Re: ant dependency on jython and antlr

[Daniel Bonniot <Daniel.Bonniot@inria.fr>]

[I posted this on Dec 22nd, but it seems it never got through, so I 
repost it.

Since then I found bug #211560, which raises a similar concern with ant. 
It was sent in mid September, and did not get any answer. We really need 
to address this issue.]


Ben Burton wrote:

> Let me preface this by stating that I know very little about ant, and I
> have no idea how ant specifically interacts with jython.
>  
I'm not familiar with jython either. What I am assuming is the
following: that the jython ant task provides a way to write build files
for jython projects.  That is, if you develop with jython and ant, you
can write a build.xml file with:
 <jython source="myfile.py" destination="build optimize="all" />
that will compile your source file (invented attributes, that's just the
idea).

If this is not the case, what I'm saying is irrelevant, but after
looking at the info I could find on the web, I believe this is close to
truth. If someone knows better, please comment.

> Jython itself is an implementation of the python scripting language.
> The python package does not suggest every package that uses python
> scripting, nor does it suggest every package that provides a simple
> interface for editing python scripts.  The relationship works the other
> way around: packages that use python scripting will need to depend on
> python, 
I don't believe ant can be scripted by jython. Is it the case?

> and packages that provide script writing facilities will
> probably suggest python also.
>  
I would not mind ant suggesting jython. I object it _depending_ on
jython, because it causes bloat (and a policy violation).

> In a similar way, it does not seem sensible for the jython maintainer to
> hunt down every package that might use python scripting 
Agreed, but this is not the case.

> or offer a fancy
> interface for writing python scripts.
>  
He does not _need_ to track those. But ant is a very common tool, and if
you want to program in python, you could _consider_ using ant. Hence the
'Jython suggests ant'. I would not mind if it was not there either, btw.
I just think it makes sense.

> You might also find bash a useful tool (for the same purpose).  Or make.
>  
Both are essential, so we are not used to think about relations with them.

> You might find vim a useful editor.  Or emacs, or kate.  Or anything
> else with python-specific support.  You might find idle a useful IDE
> (this is python after all).  I could go on.
>  
I think it makes sense to _suggest_ packages that provide specific
support for jython. The reasoning goes like this: I found a reference to
jython on the web, and I think I will try it. So I do: apt-get install
python. Then I wonder: are there any tools that will help me write
jython programs? Well, let's look at the suggests list (which apt-get
would have shown you).
I you look for packages X that suggest jython (the other way around),
you get a very different information: it would mean that jython helps
you use X. But I want to use jython, not X!

From the policy: "Using this field [suggests] tells the packaging
system and the user that the listed packages are related to this one and
can perhaps enhance its usefulness, but that installing this one without
them is perfectly reasonable."

> > It's the same relationship as between a compiler and an editor: the 
> > editor helps me use the compiler, the compiler does not help me use the 
> > editor.
> >    
>
> But neither suggests the other.
>  
ocaml and octave2.1 suggest emacs.

> Can you please write a more detailed account of how ant relates to
> jython?  This would probably be a helpful addition to this discussion. :)
>  
See above, especially the first answer.

Cheers,

Daniel

PS: Ben, no need to CC me, I am on the list :-)