Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-11-01 Thread David Tweed
...@cs.uiuc.edu] On Behalf Of Rafael Espíndola Sent: 30 October 2012 03:58 To: llvm cfe Subject: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum There is a lot of char* being passed around in the clang driver. At least part of that seems to be because we have an enum (llvm::Triple::ArchType) that can

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-11-01 Thread Matthieu Monrocq
cfe Subject: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum There is a lot of char* being passed around in the clang driver. At least part of that seems to be because we have an enum (llvm::Triple::ArchType) that can distinguish an ARM from an X86, but we don't have an enum that can

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-11-01 Thread David Tweed
. Cheers, Dave -Original Message- From: cfe-commits-boun...@cs.uiuc.edu [mailto:cfe-commits-boun...@cs.uiuc.edu] On Behalf Of Rafael Espíndola Sent: 30 October 2012 03:58 To: llvm cfe Subject: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum There is a lot of char* being passed around

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-11-01 Thread Matthieu Monrocq
On Thu, Nov 1, 2012 at 1:39 PM, David Tweed david.tw...@arm.com wrote: This looks like a reasonable goal, but obviously the ARM architecture being quite modular there are a lot more actual variants (and a huge number more potential variants) than for other CPUs (and those listed in the

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-31 Thread Daniel Dunbar
I think very little of the char* handling in the driver is related to this particular issue. I'm hesitant to promote the very specialized Darwin CPU arch naming scheme higher in the API. This isn't really a useful scheme (i.e., no one else should use this notion), but is required for Darwin

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-31 Thread Rafael Espíndola
On 31 October 2012 13:55, Daniel Dunbar dan...@zuster.org wrote: I think very little of the char* handling in the driver is related to this particular issue. I'm hesitant to promote the very specialized Darwin CPU arch naming scheme higher in the API. This isn't really a useful scheme (i.e.,

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-31 Thread Daniel Dunbar
Yes, but still seems to add API and code around something that isn't worth promoting: -- diffstat t.patch ToolChains.cpp | 100 +++-- Tools.cpp | 97 +-- Tools.h| 34

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-31 Thread Rafael Espíndola
On 31 October 2012 19:26, Daniel Dunbar dan...@zuster.org wrote: Yes, but still seems to add API and code around something that isn't worth promoting: Well, my idea is to propagate it out of TranslateArgs into the callers. Doing that we reach cases where we have char* as keys in densmaps, which

Re: [cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-31 Thread Rafael Espíndola
Well, my idea is to propagate it out of TranslateArgs into the callers. Doing that we reach cases where we have char* as keys in densmaps, which is *really* scary and where I actually started looking at this. I will just sent a larger patch to better illustrate the impact tonight or

[cfe-commits] [patch] Add a llvm::Triple::CPUArch enum

2012-10-29 Thread Rafael Espíndola
There is a lot of char* being passed around in the clang driver. At least part of that seems to be because we have an enum (llvm::Triple::ArchType) that can distinguish an ARM from an X86, but we don't have an enum that can distinguish a ARMv6 form an ARMv7 for example. The attached patches add