[Po4a-devel]Release, and other considerations

[Nicolas François]

On Mon, Feb 14, 2005 at 09:03:55AM +0100, Martin Quinson wrote:
> On Sat, Feb 12, 2005 at 03:54:16PM +0100, Jordi Vilalta wrote:
> > Three months have past since the last release and we have support for many 
> > new formats, and some bug fixes. Should we begin thinking on a new release 
> > soon?
> 
> I strongly second you. Moreover, the email address given as maintainer of
> 0.19 is my tuxfamily.org one, which is dead. I vote for a rapid upload.
> 
> I think that this is Nekral's call. Could we release as is, or you still
> have something to fix in TeX ? It's not released yet, but I think it should
> since it almost works, if I understood well. 

Short version: some small things to commit, I can probably be ready in one
or two weeks.


Long version:
Here is what is mostly lacking to TeX* modules:
  - Documentation
    (I'm probably the only one who can use them)
    I may manage to add a minimal documentation and propose support on the
    list
    => This should not stop/delay a release

  - tabular environment
    It may change users PO in later versions.
    I've failed my first try. It may require a design change of the
    parser. => This should not stop/delay a release (but may change
    its releasable state;)

  - Tests
    A release could provide these tests. => This should not stop/delay a
    relese

  - Known bug, not fixed, which may require a design change:
    % in verbatim environment are handled as comments
    (this is triggered by Python)
    => This should not stop/delay a release (it could be documented)

I've got the feeling that they are not perfect, but "usable".

I may have done an error by committing the PythonDoc module: it is
probably a pure LaTeX module (this only permits to avoid some "% po4a:"
lines)

I've got some updates that I can push now or really soon:
  - a "generic" command subroutine, which permits to specify which
    arguments have to be translated.
  - definitions for all LaTeX2 commands I could find, using this "generic"
    command processing
  - a way to specify which commands the parser is allowed to separate
    (I'm using a '*' prefix for the command)
  - handling of unterminated arguments (when an arguments contains an empty
    line)
  - including files with \input
  - adding \clearpage before and after inclusions done by \include
    I discovered this while trying to understand why the TOC of Nicolas'
    book was modified after a gettextization. (I will stop investigation
    on this issue, since manually inserting the file, with and without
    \clearpage, also produce the TOC diff)

There is another design issue: I assumed that commands are always followed
by optional arguments and then by mandatory arguments. This is wrong: some
are followed by arguments between parenthesis (e.g. dashbox); some have
optional arguments preceded by mandatory arguments (e.g. newcommand,
savebox). I have an idea on how to fix this, but this will change the way
user provide "generic" commands (this should not produce any PO diff).

That's all folks!

To conclude, I advocate for a release (maybe with a 20-1 ;), and think
the LaTeX module can be included (with TeX). I need to test the Python
Documentation with the LaTeX module, and see if there really are some
Python specific commands.

> We are not in such a hurry either. If you need another month, that's cool
> for me. Next month will be completely crazzy for me anyway. Moving to
> another city again, soon a new baby and such.

congratulations!

-- 
Nekral