ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <dona...@apache.org>
Subject RE: TASTE QUESTION: optional description attribute for ant projects?
Date Thu, 12 Apr 2001 00:42:15 GMT
Hi,

At 10:25  12/4/01 +1000, Conor MacNeill wrote:
>> However more importantly is the fact that we will eventually want to doc
>> targets or tasks etc.
>
>Agreed. I think, however, that we should be using XML by allowing a <doc> or
><description> sub-elements to each of these elements (project and target,
>maybe tasks). Whether the content of that element has any structure (@link,
>etc) is, for me, a separate issue.

Right - but sure enough as soon as you add doc elements there will be
people asking for link sub-elements - I mean we are already using XML for
describing docs - why not use it "properly". After we have link
sub-elements we will have a whole host of other elements for doc elements. 

>> It would be better if there was a consistent doccing
>> standard.
>
>Agreed - I just don't think we should use XML comments for that.

Partially agree - I think it should be namespaced elements - but thats not
viable for ant1.x.

>> To achieve this you can easily and transparently use comments or
>> alternatively use magic element names throughout the build.xml.
>
>Again, I see nothing magical about these elements - we are just giving some
>more structure to the build file - defining its content model, if you like.

Sure they are magical - they completely go against the whole structure of
tasks and targets. Tasks no longer are directly mapped onto their task
objects and targets can now contain more than just tasks.

And this is assuming we don't end up adding extra other magic sub-elements.
(DBC for Ant could end up using another sub-element etc).

>What is the advantage of making this "transparent"?

Transparent is good as then we will not mess with our current users. They
have an expectation that if there is a sub-element foo that the
coresponding createFoo() or addFoo() will be called. Some may have even
defined tasks that use the name foo for their sub-element name (of course
replace foo with whatever we call this magic element).

Cheers,

Pete

*-----------------------------------------------------*
| "Faced with the choice between changing one's mind, |
| and proving that there is no need to do so - almost |
| everyone gets busy on the proof."                   |
|              - John Kenneth Galbraith               |
*-----------------------------------------------------*


Mime
View raw message