[PATCH:libXi] Make shadow man pages generated by asciidoc work with Solaris man

[Peter Hutterer]

On Mon, Sep 19, 2011 at 04:43:48PM -0400, Gaetan Nadon wrote:
> On Mon, 2011-09-19 at 13:12 -0700, Alan Coopersmith wrote:
> 
> > On 09/17/11 13:59, Gaetan Nadon wrote:
> > > On Fri, 2011-09-16 at 21:00 -0700, Alan Coopersmith wrote:
> > >> Solaris man requires .so entries in man pages include the man section
> > >> directory, not just the man page filename, even when it's in the same
> > >> directory, so use sed to add it in.
> > >>
> > >> Signed-off-by: Alan Coopersmith<alan.coopersmith at oracle.com  <mailto:alan.coopersmith at oracle.com>>
> > >> ---
> > >>   man/Makefile.am |    3 +++
> > >>   1 files changed, 3 insertions(+), 0 deletions(-)
> > >>
> > >> diff --git a/man/Makefile.am b/man/Makefile.am
> > >> index d4c37b3..28211a1 100644
> > >> --- a/man/Makefile.am
> > >> +++ b/man/Makefile.am
> > >> @@ -159,6 +159,9 @@ MAN_SUBSTS += -e 's/libmansuffix/$(LIB_MAN_SUFFIX)/g'
> > >>   MAN_SUBSTS += -e 's/\[FIXME: source\]/$(XORG_MAN_PAGE)/'
> > >>   MAN_SUBSTS += -e 's/\[FIXME: manual\]/XINPUT FUNCTIONS/'
> > >>
> > >> +# asciidoc generates shadow page references without the man section directory
> > >> +MAN_SUBSTS += -e 's/^\.so X/.so man$(LIB_MAN_SUFFIX)\/X/'
> > >> +
> > >>   SUFFIXES = .man .$(LIB_MAN_SUFFIX)
> > >>   .man.$(LIB_MAN_SUFFIX):
> > >>   	$(AM_V_GEN)$(SED) $(MAN_SUBSTS)<  $<  >  $@
> > >
> > > On Debian, the patch is a noop. I see no difference in the generated shadow man
> > > pages. That would mean that what asciidoc or xmlto generates is different on
> > > Solaris.
> > 
> > Actually, that would be on Fedora, since this is the patch I use on Solaris 
> > systems without xmlto or asciidoc to build the tarballs Peter produces and
> > publishes.
> > 
> > If you download http://www.x.org/releases/individual/lib/libXi-1.4.3.tar.bz2
> > you'll see in it, libXi-1.4.3/man/XFreeDeviceList.man has:
> > 
> > .so XListInputDevices.libmansuffix
> > 
> > Perhaps we need to get Peter building with a fixed version of asciidoc then?
> 
> That would be a sensible test. Perhaps some listeners on the list can
> have a look in their libXi/man and report the version of
> asciidoc/xmlto/xsltproc if they have such a shadow manpage without the
> directory. This a fairly complex stack of tools and it is not easy to go
> back in time, particularly when we don't know exactly where the problem
> is. This would help Peter find out if it is a local workstation problem
> or a tool problem.

which version of asciidoc has the fixes? 

Cheers,
  Peter