llvm-svn: 187872

llvm-svn: 187870

This value may be used uninitialized in SIInsertWaits::insertWait.
Found with MemorySanitizer.

llvm-svn: 187869

lld has a hashtable with StringRef keys; it needs to iterate over the keys in
*insertion* order.  This is currently implemented as std::vector<StringRef> +
DenseMap<StringRef, T>.  This will probably need a proper
DenseMapInfo<StringRef> if we don't want to lose memory/performance by
migrating to a different data structure.

llvm-svn: 187868

llvm-svn: 187867

for StringRef with a StringMap

The bug is that the empty key compares equal to the tombstone key.

Also added an assertion to DenseMap to catch similar bugs in future.

llvm-svn: 187866

As of this revision, all functions of LTOCodeGenerator are consistent in
ret-true-on-succ.

Tested on multiple OSes.

llvm-svn: 187864

llvm-svn: 187863

llvm-svn: 187862

llvm-svn: 187861

llvm-svn: 187860

llvm-svn: 187859

llvm-svn: 187858

llvm-svn: 187857

llvm-svn: 187856

llvm-svn: 187855

 - Since we only have a few of these, use the cumbersome method of getting the
   exception object from 'sys' to retain the current pre-2.6 compatibility.

llvm-svn: 187854

llvm-svn: 187853

llvm-svn: 187852

llvm-svn: 187851

llvm-svn: 187850

llvm-svn: 187849

llvm-svn: 187848

llvm-svn: 187847

This fixes a bug from r187826.

Reviewers: hans

Differential Revision: http://llvm-reviews.chandlerc.com/D1304

llvm-svn: 187846

One use needs to copy the alloca into a std::string, and the other use
is before calling CreateProcess, which is very heavyweight anyway.

llvm-svn: 187845

Use it when we're looking for a string in particular. Update comments
as well.

llvm-svn: 187844

No functional change.

llvm-svn: 187843

columnWidth), and partially disable tests on MACOSX (which handles some
characters differently).

llvm-svn: 187842

Previously this check was guarded by MSVC, which doesn't distinguish
between the compiler and the headers/library.  This enables clang to
compile more of LLVM on Windows with Microsoft headers.

Remove some unused macros while I'm here: error_t and LTDL stuff.

llvm-svn: 187839

llvm-svn: 187838

Implemented llvm::sys::locale::columnWidth and isPrint for the case of generic UTF8-capable terminal.

Summary:
This is a second attempt to get this right. After reading the Unicode
Standard I came up with the code that uses definitions of "printable" and
"column width" more suitable for terminal output (i.e. fixed-width fonts and
special treatment of many control characters).
The implementation here can probably be used for Windows and MacOS if someone
can test it properly.
The patch addresses PR14910.

Reviewers: jordan_rose, gribozavr

CC: llvm-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D1253

llvm-svn: 187837

Replacing /GR with /GR- instead of applying both options to the project.  This should reduce some build bot warnings (D9025: "overriding '/GR' with '/GR-'").

llvm-svn: 187836

llvm-svn: 187834

llvm-svn: 187832

Since the VSrc_* register classes contain both VGPRs and SGPRs, copies
that used be emitted by isel like this:

SGPR = COPY VGPR

Will now be emitted like this:

VSrC = COPY VGPR

This patch also adds a pass that tries to identify and fix situations where
a VGPR to SGPR copy may occur.  Hopefully, these changes will make it
impossible for the compiler to generate illegal VGPR to SGPR copies.

llvm-svn: 187831

Also factor out the register class lookup to its own function.

llvm-svn: 187830

[NVPTX] We dont have any target specific flags yet for generating symbol references, so get rid of the default-only switch statement.  Fixes an MSVC warning.

llvm-svn: 187829

instructions defined in MipsInstrInfo.td as codegen-only instructions.

llvm-svn: 187828

The globals being generated here were given the 'private' linkage type. However,
this caused them to end up in different sections with the wrong prefix. E.g.,
they would be in the __TEXT,__const section with an 'L' prefix instead of an 'l'
(lowercase ell) prefix.

The problem is that the linker will eat a literal label with 'L'. If a weak
symbol is then placed into the __TEXT,__const section near that literal, then it
cannot distinguish between the literal and the weak symbol.

Part of the problems here was introduced because the address sanitizer converted
some C strings into constant initializers with trailing nuls. (Thus putting them
in the __const section with the wrong prefix.) The others were variables that
the address sanitizer created but simply had the wrong linkage type.

llvm-svn: 187827