Sven Brauch svenbrauch at
Fri Oct 15 22:25:11 UTC 2010

That indeed looks like a good idea.
I read the parser documentation and generated an XML file out of the AST the
parser provides; for this script , it looks like
I'd now read that into the C++ AST (this will require quite some changes in
class structures, as the AST this parser provides is much more clear and
detailed; for example, instead of "Identifier", it explicitly gives things
like "FunctionDef", "Attribute" and "Name", and even tells wether the
attribute is being written or read. Nice!). Do you think that's a bearable
plan (PovAddict already disliked the idea of using XML for this...)?

Best regards,

2010/10/15 Sven Brauch <svenbrauch at>

> Thanks for your answers.
> Hmm, actually that would definitely make sense, less code to maintain, and
> it'd probably be faster, too... I'll have a look at that before I try to fix
> the other problem (there's several others, too, and some of them would
> probably take much debugging work)...
> Best regards,
> Sven
> 2010/10/15 Aleix Pol <aleixpol at>
> On Fri, Oct 15, 2010 at 6:11 PM, Sven Brauch <svenbrauch at>wrote:
>>> Hey!
>>> I'm using kdevelop-pg-qt to generate a parser for python. In
>>> astvisitor.cpp (
>>> although the type of the AST is definitely BooleanNotOperation, it steps
>>> into visitBooleanAndOperation for code like "if not foo: pass" and then
>>> segfaults. Can anyone tell me why it steps into visitBooleanAndOperation and
>>> not into visitBooleanNotOperation? I really can't figure that out...
>>> Thanks,
>>> Sven
>>> --
>>> KDevelop-devel mailing list
>>> KDevelop-devel at
>> Are you sure you can't use the parser provided by libpython?
>> That would make sense...
>> --
>> KDevelop-devel mailing list
>> KDevelop-devel at
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the KDevelop-devel mailing list