[PATCH/RFC 0/2] no-op shell script i18n infrastructure

While the no-op C patch series is cooking I thought I'd send an RFC for the no-op shellscript infrastructure. So here (as threatened in <AANLkTinUtqJJHNyS9CxrC=VnS87v=GH=pOw9yr4r=pii@mail.gmail.com>[1]) is the infrastructure needed to gettextize the shell script programs.

To quote my previous E-Mail[1] here are the things that are DONE and
still need to be done.
    
    Open issues:
    
       * Write documentation for git-sh-i18n.sh and git-sh-i18n--envsubst
         like we have for git-sh-setup (already in WIP form).

It now has documentation. So this isn't a problem anymore.
    
       * git-sh-i18n--envsubst is still too fat:
    
            $ ldd -r git-sh-i18n--envsubst
                linux-vdso.so.1 =>  (0x00007fffc60fd000)
                libz.so.1 => /usr/lib/libz.so.1 (0x00007f25cff9e000)
                libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8
    (0x00007f25cfbfd000)
                libpthread.so.0 => /lib/libpthread.so.0 (0x00007f25cf9e0000)
                libc.so.6 => /lib/libc.so.6 (0x00007f25cf67f000)
                libdl.so.2 => /lib/libdl.so.2 (0x00007f25cf47b000)
                /lib64/ld-linux-x86-64.so.2 (0x00007f25d01c0000)
    
         It only needs to link to libc, but I didn't find out when I last
         checked how to convince the Makefile to only link against
         that. Help welcome :)

This is still an issue. It needs to link to libgit for xmalloc() and
friends, but it doesn't need libz, libcrypto etc.
    
       * Deal with the changes in 92c62a3f4f93432c0c82e3031a9e64e03ba290f7:
    
            $ git --no-pager  grep -A1 abomination *.sh
            git-pull.sh:            # XXX: This is an abomination
            git-pull.sh-            require_clean_work_tree "pull with
    rebase" "Please commit or stash them."
    
         The changes Ramkumar Ramachandra made in 92c62a3f4f, while good,
         are hard to square with i18n.
    
         I think I'll just leave those bits untranslated for now and deal
         with them later, since I'm trying to keep this minimal.

This is not part of this RFC series, but I still haven't dealt with it
in my branch.
    
    And then there's the issue that unlike the C patches these will not be
    a no-op that'll be optimized away by the compiler. We'll be calling an
    external program for displaying messages. While this is a trivial cost
    on Unix (especially in the context we're using it, i.e. not in tight
    loops) it's more expensive on Windows.
    
    I don't see any way to deal with that short of implementing some
    pre-processor, but I think the cost is worth it, but others might
    disagree of course.

This is still a problem I think. Although there's been some work on
this on the MinGW front from what I can gleam from the mailing list.
    
1. http://www.spinics.net/lists/git/msg151971.html

Ævar Arnfjörð Bjarmason (2):
  git-sh-i18n--envsubst: our own envsubst(1) for eval_gettext()
  git-sh-i18n.sh: add no-op gettext() and eval_gettext() wrappers

 .gitignore                              |    3 +
 Documentation/git-sh-i18n--envsubst.txt |   36 +++
 Documentation/git-sh-i18n.txt           |   57 ++++
 Makefile                                |    2 +
 git-sh-i18n.sh                          |   17 ++
 sh-i18n--envsubst.c                     |  444 +++++++++++++++++++++++++++++++
 6 files changed, 559 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/git-sh-i18n--envsubst.txt
 create mode 100644 Documentation/git-sh-i18n.txt
 create mode 100644 git-sh-i18n.sh
 create mode 100644 sh-i18n--envsubst.c