Re: master 5a125fb5a97 1/2: Update to Org 9.7.3

[Eli Zaretskii <eliz@gnu.org>]

> From: Ihor Radchenko <yantar92@posteo.net>
> Cc: Stefan Monnier <monnier@iro.umontreal.ca>, kyle@kyleam.com,
>  stefankangas@gmail.com, emacs-devel@gnu.org
> Date: Thu, 13 Jun 2024 15:44:02 +0000
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> I don't think this argument is very convincing: getting the info "from
> >> org.org via an awk script" is not that different from getting it "from
> >> org.org via org-export plus install-info".
> >
> > If we could use install-info, we didn't need that script at all.  The
> > comments in the top-level Makefile explain why we don't use
> > install-info for generating DIR in the build tree.
> 
> I am seeing
> 
> ## FIXME:
> ## If info/dir is missing, but we have install-info, we should let
> ## that handle it.  If info/dir is present and we do not have install-info,
> ## we should check for missing entries and add them by hand.
> 
> Does it mean that making use of install-info would be fine, if it is available?

No:

  ## It would be much simpler if info/dir was only created in the
  ## installation location by the install-info rule, but we also
  ## need one in the source directory for people running uninstalled.
  ## FIXME it would be faster to use the install-info program if we have it,
  ## but then we would need to depend on ${INFOS}, which would
  ## slow down parallelization.

So there are two problems:

  . install-info doesn't have to be available
  . even if it is available, it gets in the way of highly-parallel builds