It enables clang to compile Mingw's headers.

llvm-svn: 116184

llvm-svn: 116183

  - lib/gcc/include should not be included. Clang oughta have alternatives. (PR7956)
  - Cygwin: /usr/include/w32api should be included. gcc/cygwin does.
  - gcc/cygwin uses gcc paths as /usr/lib not /lib. They are same on Cygwin environment, though, We have to take what gcc/cygwin does.
  - Cygwin-1.7: Use 4.3.4.
  - Cygwin-1.5: Use 4.3.2 for gcc-4.

llvm-svn: 116182

forward declarations).

llvm-svn: 116181

llvm-svn: 116180

what gdb uses)) so we can tell our "lldb" driver program to not automatically
parse any .lldbinit files. 

llvm-svn: 116179

lldb_private::RegularExpression compiles and matches with:

    size_t
    RegularExpression::GetErrorAsCString (char *err_str, 
                                          size_t err_str_max_len) const;
    
Added the ability to search a variable list for variables whose names match
a regular expression:

    size_t
    VariableList::AppendVariablesIfUnique (const RegularExpression& regex, 
                                           VariableList &var_list, 
                                           size_t& total_matches);


Also added the ability to append a variable to a VariableList only if it is 
not already in the list:

    bool
    VariableList::AddVariableIfUnique (const lldb::VariableSP &var_sp);

Cleaned up the "frame variable" command:
- Removed the "-n NAME" option as this is the default way for the command to
  work.
- Enable uniqued regex searches on variable names by fixing the "--regex RE"
  command to work correctly. It will match all variables that match any
  regular expressions and only print each variable the first time it matches.
- Fixed the option type for the "--regex" command to by eArgTypeRegularExpression
  instead of eArgTypeCount

llvm-svn: 116178

llvm-svn: 116177

Added frame relative frame selection to "frame select". You can now select
frames relative to the current frame (which defaults to zero if the current
frame hasn't yet been set for a thread):

The gdb "up" command can be done as:
(lldb) frame select -r 1
The gdb "down" command can be done as:
(lldb) frame select -r -1

Place the following in your ~/.lldbinit file for "up" and "down":

command alias up frame select -r 1
command alias down frame select -r -1

llvm-svn: 116176

llvm-svn: 116175

llvm-svn: 116174

llvm-svn: 116173

as binary bytes or as an ASCII text dump. 
- The output file is specified with the  "--outfile FILE" option.
- The memory can be appended to an existing file using the "--append" option. 
- The memory will be written as an ASCII text dump by default, or as 
  binary with the "--binary" option. 

Added new options to memory write to allow writing all or part of
a file on disk to target memory:
- The input file is specified using the "--infile FILE" option
- The offset at which to start in the file defaults to zero, but
  can be overridden using the "--offset OFFSET" option. If the
  size is not specified, the remaining number of bytes in the file
  will be used as the default byte size.
- The number of bytes to write defaults to the entire file byte
  size, but can be changed with the "--size COUNT" option.
 

llvm-svn: 116172

with the function name 'lldb_iter'.  Example:

def disassemble_instructions (insts):
    from lldbutil import lldb_iter
    for i in lldb_iter(insts, 'GetSize', 'GetInstructionAtIndex'):
        print i

llvm-svn: 116171

llvm-svn: 116170

truncate the stack pointer to 32-bits on a 64-bit machine.

llvm-svn: 116169

alignment for PPC32/64, avoiding some masking operations.

llvm-gcc expands vaarg inline instead of using the instruction
so it has never hit this.

llvm-svn: 116168

llvm-svn: 116167

llvm-svn: 116166

by Pierre Habouzit!

llvm-svn: 116165

llvm-svn: 116164

Lei Zhang.

llvm-svn: 116163

llvm-svn: 116162

llvm-svn: 116161

Now using a variant of the existing inlining heuristics to decide whether to create a given specialization of a function in PartialSpecialization.  If the total performance bonus across all callsites passing the same constant exceeds the specialization cost, we create the specialization.

llvm-svn: 116158

llvm-svn: 116157

llvm-svn: 116156

llvm-svn: 116155

llvm-svn: 116154

Store pointer and length of the message in DelayedDiagnostic and hide the gory union details.

llvm-svn: 116153

llvm-svn: 116152

llvm-svn: 116151

llvm-svn: 116150

llvm-svn: 116149

llvm-svn: 116148

  __builtin_ia32_vec_init_v8qi
  __builtin_ia32_vec_init_v4hi
  __builtin_ia32_vec_init_v2si

They are lowered to bitcasts. (These are all ready tested by the gcc testsuite.)
<rdar://problem/8529957>

llvm-svn: 116147

llvm-svn: 116146

llvm-svn: 116145

llvm-svn: 116144

llvm-svn: 116143