On Jan 8, 2007, at 11:08 AM, Devang Patel wrote:
On Jan 8, 2007, at 11:05 AM, Reid Spencer wrote:
OK. I suspect this may not be straight forward. Let's see.
-
Devang
config.gcc selects target specific C source file name based on
cpu_type (by default). So, I thought it may not be super
On Jan 8, 2007, at 11:05 AM, Reid Spencer wrote:
OK. I suspect this may not be straight forward. Let's see.
-
Devang
config.gcc selects target specific C source file name based on
cpu_type (by default). So, I thought it may not be super easy to trace
all uses of i386.c.
Your patch sh
The attached patch worked for me on x86-linux and for Chandler on
x86-64-linux. This is a delta from r240 of the mirror .. YMMV.
Reid.
On Mon, 2007-01-08 at 10:51 -0800, Devang Patel wrote:
> On Jan 8, 2007, at 10:45 AM, Chris Lattner wrote:
>
> >> After quick look, I think makefiles are ready.
On Jan 8, 2007, at 10:45 AM, Chris Lattner wrote:
>> After quick look, I think makefiles are ready. However, if any target
>> relies on llvm-i386.cpp to provide these undefined symbols then
>> config.gcc needs to be updated to inform this to makefiles for such
>> targets. Each target can use the
> After quick look, I think makefiles are ready. However, if any target
> relies on llvm-i386.cpp to provide these undefined symbols then
> config.gcc needs to be updated to inform this to makefiles for such
> targets. Each target can use their own target specific C++ source
> file, no need to put
On Jan 7, 2007, at 6:24 PM, Reid Spencer wrote:
> I wish I could retract email.
>
> The patch I just sent didn't have differences for sub-directories in
> it.
> Please don't apply it. Instead use this one which should bring you
> up to
> date with Apple's changes. Again, this applies to r240
