[Caml-list] ocamlbuild, ocamldoc and module packs

[Caml-list] ocamlbuild, ocamldoc and module packs

[Guillaume Yziquel]

Hi.

I've been trying recently to use ocamldoc with ocamlbuild on module
packs.

I have a prog.odocl file containing things like

	ModuleOne/A
	ModuleOne/B
	ModuleTwo/C
	ModuleTwo/D

and the ocamldoc documentation has at its root the module A, B, C and D.
Not ModuleOne and ModuleTwo.

Replacing prog.odocl naively by

	ModuleOne
	ModuleTwo

fails with a rather verbose ocamldoc solver error.

My _tags file contains something like:

	<ModuleOne/*.cmx> : for-pack(ModuleOne)
	<ModuleTwo/*.cmx> : for-pack(ModuleTwo)

Is there any way to make ocamldoc behave more nicely with respect to
module packs (if possible via ocamlbuild)?

-- 
     Guillaume Yziquel

Re: [Caml-list] ocamlbuild, ocamldoc and module packs

[Pietro Abate]

Hello.

As far as I know ocamldoc doesn't play very well with -pack. There is 
a very old bug on mantis about this.

I'm in a situation that is worst then yours where I use pack to hide
modules with the same name as in ModuleOne/A and ModuleTwo/A and
ocamldoc refuses altogether to generate doc for both modules.  I've not
found a solution neither for my problem, or more in general for the one
you describe below.

I think the good ppl that put Batteries together had similar issues.

If you look at the Batteries source there is a lot of code aimed at
solving this problem by effectively creating new source files that can
be properly digested by ocamldoc. For what I gathered is pretty specific
to their case, but probably you can use it as an example to write an
appropriate ocamldoc plugin.

Maybe the recent release of Argot by XC solves some of these problem
(just a wild guess, I didn't look at it very carefully).

Hope this helps.

p

On Thu, Nov 10, 2011 at 09:52:59PM +0100, Guillaume Yziquel wrote:
> Hi.
> 
> I've been trying recently to use ocamldoc with ocamlbuild on module
> packs.
> 
> I have a prog.odocl file containing things like
> 
> 	ModuleOne/A
> 	ModuleOne/B
> 	ModuleTwo/C
> 	ModuleTwo/D
> 
> and the ocamldoc documentation has at its root the module A, B, C and D.
> Not ModuleOne and ModuleTwo.
> 
> Replacing prog.odocl naively by
> 
> 	ModuleOne
> 	ModuleTwo
> 
> fails with a rather verbose ocamldoc solver error.
> 
> My _tags file contains something like:
> 
> 	<ModuleOne/*.cmx> : for-pack(ModuleOne)
> 	<ModuleTwo/*.cmx> : for-pack(ModuleTwo)
> 
> Is there any way to make ocamldoc behave more nicely with respect to
> module packs (if possible via ocamlbuild)?
> 
> -- 
>      Guillaume Yziquel
> 
> -- 
> Caml-list mailing list.  Subscription management and archives:
> https://sympa-roc.inria.fr/wws/info/caml-list
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs

-- 
----
http://en.wikipedia.org/wiki/Posting_style

Re: [Caml-list] ocamlbuild, ocamldoc and module packs

[Guillaume Yziquel]

Le Friday 11 Nov 2011 à 14:08:47 (+0100), Pietro Abate a écrit :
> Hello.
> 
> As far as I know ocamldoc doesn't play very well with -pack. There is 
> a very old bug on mantis about this.
> 
> I'm in a situation that is worst then yours where I use pack to hide
> modules with the same name as in ModuleOne/A and ModuleTwo/A and
> ocamldoc refuses altogether to generate doc for both modules.  I've not
> found a solution neither for my problem, or more in general for the one
> you describe below.

Well, I'm having the same issue, as ModuleOne and ModuleTwo satisfy a
signature implying a common module.

> I think the good ppl that put Batteries together had similar issues.
> 
> If you look at the Batteries source there is a lot of code aimed at
> solving this problem by effectively creating new source files that can
> be properly digested by ocamldoc. For what I gathered is pretty specific
> to their case, but probably you can use it as an example to write an
> appropriate ocamldoc plugin.

I guess workarounds or ugly hacks are way too overkill for my current
purpose.

> Maybe the recent release of Argot by XC solves some of these problem
> (just a wild guess, I didn't look at it very carefully).

I'll have a try at Argot as soon as the current situation gets too
frustrating.

> Hope this helps.

Thanks.

> p

-- 
     Guillaume Yziquel