[cfe-dev] Default header search paths not included when compiling from libclang?

[Manuel Klimek]

+chandler, as he knows a lot about the driver...

On Wed, Nov 20, 2013 at 1:52 AM, Val Markovic <val at markovic.io> wrote:

> From what I can gather, it appears that when compiling a C++ file with the
> clang binary, clang will use some default header search paths. But when
> trying to do the same from libclang, it won't. Is this caused by the clang
> driver/frontend distinction (I don't know much about it)?
>
> Specifically, I'm talking about this issue affecting the YouCompleteMe
> plugin: https://github.com/Valloric/YouCompleteMe/issues/303
>
> So, is there a way to get libclang to use the same header search paths
> that the clang binary does? If not, why not and could this be changed in
> future versions of libclang?
>
> If there's no way to get libclang to use the same header search paths by
> default, is there a way to extract them out of the libclang API so that the
> paths can then be provided in other libclang API calls? If not now, could
> libclang be changed to support this in the future?
>
> The current situation where libclang users are forced to recreate the
> default header search path logic for every OS is incredibly frustrating.
> YCM can't even call out to clang to extract the headers out of it with
> "echo | clang -v -E -x c++ -" because clang might not even be installed on
> the machine. We only have libclang to work with.
>

I guess it would make sense to add some way to query the driver logic to
libclang.
What I don't know enough about is relative to which "clang" version the
driver should search in that case - perhaps the right solution is to either
hand one in from the outside, or use the unix search path?

Cheers,
/Manuel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20131128/808e17fa/attachment.html>