[Docutils-develop] Re: Dependency scanner

[Felix W. <Fel...@gm...>]

Beni Cherniavsky wrote:

> I think a dependency scanner that is separate from the processing
> program is generally a messy, suboptimal and unnecessary approach.
> The trick is that until you have processed the document first time,
> there is no need to know the exact dependencies -- you are going to
> process it anyway.  A doubt requiring to know the dependencies only
> arises when you have processed it already and want to detect when the
> output becomes out-of-date.

I see.

> All this dependency information can be generated as a by-product of
> processing the source.

I just added a --dependency-file option:

------------------------------------------------------------------------
~/tmp $ cat dep.txt
.. include:: a.txt

.. raw:: html
   :file: b.txt

.. csv-table::
   :file: csv.txt

.. image:: image.png

~/tmp $ touch a.txt b.txt image.png

~/tmp $ echo content for the csv-table > csv.txt

~/tmp $ rst2html.py --dependency-file=dep-file.txt dep.txt > /dev/null

~/tmp $ cat dep-file.txt 
a.txt
b.txt
csv.txt
image.png
------------------------------------------------------------------------

Is the realization and implementation okay?  If yes, I'll add tests and
docs.

-- 
When replying to my email address, please ensure
that the mail header contains 'Felix Wiemann'.

http://www.ososo.de/