[cfe-dev] libclang additions, how to integrate?

Manuel Klimek klimek at google.com
Sun Dec 21 01:54:27 PST 2014


On Sun Dec 21 2014 at 3:44:26 AM mobi phil <mobi at mobiphil.com> wrote:

> still fighting the following inconsistencies:
>
> wanted to reuse CXCursor to present to the libclang client the ast matches
> in form of cursor. I see that it is impossible to create cxcursor from
> Type, TypeLoc, NestedNameSpecifer etc. to cover all matcher result types.
> Were this intentionally left out, or still missing sthg. from the full
> picture?
>

Usually the answer to questions like this is "nobody had time to
design/implement it yet" :)
Often what is left out is a bit of a challenge to design though (but I
don't know whether that's the case here...)


> The usual idea is to first write a little "design" email to cfe-dev,
>> outlining what you're trying to do, what the interfaces would look like,
>> etc.
>>
>> Note that libclang is considered to be a stable interface, so it's very
>> important to put some thought into designing new features - we'll be stuck
>> with them for a while...
>>
>> On Sun Dec 14 2014 at 9:00:54 PM mobi phil <mobi at mobiphil.com> wrote:
>>
>>> Hi,
>>>
>>> added in my local repo some libclang functions for:
>>>
>>> -> ast matcher completion
>>> -> ast matcher
>>>
>>> they do the basic things, but before bringing them to "standard" (some
>>> additional code) would like to know what are the conditions/procedure to
>>> follow
>>>
>>> any python/liblang developer around to discuss some details?
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20141221/8e93498a/attachment.html>


More information about the cfe-dev mailing list