[cfe-dev] Precompiled headers

[Russell Wallace via cfe-dev]

That sounds good. After all, when creating a precompiled header file, you
don't even want an object file at that time anyway, and if the precompiled
header fails to build, you expect to have to stop and fix it before you can
usefully do anything else.

On Fri, Dec 11, 2015 at 7:35 PM, Nico Weber via cfe-dev <
cfe-dev at lists.llvm.org> wrote:

> On Fri, Dec 11, 2015 at 1:42 PM, Reid Kleckner via cfe-dev <
> cfe-dev at lists.llvm.org> wrote:
>
>> On Fri, Dec 11, 2015 at 10:26 AM, Richard Smith via cfe-dev <
>> cfe-dev at lists.llvm.org> wrote:
>>
>>> -Xclang passes only the immediately following argument to cc1. You need
>>>
>>> clang-cl -fms-compatibility-version=19 -Xclang -include-pch -Xclang
>>> a.h.gch a.cpp
>>>
>>> Also, the -Xclang / -cc1 interface is not really supported and has no
>>> stability guarantees; do the cl.exe flags for PCH not work in clang-cl? (If
>>> not, please file a bug on that.)
>>>
>> Nico was going to take a shot at making the /Y preprocessor flags work
>> with clang-cl, but I don't think he got around to it. I think it's not as
>> easy as adding more aliases, you have to add some driver logic.
>>
>
> I actually have a pretty complete patch that seems to work modulo testing
> and error handling (the most significant problem is that if your
> precompiled header fails to build, the driver currently still goes ahead
> and tries to use the (non-existent) gch file to produce the .obj file in
> /Yc mode).
>
>
>>
>> _______________________________________________
>> cfe-dev mailing list
>> cfe-dev at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
>>
>>
>
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20151211/5269eb84/attachment.html>