Proposed patch to have dvc understand MTN even in a foreign locale

Hello,

In the hope that this is an acceptable way to send changes for DVC,
here's a small one.

message:
  Monotone has several non-english translations.  It means that it might
  output things that are incomprehensible to DVC.  Therefore, we need to
  set a locale that DVC has a chance to understand, such as "C".

  * lisp/xmtn-dvc.el (xmtn-dvc-prepare-environment): Added.
  * lisp/xmtn-run.el (xmtn-prepare-environment): Added to prepare the
    environment as necessary.  For the moment, it's to add LC_MESSAGES=C.

Cheers,
Richard

--

-- 
Richard Levitte                         richard <at> levitte.org
                                        http://richard.levitte.org/

"Life is a tremendous celebration - and I'm invited!"
-- from a friend's blog, translated from Swedish

# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: richard <at> levitte.org-20110204135157-fik6w33thyppv17y
# target_branch: http://bzr.xsteve.at/dvc/
# testament_sha1: 93569d0d765e0f72bbf6b1724ee744f550b38831

Re: Proposed patch to have dvc understand MTN even in a foreign locale

Richard Levitte <richard <at> levitte.org> writes:

> Hello,

Hi Richard!

> In the hope that this is an acceptable way to send changes for DVC,

Perfectly.

> here's a small one.
>
> message:
>   Monotone has several non-english translations.  It means that it might
>   output things that are incomprehensible to DVC.  Therefore, we need to
>   set a locale that DVC has a chance to understand, such as "C".
>   
>   * lisp/xmtn-dvc.el (xmtn-dvc-prepare-environment): Added.
>   * lisp/xmtn-run.el (xmtn-prepare-environment): Added to prepare the
>     environment as necessary.  For the moment, it's to add LC_MESSAGES=C.

Excellent point. 

This will show up in the ada-france server
(mtn://www.ada-france.org?org.emacs.dvc) soon, and after that in the bzr
repository.

Thanks for trying DVC!

--

-- 

dvc and xemacs 21.5-b29

I'm trying to get dvc to work with xemacs 21.5-b29.  I found a couple of
minor issues and a show-stopper.

First, I can't build dvc at all because xemacs erroneously thinks there
is a badly formatted Local Variables section in dvc-build.el.  I fixed
this by placing a page-break (Ctrl-L) at the end of dvc-build.el, just
before the ending comment.

Second, some packages need to be loaded to be able to build.  In
dvc-unified.el, I had to require pp, debug.  XEmacs doesn't have
dired-x, so I just required dired-xemacs instead.  (Maybe I should use
just plain dired?)  Also, xemacs doesn't have
uniquify-list-buffers-directory-modes, so I just added a defvar for
that.  This is probably not right, but it lets dvc build.

Third, xhg.el requires dired-x, which I just changed to require
dired-xemacs instead.

With these changes, xemacs can now compile dvc.  Yay!  However, the
result doesn't work.  I ran make and make install to install all the
files in ~/.xemacs/lib/...  Now when I start xemacs, I get lots of
errors about basically every package having it's auto-autoloads already
loaded.  Other than the 100+ warnings (and slower startup), this is
mostly harmless, I think.

The show-stopper is that when I do (load "dvc-load"), nothing works.
When I try to load any file, I get an error about dvc-buffers-tree being
not defined.  If I defvar that, then there's an error about
dvc-process-running being undefined.  I don't understand how the
auto-autoloads work, so I'm kind of stuck here.

hg diff from dvc-status

I'm having problems with getting diffs from dvc-status. If I do this on
my repository, it all works...

hg status for *snip*/src/ht/knowledgeblog/devel/

 M trunk/plugins/kcite/kcite-todo.org
 M trunk/plugins/kcite/kcite.php
 ? trunk/plugins/kcite/gorun_settings.py
 ? trunk/plugins/kcite/gorun_settings.pyc
 ? trunk/plugins/test-plugin.php

But running dvc-dvc-diff on one of the modified files, gives me a no
such file or directory. 

dvc-dvc-file-diff: Setting current directory: no such file or directory,
trunk/plugins/kcite/

Or this backtrace...

Debugger entered--Lisp error: (file-error "Setting current directory" "no such file or directory" "trunk/plugins/kcite/")
  call-process("diff" nil #<buffer *xhg-file-diff*<2>> nil "-u"
"-La/trunk/plugins/kcite/kcite.php" "-Lb/trunk/plugins/kcite/kcite.php"
"/tmp/DVC-file-diff-base2582N0c" "/tmp/DVC-file-diff-mod2582a-i")
  dvc-dvc-file-diff("trunk/plugins/kcite/kcite.php" (xhg (last-revision
"/mnt/sdb1/home/phillord/src/ht/knowledgeblog/devel/" 1)) (xhg (local-tree
"/mnt/sdb1/home/phillord/src/ht/knowledgeblog/devel/")) t)
  apply(dvc-dvc-file-diff ("trunk/plugins/kcite/kcite.php" (xhg (last-revision
"/mnt/sdb1/home/phillord/src/ht/knowledgeblog/devel/" 1)) (xhg (local-tree
"/mnt/sdb1/home/phillord/src/ht/knowledgeblog/devel/")) t))

Re: dvc and xemacs 21.5-b29

Raymond Toy writes:
> 
> I'm trying to get dvc to work with xemacs 21.5-b29.  I found a couple of
> minor issues and a show-stopper.

	I attempted the same thing using a checkout from
http://bzr.xsteve.at/dvc/ (and my xemacs claims to be the same
version).  I ran into most of your minor issues and some others
besides (see below), but eventually got a compiling dvc.

> I ran make and make install to install all the
> files in ~/.xemacs/lib/...
...
> The show-stopper is that when I do (load "dvc-load"), nothing works.

	I didn't attempt to install dvc; instead, I just did
(load-file "/path/to/dvc/dvc-load.el"), pointing at the directory
where I had compiled using xemacs.  Some things work: dvc-status
worked on an hg directory (and on dvc's source directory, under bzr),
and within that I could do 'e' for an ediff (although sometimes I'd
get an error about "display-buffer switched frames on its own!!").

	But when I tried 'i' or 'c', I get:

Symbol's function definition is void: dvc-derived-mode-p.

This turns out to be unsurprising since dvc-derived-mode-p is defined
in dvc-emacs (only).  Copying the definition over to dvc-xemacs.el
seems to get rid of that error.  But there are other problems; for
instance, xhg-push fails with an error about try-completion.  I