Re: [Doxygen-users] Adding Custom Commands to execute Shell Scripts

[Adam N. <a.n...@sh...>]

>> Couldn't you use INPUT_FILTER to write a parser that runs these \shell
>> commands instead?
> 
> that means we would need to run a lexer and figure out the comments,
> and then parse the language tags. do you know of such a script? a
> \shell tag is so much easier and natural.

Depends how flexible you need it to be.  Unless you want to run hundreds of
different shell commands you could just replace it with some generic tokens
like $OUTPUT_GOES_HERE and then use sed or something in INPUT_FILTER to do a
search and replace with the output of some predefined commands.  The advantage
there being if you change one of the programs that generates the output, you
only need to update your INPUT_FILTER script rather than hundreds of \shell
commands scattered throughout the docs/code.

> i'm not sure why we are talking about 'platform independent'
> documentation generation when clearly doxygen is filled with many
> interesting features that execute shell commands at a user's demand,
> like the INPUT_FILTER configuration.

Executing shell commands isn't a problem, but shell scripts usually run
external programs and these can be configured differently on different
systems.  That's why autoconf and other build environments have tests to make
sure programs exist, are the correct versions, etc.  I would imagine that in
an ideal world INPUT_FILTER would call a local script which has been generated
as part of the configure process.

> anyway, the information we put in
> our doxygen comments and how we generate the actual documentation
> heavily depends on our environment, even though our project is
> platform independent...

Hope your documentation generation procedure is well documented then ;-)

> as my original email said, all doxygen has to do is pass the "progname
> args" to the portable_system function.

Well if it's that simple I'm sure a patch would be cheerfully accepted :-)

Cheers,
Adam.