Use root non-virtual primary base class, not just immediate primary base class, for AT_containing_type.
This is tested by virtfunc.exp in gdb testsuite.

llvm-svn: 116535

and let the ARMExpandPseudoInsts pass fix them up into the real (MOVs)
instruction form.

llvm-svn: 116534

llvm-svn: 116533

debug information and you evaluated an expression, a crash would occur as a
result of an unchecked pointer.

Added the ability to get the expression path for a ValueObject. For a rectangle
point child "x" the expression path would be something like: "rect.top_left.x".
This will allow GUI and command lines to get ahold of the expression path for
a value object without having to explicitly know about the hierarchy. This
means the ValueObject base class now has a "ValueObject *m_parent;" member.
All ValueObject subclasses now correctly track their lineage and are able
to provide value expression paths as well.

Added a new "--flat" option to the "frame variable" to allow for flat variable
output. An example of the current and new outputs:

(lldb) frame variable 
argc = 1
argv = 0x00007fff5fbffe80
pt = {
  x = 2
  y = 3
}
rect = {
  bottom_left = {
    x = 1
    y = 2
  }
  top_right = {
    x = 3
    y = 4
  }
}
(lldb) frame variable --flat 
argc = 1
argv = 0x00007fff5fbffe80
pt.x = 2
pt.y = 3
rect.bottom_left.x = 1
rect.bottom_left.y = 2
rect.top_right.x = 3
rect.top_right.y = 4


As you can see when there is a lot of hierarchy it can help flatten things out.
Also if you want to use a member in an expression, you can copy the text from
the "--flat" output and not have to piece it together manually. This can help
when you want to use parts of the STL in expressions:

(lldb) frame variable --flat
argc = 1
argv = 0x00007fff5fbffea8
hello_world._M_dataplus._M_p = 0x0000000000000000
(lldb) expr hello_world._M_dataplus._M_p[0] == '\0'

llvm-svn: 116532

to CodeGenOption flags.

llvm-svn: 116530

llvm-svn: 116529

identifiers to determine good typo-correction candidates. Once we've
identified those candidates, we perform name lookup on each of them
and the consider the results. 

This optimization makes typo correction > 2x faster on a benchmark
example using a single typo (NSstring) in a tiny file that includes
Cocoa.h from a precompiled header, since we are deserializing far less
information now during typo correction.

There is a semantic change here, which is interesting. The presence of
a similarly-named entity that is not visible can now affect typo
correction. This is both good (you won't get weird corrections if the
thing you wanted isn't in scope) and bad (you won't get good
corrections if there is a similarly-named-but-completely-unrelated
thing). Time will tell whether it was a good choice or not.

llvm-svn: 116528

llvm-svn: 116527

llvm-svn: 116524

llvm-svn: 116522

It's more stress-free without access checking though..

llvm-svn: 116521

if no update commands are specified it just lists the current values, and show that
it always shows the new values for a signal after it has been updated.  Also updated
the help text to match the new functionality.

llvm-svn: 116520

llvm-svn: 116519

validating the variables set on the make command line.

llvm-svn: 116517

llvm-svn: 116516

llvm-svn: 116515

The failing was due to this:
1. preamble.c contains CR+LF new lines
2. write() is called with a buffer containing the original (CR+LF) to output the result on the console.
3. In text mode(the default), write() convert LF to CR+LF even if LF is preceded by CR, hence we have CR+CR+LF which filecheck interprets as 2 lines.

llvm-svn: 116513

pseudonym.

llvm-svn: 116512

solely based on the names it sees, rather than actual declarations it
gets. In essence, we determine the set of names that are "close
enough" to the typo'd name. Then, we perform name lookup for each of
those names, filtering out those that aren't actually visible, and
typo-correct from the remaining results.

Overall, there isn't much of a change in the behavior of typo
correction here. The only test-suite change comes from the fact that
we make good on our promise to require that the user type 3 characters
for each 1 character corrected. 

The real intent behind this change is to set the stage for an
optimization to typo correction (so that we don't need to deserialize
all declarations in a translation unit) and future work in finding
missing qualification ("'vector' isn't in scope; did you mean
'std::vector'?). Plus, the code is cleaner this way.

llvm-svn: 116511

Always use binary mode for output stream. This is important to prevent unwanted end of line conversion on Windows. Should not affect Unix where O_BINARY is not defined. This fix /clang/test/lexer/preamble.c XFAIL on WIN32.

llvm-svn: 116509

Store in PCH the key function of C++ class to avoid deserializing the complete declaration context in order to compute it.
Progress for rdar://7260160.

llvm-svn: 116508

instead of deserializing the complete declaration context of the record.

Iterating over the fields of a record is very common (e.g to determine the layout), unfortunately we needlessly deserialize every declaration
that the declaration context of the record contains; this can be bad for large C++ classes that contain a lot of methods.
Fix this by allow deserialization of just the fields when we want to iterate over them.
Progress for rdar://7260160.

llvm-svn: 116507

llvm-svn: 116506

and emits an error if a declaration with this name is deserialized from PCH.

This is for testing, to make sure that we don't deserialize stuff needlessly.

llvm-svn: 116505

Don't add the injected class name to the redeclarations chain; the chain should contain actual redeclarations, not implicits.
As a bonus, now we don't deserialize it unless we need it.

llvm-svn: 116504

Introduce command line option -dump-deserialized-decls which is used to print the PCH decls that got deserialized, for testing purposes.

llvm-svn: 116503

llvm-svn: 116501

llvm-svn: 116500

llvm-svn: 116498

Before we would also split around a loop if any peripheral block had multiple
uses. This could cause repeated splitting when splitting a different live range
would insert uses into the periphery.

Now -spiller=inline passes the nightly test suite again.

llvm-svn: 116494

llvm-svn: 116493

Try again at implementing thread-safe lazy pass initialization, without depending on static local initialization
being threadsafe AND ensuring that initialization is complete by the time the initializeFooPass method returns.

llvm-svn: 116492

llvm-svn: 116491

llvm-svn: 116490

Apparently MSVC doesn't support thread-safe static local initialization.  Roll our own solution instead.

llvm-svn: 116489

llvm-svn: 116488

llvm-svn: 116487

llvm-svn: 116486

Inside the lldb module, there's no need (and as a matter of fact, incorrect) to specify the 'lldb'
module name.

Comment out the call to lldb.SBDebugger.Initialize() within the test driver itself, since it is
already done when we import the lldb.py module.

llvm-svn: 116485

llvm-svn: 116484