Re: [Caml-list] Package installation assumptions made by odb

[Edgar Friendly <thelema314@gmail.com>]

On 02/13/2012 08:39 AM, Romain Bardou wrote:
> For instance, I'm currently working on a suite of tools named
> like foo-feature1, foo-feature2, foo-feature3 and so on. The package
> would obviously be named foo, but it does not make sense to have a
> single "foo" executable. It does not make sense to have once package per
> executable either.

If it's all compiled at once, it's probably one package - don't split it 
up.  I see three options:
1) name your package (in odb) foo-feature1 (ugly, but simple and works)
2) provide a dummy executable `foo` for odb to detect
3) convince me that your requirements are common enough to deserve more 
complexity in odb

There's no need for odb to detect (or even know about) all three 
foo-features, it's perfectly fine for it to only know about one of them 
and base its decision to recompile on that one.

> So I guess you need some way to override this
> behavior when needed. Can odb extract this kind of information from an
> _oasis file, or from a META file? I'm not very familiar with those.
>
odb doesn't directly access either of those files.  Oasis-db extracts 
lots of info from _oasis files and make them available to odb in a form 
it understands.  Findlib also handles META files on behalf of odb, 
although for different uses.

> Regarding Melt, I'm a bit confused: you say in this mail that "its odb
> package name could be melt or melt-build, depending on which part of
> melt should be used to detect its installation". But in your guidelines
> it is written: "Programs that are both should do both". By the way, Melt
> actually installs two libraries: Melt and Latex.
>
At the moment, the library/executable distinction is important only for 
detecting installed packages.  If a package is a library, it is 
installed only if ocamlfind knows of an ocamlfind package with the same 
name as the odb package.  Similarly, if a package is a executable, it is 
installed only if `which` knows of that executable name.

> Regarding the version number, if you need to define a standard I think
> it would be good if this standard was to call the executable with the
> "-version" option. It is, after all, the behavior of "ocamlc". This does
> not work for libraries though;

Libraries are not a problem - ocamlfind provides a version for them 
trivially (from their META file).  IIRC, I wrote some code that tried to 
run `foo -version` to get the version of foo, but it looks like I'll 
have to parse the output of most programs looking for a version number.

E.