Leon Messerschmidt [SMTP:[EMAIL PROTECTED]] wrote :

> [geir wrote : ]
>> Sure.  To be precise, the #macro()'s are nothing more than template code,
so
>> I don't quite see why they are worse than using #include.

> What I'm wary of here is that people would start using #include to import
> #macros.  #macro directives would then begin to be treated as black-box
> implementations.  If a bug is detected you will need to debug both the
> #included stuff and your java code.  I'm yet to think of a difficult to
fix
> #macro bug :-)  Maybe I'm just paranoid here...

A few things :

1)  You can shut that down with current proposed implementation.  There is a
property to turn of 'inline', which  includes both explicit #macro()
defintions in the template, as well as #macros in #included templates.
Maybe it's misnamed, in that it should be clarified in the property name
that it really controls any attempts at a #macro() definition after the
engine loads the central library #macro() VMs, if any.

2) Theoretically, I think that you can hang yourself on #include as well, as
you can also chain more templates together with #includes w/in the included
template, in theory.

I think that the central library idea allows the 'black box' implementation,
if I understand you correctly.  I personally don't think that is a bad thing
as long as administration process is sound, and again, it can be turned off.

Paranoia is good.

geir


Reply via email to