
> > The patch I did amounts to checking the integrity of the platform installed
> > software.
> > Not something we should be doing, as users of groff. On the other we try to
> > be helpful
> > to our users when things go wrong.
> >
> > Had you been the author of the patch, I might have made the same objection.
> > Infinite loop :-)
> 
> Right. I don't really want to be in the business of checking that the
> distro has setup their package correctly. (I just checked fedora that
> I'm on and they don't make groff depend on netpbm). Also, it requires
> that we track how grohtml works internally.
> 
> > I don't think we want to be checking the internal details of
> > grohtml. It would probably be preferred to just try to generate html
> > from a simple source file and see if it produced useful results.
> >
> > Are you suggesting using another tool?
> 
> No, just testing the tools capabilities like libtool does to $CC. I'm
> not checking the details (especially about the troff, which I know
> nothing about), but here's what I basically had in mind.
> 
> AC_MSG_CHECKING([for working grohtml])
> cat > conftest.ms << "EOF"
> ... some extremely minimal troff stub ...
> EOF
> if AC_RUN_LOG([$GROFF -T html conftest.ms]); then
>     have_grohtml=yes
> else
>     have_grohtml=no
> fi
> # clean up output (not sure about the names)
> rm -f conftest.html conftest*.png
> AM_CONDITIONAL([HAVE_GROFF_HTML], [test $have_grohtml = yes])


Ah, ok. Totally missed. Same principle used to check for mm and ms
macros. Yes, better.

> 
> > However, if we do want to just check the program, I'd suggest
> > borrowing heavily from groff's internal GROFF_HTML_PROGRAMS macro.
> > See:
> >
> > Exactly what I did. I simplified for our situation. I picked one  command
> > per package rather than checking all of them.
> >
> > http://cvs.savannah.gnu.org/viewvc/groff/m4/groff.m4?root=groff&view=markup
> 
> Awesome.
> 
> --
> Dan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg-devel/attachments/20100620/7cdccebc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg-devel/attachments/20100620/7cdccebc/attachment.pgp>