[llvm-commits] [llvm] r114684 - /llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp

[Benjamin Kramer]

On 23.09.2010, at 21:34, Jim Grosbach wrote:

> 
> On Sep 23, 2010, at 11:57 AM, Benjamin Kramer wrote:
> 
>> Author: d0k
>> Date: Thu Sep 23 13:57:26 2010
>> New Revision: 114684
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=114684&view=rev
>> Log:
>> Unbreak build. Jim, please review.
>> 
>> Modified:
>>   llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp
>> 
>> Modified: llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp
>> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp?rev=114684&r1=114683&r2=114684&view=diff
>> ==============================================================================
>> --- llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp (original)
>> +++ llvm/trunk/lib/Target/ARM/ARMAsmPrinter.cpp Thu Sep 23 13:57:26 2010
>> @@ -1216,7 +1216,7 @@
>>  // FIXME: This really should be in AsmPrinter/ARMInstPrinter.cpp, not here.
>>  //        Need a way to ask "isTargetDarwin()" there, first, though.
>>  if (MI->getOpcode() == ARM::TRAP && !Subtarget->isTargetDarwin()) {
>> -    OS << "\t.long\t2147348462\t\t" << MAI->getCommentString() << "trap";
>> +    OS << "\t.long\t3892305662\t\t" << MAI->getCommentString() << "trap";
>>  } else if (MI->getOpcode() == ARM::tTRAP && !Subtarget->isTargetDarwin()) {
>>    OS << "\t.short\t57086\t\t\t" << MAI->getCommentString() << " trap";
>>  } else
>> @@ -1729,7 +1729,7 @@
>>    // FIXME: Remove this special case when they do.
>>    if (!Subtarget->isTargetDarwin()) {
>>      //.long 0xe7ffdefe ${:comment} trap
>> -      uint32_t Val = 0xe7ffdefee;
>> +      uint32_t Val = 0xe7ffdefe;
> 
> The value was correct before. Clang is just being picky about the constant type. Not sure why I can't reproduce locally; I am building with clang. Odd. From the bot emails, looks like it only issues that diagnostic for this line on non-Darwin hosts?
> 
> Fixed the value in r114687, with an explicit "UL" suffix on the constant this time. Hopefully the bots will like that...

You had 3 different values in your code.

1. 0x7ffdefee (.long\t2147348462)
2. 0xe7ffdefe (in the comment, which I believe is the correct one)
3. 0xe7ffdefee (the buildbots complained about this one, doesn't even fit in 32 bits)

You reintroduced #3 in r114687. I don't think that's right, it will truncate to #1.