> From: Peter Reilly [mailto:[EMAIL PROTECTED]
>
> Jose Alberto Fernandez wrote:
>
> >As per other approaches to local properties, unless we go
> and define a
> >real semantic for them (like any other well design
> programming language
> >out there) I see they creating more problems than solutions, in
> >particular in our BC constrained world.
> >
> >
> I do not think that the problems are too bad (but I would say that!).
>
> I would like to have the thread-local property
> implementation, however there was a lot of discussions about syntax.
>
How about haing the two, and see which one works better? As I said
before <let/> is a very small change to <macrodef/> with no other
dependencies.
> As the main use case for local properties is <macrodef> we
> could just implement them for macrodefs, and if necessary
> extend them later to be the free style properties.
>
> So the syntax would be:
> <macrodef name="show_length">
> <attribute name="filename">
> <local-property name="local-prop"/>
> <sequential>
> <length file="@{filename}" property="${local-prop}"/>
> <echo>The length of file '@{filename}' is ${local-prop}</echo>
> </sequential>
> </macrodef>
>
In your example above, are you sure you mean 'property="${local-prop}"'
and not 'property="local-prop"'?
We could write this with let as:
<macrodef name="show_length">
<attribute name="filename">
<let name="local-prop"/>
<sequential>
<length file="@{filename}" property="@{local-prop}"/>
<echo>The length of file '@{filename}' is [EMAIL PROTECTED]</echo>
</sequential>
</macrodef>
And in this case the scope of "@{local-prop}" is well defined as
the body of "show-length".
>
> We could implement this as a trial in ant cvs and pull it if
> there are
> too many issues.
>
> Peter
Wouldn't mind having the two, and see what works best. :-)
Jose Alberto
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
