On Windows, character encoding of multibyte environment variable varies
depending on settings. The only reliable way to handle it I think is to use
GetEnvironmentVariableW().
GetEnvironmentVariableW() works on wchar_t string, which is on Windows UTF16
string. That's not ideal because we use UTF-8 as the internal encoding in LLVM.
This patch defines a wrapper function which takes and returns UTF-8 string for
GetEnvironmentVariableW().
The wrapper function does not do any conversion and just forwards the argument
to getenv() on Unix.
Differential Revision: http://llvm-reviews.chandlerc.com/D1612
llvm-svn: 190423
We try to create the scope children DIEs after we create the scope DIE. But
to avoid emitting empty lexical block DIE, we first check whether a scope
DIE is going to be null, then create the scope children if it is not null.
From the number of children, we decide whether to actually create the scope DIE.
This patch also removes an early exit which checks for a special condition.
It also removes deletion of un-used children DIEs that are generated
because we used to generate children DIEs before the scope DIE.
Deletion of un-used children DIEs may cause problem because we sometimes keep
created DIEs in a member variable of a CU.
llvm-svn: 190421
It was removed in r189130, but it turns out this makes life hard for
folks packaging LLVM and Clang and building the latter based on the
LLVM package.
Note that this only adds back the LLVM tblgen, and it's obviously
not included when LLVM_INSTALL_TOOLCHAIN_ONLY is set.
llvm-svn: 190419
Specialize the constructors for DIRef<DIScope> and DIRef<DIType> to make sure
the Value is indeed a scope ref and a type ref.
Use DIScopeRef for DIScope::getContext and DIType::getContext and use DITypeRef
for getContainingType and getClassType.
DIScope::generateRef now returns a DIScopeRef instead of a "Value *" for
readability and type safety.
llvm-svn: 190418
We were figuring out whether to use tPICADD or PICADD, then just using
tPICADD unconditionally anyway. Oops.
A testcase from someone familiar enough with ELF to produce one would
be appreciated. The existing PIC testcase correctly verifies the .s
generated, but that doesn't catch this bug, which only showed up in
direct-to-object mode.
http://llvm.org/bugs/show_bug.cgi?id=17180
llvm-svn: 190417
Dotfiles are impractical on Windows. This makes clang-format search
for the style configuration file as '_clang-format' in addition to
the usual '.clang-format'. This is similar to how VIM searches for
'_vimrc' on Windows.
Differential Revision: http://llvm-reviews.chandlerc.com/D1629
llvm-svn: 190413
LibXML2 config doesn't specify lzma as a dependency, which breaks
cross-compilation builds using new linkers (ld 2.21 or higher).
There is a bug on libxml2 to fix that, but since it's going to take
a while for things to go round and back, so we should have a harmless
addition of the library until then.
llvm-svn: 190409
Summary: This also unifies the handling of escaped newlines for all tokens.
Reviewers: djasper
Reviewed By: djasper
CC: cfe-commits, klimek
Differential Revision: http://llvm-reviews.chandlerc.com/D1638
llvm-svn: 190405
The main complication here is that TM and TMY (the memory forms) set
CC differently from the register forms. When the tested bits contain
some 0s and some 1s, the register forms set CC to 1 or 2 based on the
value the uppermost bit. The memory forms instead set CC to 1
regardless of the uppermost bit.
Until now, I've tried to make it so that a branch never tests for an
impossible CC value. E.g. NR only sets CC to 0 or 1, so branches on the
result will only test for 0 or 1. Originally I'd tried to do the same
thing for TM and TMY by using custom matching code in ISelDAGToDAG.
That ended up being very ugly though, and would have meant duplicating
some of the chain checks that the common isel code does.
I've therefore gone for the simpler alternative of adding an extra
operand to the TM DAG opcode to say whether a memory form would be OK.
This means that the inverse of a "TM;JE" is "TM;JNE" rather than the
more precise "TM;JNLE", just like the inverse of "TMLL;JE" is "TMLL;JNE".
I suppose that's arguably less confusing though...
llvm-svn: 190400
Summary:
This fixes various issues with mixed tabs and spaces handling, e.g.
when realigning block comments.
Reviewers: klimek, djasper
Reviewed By: djasper
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1608
llvm-svn: 190395
PR16752: 'mode' attribute for unusual targets doesn't work properly
Description:
Troubles could be happened due to some assumptions in handleModeAttr function (see SemaDeclAttr.cpp).
For example, it assumes that 32 bit integer is 'int', while it could be 16 bit only.
Instead of asking target: 'which type do you want to use for int32_t ?' it just hardcodes general opinion. That doesn't looks pretty correct.
Please consider the next solution:
1. In Basic/TargetInfo add getIntTypeByWidth and getRealTypeByWidth methods. Methods asks target for proper type for given bit width.
2. Fix handleModeAttr according to new methods in TargetInfo.
Fixes:
1st Commit (Done): Add new methods for TargetInfo:
getRealTypeByWidth and getIntTypeByWidth
for ASTContext names are almost same(invokes new methods from TargetInfo):
getIntTypeForBitwidth and getRealTypeForBitwidth
2nd Commit (Current): Fix SemaDeclAttr, handleModeAttr function.
llvm-svn: 190391
This is a part of a series of patches that have been sitting fallow on a
personal branch that I have been messing with for a bit.
The patches start to flesh out the python llvm-c wrapper to the point where you can:
1. Load Modules from Bitcode/Dump/Print them.
2. Iterate over Functions from those modules/get their names/dump them.
3. Iterate over the BasicBlocks from said function/get the BB's name/dump it.
4. Iterate over the Instructions in said BasicBlocks/get the instructions
name/dump the instruction.
My main interest in developing this was to be able to gather statistics about
LLVM IR using python scripts to speed up statistical profiling of different IR
level transformations (hence the focus on printing/dumping/getting names).
This is a gift from me to the LLVM community = ).
I am going to be committing the patches slowly over the next bit as I have time
to prepare the patches.
The overall organization follows the c-api like the bindings that are already
implemented.
llvm-svn: 190388
Enabling sse4.2 will implicitly enable popcnt unless popcnt is explicitly disabled.
Disabling sse4.2 will not disable popcnt if popcnt is explicitly enabled.
llvm-svn: 190387
Use 0 >= 1 instead of 0 != 0 to represent 'false'. This might be slightly more
efficient as isl may create a union of sets for 0 != 0, whereas this is never
needed for the expression 0 >= 1.
Contributed-by: Alexandre Isoard <alexandre.isoard@gmail.com>
llvm-svn: 190384
This patch does a few different things.
This patch improves unused var diags for const vars: we no longer
unconditionally suppress diagnostics for const vars, instead only suppressing
the diagnostic when the declaration appears to be useful.
This patch also makes us more consistently use whether a variable/function
is declared in the main file to suppress diagnostics where appropriate.
Fixes <rdar://problem/14907887>.
llvm-svn: 190382
that /bin/sh re-exec's itself to /bin/bash, so it needs one more resume when you
are using it as the shell than /bin/bash did or you will stop at the start of your
program, rather than running it.
So I added a Platform API to get the number of resumes needed when launching with
a particular shell, and set the right values for Mac OS X.
<rdar://problem/14935282>
llvm-svn: 190381
"coalesce the line ranges for a file & line breakpoint to the first range in each block". We were still setting a silly number
of independent breakpoints sometimes, and until we get a compiler that emits trustworthy is_stmt flags in the line table, we
need to do something to reduce the noise.
<rdar://problem/14920404>
llvm-svn: 190380
Summary:
Calling conventions are inherited during decl merging. Before this
change, deduction would fail due to a type mismatch between the template
and the specialization. This change adjusts the CCs to match before
deduction, and lets the decl merging logic diagnose mismatch or inherit
the CC from the template.
This allows specializations of static member function templates in the
Microsoft C++ ABI.
Reviewers: rsmith
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1570
llvm-svn: 190377
The vselect mask isn't a setcc.
This breaks in the case when the result of getSetCCResultType
is larger than the vector operands
e.g. %tmp = select i1 %cmp <2 x i8> %a, <2 x i8> %b
when getSetCCResultType returns <2 x i32>, the assertion
that the (MaskTy.getSizeInBits() == Op1.getValueType().getSizeInBits())
is hit.
No test since I don't think I can hit this with any of the current
targets. The R600/SI implementation would break, since it returns a
vector of i1 for this, but it doesn't reach ExpandSELECT for other
reasons.
llvm-svn: 190376
Rui Ueyama