Le 06/02/2012 00:48, lkcl luke a écrit :
> rright.
> lkcl@teenymac:~/pyjamas/pyjs/src/pyjs$ diff -uwbB
> translator_proto-KEES.py translator_proto-KLUEV.py> f
>
> vi f
> .....
> .....
>
> @@ -2255,7 +2272,9 @@
> else:
> fail = ''
> self.w( self.spacing() + "if (!( " + expr + " )) {")
> - self.w( self.spacing() + " throw @{{AssertionError}}(%s);" % fail)
> + self.w( self.spacing() + " throw %(AssertionError)s(%(fail)s);" %
> + dict(AssertionError=self.pyjslib_name('AssertionError'),
> + fail=fail))
> self.w( self.spacing() + " }")
>
> def _return(self, node, current_klass):
> @@ -2338,7 +2357,7 @@
>
>
> ok. right. %(AssertionError)s.... that's a python string/dict
> formatting thing. "%(AssertionError)" % {'AssertionError': 'hello'} =>
> "hello"
>
> iiinteresting. it looks like daniel decided to use python formatting,
> here. perhaps to reduce space, perhaps to optimise pricessing
> requirements, i don't know - but it looks like it's ok.
>
> i wish he'd bloody well discussed this stuff. that was the whole
> point of gsoc. *sigh*. hey daniel, have you dropped off the face of
> the planet? we're missing you! :)
>
> l.
>
IIRC, in then I left everything as @{{}} while merging, and just put the
two base versions aside in case string formatting actually had an
interest over template vars.
Maybe Daniel had some long term idea in the head concerning
autocompilation of pyjs to javascript (parser included), but I can't see
what.
Regards,
PKL
