I noticed that the existing RUN test cases all kept x86_64-apple-darwin10 even in the case of i386. I can't claim to understand why, I merely (unintelligently) followed the pattern with minimal changes.
I rather liked the simple commands we tested offline yesterday:

e.g.:
clang -### -target powerpc-apple-darwin8 foo.m -o foo.o

as that was sufficient to see the difference in -fobjc-runtime=.
Feel free to amend/append the test as you see fit.

David

Hi David,

On 22 Jan 2014, at 08:00, David Fang wrote:

How about this addition to an existing testcase?  (attached)

+// ppc
+
+// RUN: %clang -target x86_64-apple-darwin10 -S -### %s \
+// RUN:   -arch ppc -mmacosx-version-min=10.4 2> %t
+// RUN: FileCheck --check-prefix CHECK-CHECK-PPC_OSX10_4 < %t %s

It's somewhat confusing to declare the target as x86_64- and then the arch as 
ppc.

Would there be some reason that " -target powerpc-apple-darwin8 " without the 
-arch and -mmacosx-version-min is not going to work here?
(after all we are only testing that the driver translates c/l args properly - 
there is no need to invoke a back end).

+// CHECK-CHECK-PPC_OSX10_4: "-cc1"
+// CHECK-CHECK-PPC_OSX10_4: -fobjc-runtime=macosx-fragile-10.4
+// CHECK-CHECK-PPC_OSX10_4-NOT: -fobjc-dispatch-method
+// CHECK-CHECK-PPC_OSX10_4: darwin-objc-defaults
+
+// RUN: %clang -target x86_64-apple-darwin10 -S -### %s \
+// RUN:   -arch ppc -mmacosx-version-min=10.5 2> %t
+// RUN: FileCheck --check-prefix CHECK-CHECK-PPC_OSX10_5 < %t %s

likewise " -target powerpc-apple-darwin9 "

+
+// CHECK-CHECK-PPC_OSX10_5: "-cc1"
+// CHECK-CHECK-PPC_OSX10_5: -fobjc-runtime=macosx-fragile-10.5
+// CHECK-CHECK-PPC_OSX10_5-NOT: -fobjc-dispatch-method
+// CHECK-CHECK-PPC_OSX10_5: darwin-objc-defaults

Iain

On 8 January 2014 12:10, Iain Sandoe <[email protected]> wrote:
bump?

Needs a testcase.

I know nothing about objc ABIs, but you and David Fang own darwin-ppc
these days, so LGTM with the test.

Cheers,
Rafael
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits


--
David Fang
http://www.csl.cornell.edu/~fang/
<objc-ABI.patch>


--
David Fang
http://www.csl.cornell.edu/~fang/

_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to