Path changes in 21.5.29

[Ron Isaacson]

All grumbling aside, I'm still struggling with this... I don't think
the Makefile was properly updated to reflect the change from lib to
share. The install-arch-dep target does:

  if test "${prefix}" != "${exec_prefix}"; then \
    $(MAKEPATH) ${exec_prefix}/lib/${instvardir}; \
    for dir in \
          lib/${inststaticdir} \
          lib/${instvardir}/etc \
          lib/${instvardir}/info \
          lib/${instvardir}/lisp; do \
      if test ! -d ${exec_prefix}/$${dir}; then \
        $(LN_S) ${prefix}/$${dir} ${exec_prefix}/$${dir}; fi; \
    done; \
  fi

However, at startup time, it seems to be looking for these directories
in share/${instvardir}, not lib. Patch attached.

<stephen at xemacs.org> wrote:
> 
> Ron Isaacson writes:
> 
>  > I've had quite a bit of trouble getting 21.5.29 working in our
>  > environment. After comparing paths.h between 21.5.28 and 21.5.29, I
>  > finally figured out why:
> 
> xemacs -debug-paths
> 
> might short-circuit that a bit.
> 
>  > In the future, I'd appreciate it if this sort of thing were
>  > highlighted right up top in the release notes, along with other things
>  > that are important to your friendly neighborhood local maintainers...
> 
> I'll do what I can, but with two years between beta releases, I can't
> really say I'm surprised something like that snuck through.  I'm not
> even sure that the actual change wasn't in 21.5.28 or even earlier;
> the logs in 21.5.29 look like bugfixes to an earlier change.
> 
> You also shouldn't take the fact of a release or its absence too
> seriously; compared to any given commit, the only extra QA that a
> release gets in 21.5 is a guarantee that it builds, runs, and shows no
> new regressions on my machines (Mac OS X with default GCC and Gentoo
> Linux with GCC 4.2 currently) in my default configuration (with Mule,
> X11, modules, debug, and error checking, and maybe a couple other
> minor features).
> 
> AFAIK, few of the active developers have the resources to regularly do
> installs, so beta testing would be very helpful for this kind of
> thing.

--
Ron Isaacson
Morgan Stanley
ron.isaacson at morganstanley.com / (212) 762-1051

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Makefile.patch
Type: application/octet-stream
Size: 735 bytes
Desc: not available
Url : http://calypso.tux.org/pipermail/xemacs-beta/attachments/20090714/2cc545b4/attachment.obj