Just to be contrarian (but not really), the "[EMAIL PROTECTED]" notation looks weird to me! "@{x}" is familiar enough, although I can't say why at the moment -- oh, yeah, doesn't Perl have a similar construct?
I've watched this discussion all the way through, and I can see the benefits of both approaches. FWIW, seems to me that a run-time definition of a property within the macro (<local> rears its ugly(?!) head again) is desirable. Although a straight textual substitution will be easily understood by folks familiar with the C/C++ pre-processor.
I feel strongly both ways! :^/
Ken
At 10:11 2003-11-19, you wrote:
> From: Jose Alberto Fernandez [mailto:[EMAIL PROTECTED] > > From: Gus Heck [mailto:[EMAIL PROTECTED] > > My (non-committer) oppion coincides with Stefan here, with a slight > > preference for @{x} > > because it looks like "put the substitution AT this location" when I > > read it to myself. > > > > Actually if we go for reading value, the advantage of @{x} notation is > that sounds like "AT(tribute) x" :-) > > I think I can live with that.
Unlike Jose Alberto, I think it's a 'good' thing than referencing an declared attribute of a <macrodef> in its body/impl resembles the XSLT referencing of a attribute of the current XML element!
The similarities are striking, and the syntax is well known and clearly documented. The <macrodef> attribute *will* be an XML element attribute when it's used actually!!!
[EMAIL PROTECTED] feels very natural, and avoids any confusion with ${x}. It can be easily escaped using the double symbol people like, so that {@@x} passes thru as the [EMAIL PROTECTED] literal. (After all, I don't think it's valid to have an XML attribute starting with an @, so it's free of conflict too.)
The point is not to resemble the existing notation for dereferencing Ant properties, since that's what it's supposed to be distinct from, which is why @{x} feels wrong to me (and looks ugly IMHO ;-).
The point is to use a widely used notation for a widely similar purpose, i.e. the XSLT notation, which as I noted above is so similar to the semantic of what's being done.
I'm not a committer and all, but to me [EMAIL PROTECTED] is the clear choice for <macrodef> attribute dereferencing. I'm sure others will disagree ;-) But no one can escape getting my opinion on the matter ;-)))) --DD
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
============================================================= J. Kenneth Gentle (Ken) | Phone: (610) 255-0361 Gentle Software, LLC | Email: [EMAIL PROTECTED] =============================================================
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
