Aaron Ballman [Thu, 13 Mar 2014 17:00:06 +0000 (17:00 +0000)]
[C++11] Replacing CXXRecordDecl iterators friend_begin() and friend_end() with iterator_range friends(). Updating all of the usages of the iterators with range-based for loops.
Aaron Ballman [Thu, 13 Mar 2014 16:51:27 +0000 (16:51 +0000)]
[C++11] Replacing CXXRecordDecl iterators ctor_begin() and ctor_end() with iterator_range ctors(). Updating all of the usages of the iterators with range-based for loops.
Ben Langmuir [Thu, 13 Mar 2014 16:46:36 +0000 (16:46 +0000)]
Prevent outputting HeaderFileInfos for files not used as headers
When building an AST file, we don't want to output HeaderFileInfo
structures for files that are not actually used as headers in the
current context. This can lead to assuming that unrelated files have
include counts of 0, defeating multiple-include prevention.
This is accomplished by adding an IsValid bit to the HFI.
Aaron Ballman [Thu, 13 Mar 2014 16:36:16 +0000 (16:36 +0000)]
[C++11] Replacing CXXRecordDecl iterators method_begin() and method_end() with iterator_range methods(). Updating all of the usages of the iterators with range-based for loops.
Aaron Ballman [Thu, 13 Mar 2014 16:15:17 +0000 (16:15 +0000)]
[C++11] Replacing CXXRecordDecl iterators vbases_begin() and vbases_end() with iterator_range vbases(). Updating all of the usages of the iterators with range-based for loops.
Aaron Ballman [Thu, 13 Mar 2014 15:41:46 +0000 (15:41 +0000)]
[C++11] Replacing CXXRecordDecl iterators bases_begin() and bases_end() with iterator_range bases(). Updating all of the usages of the iterators with range-based for loops.
Remove an unnecessary check for FormatTok not being null.
Summary:
This check hints clang analyzer, that FormatTok may be null, and it
reports a null pointer dereference error:
http://buildd-clang.debian.net/scan-build/report-827c64.html#Path28
Richard Smith [Thu, 13 Mar 2014 00:28:45 +0000 (00:28 +0000)]
PR18275: If a member function of a class template is declared with a
const-qualified parameter type and the defined with a non-const-qualified
parameter type, the parameter is not const inside its body. Ensure that
the type we use when instantiating the body is the right one. Patch by
suyog sarda!
This is still rather unsatisfactory; it seems like it might be better to
instantiate at least the function parameters, and maybe the complete function
declaration, when we instantiate the definition for such a member function
(instead of reusing the declaration from inside the instantiated class
definition).
Richard Smith [Wed, 12 Mar 2014 23:36:42 +0000 (23:36 +0000)]
Only allow streaming exactly type 'bool' to a DiagnosticBuilder, not anything
that implicitly converts to 'bool' (such as pointers, and the first operand of
?:). Clean up issues found by this. Patch by Stephan Tolksdorf!
Richard Smith [Wed, 12 Mar 2014 23:14:33 +0000 (23:14 +0000)]
Fix crash if delayed template parsing meets an erroneous trailing return type.
Based on a patch and test by Stephan Tolksdorf! Refactoring and fixing adjacent
brokenness by me.
Hans Wennborg [Wed, 12 Mar 2014 21:09:03 +0000 (21:09 +0000)]
MS intrinsics: don't declare __readeflags and __writeeflags in Intrin.h
They're already defined in ia32intrin.h, and this would cause including Intrin.h
in 64-bit mode to fail because of conflicting types. Update ms-intrin.cpp to
also run in 64-bit mode to catch things like this.
Justin Bogner [Wed, 12 Mar 2014 20:53:16 +0000 (20:53 +0000)]
CodeGen: Use a binary format for instrumentation based profiling
This updates CodeGenPGO to use the ProfileDataReader introduced to
llvm in r203703 and the new API for writing out the profile introduced
to compiler-rt in r203710.
Oliver Stannard [Wed, 12 Mar 2014 14:02:50 +0000 (14:02 +0000)]
ARM: Correct alignment of structs passed as byval pointer
When a struct has bitfields overlapping with other members
(as required by the AAPCS), clang uses a packed struct to
represent this. If such a struct is large enough for clang to
pass it as a byval pointer (>64 bytes), we need to set the
alignment of the argument to match the original type.
Unfortunately, XFAIL applies to the entire file. Split the test into two
different units, enabling the proper application of XFAIL on the test which is
expected to fail on Windows.
Ben Langmuir [Wed, 12 Mar 2014 00:06:17 +0000 (00:06 +0000)]
Add an option -fmodules-validate-system-headers
When enabled, always validate the system headers when loading a module.
The end result of this is that when these headers change, we will notice
and rebuild the module.
lex: improve include handling on Linux for Windows
Normalise the path separator character on non-windows platforms. Although this
would work on Windows as well (most newer versions of Windows support either '/'
or '\' as a path separator character), it could potentially cause problems with
full UNC paths. This change enables the use of the Windows SDK on Linux which
will not accept '\' as a path separator.
Jordan Rose [Tue, 11 Mar 2014 16:52:29 +0000 (16:52 +0000)]
[analyzer] Check all conditions in a chained if against each other.
Like the binary operator check of r201702, this actually checks the
condition of every if in a chain against every other condition, an
O(N^2) operation. In most cases N should be small enough to make this
practical, and checking all cases like this makes it much more likely
to catch a copy-paste error within the same series of branches.
Part of IdenticalExprChecker; patch by Daniel Fahlgren!
Aaron Ballman [Tue, 11 Mar 2014 13:19:42 +0000 (13:19 +0000)]
Changing this test case to use an unknown attribute, since there are not currently any type attributes which would apply to a lambda, except in MSVC compatibility mode.
Justin Bogner [Tue, 11 Mar 2014 06:49:34 +0000 (06:49 +0000)]
test: Regenerate profile data for PGO tests
Some of this data had gotten out of date, so we weren't quite testing
what we thought we were. This also moves the outdated data test to its
own file to simplify regenerating the test data.
[HeaderSearch] Fix issue where if a headermap entry maps the filename to a framework import (non-absolute path)
then we fail to find it if it is re-included later on.
Justin Bogner [Tue, 11 Mar 2014 04:37:49 +0000 (04:37 +0000)]
test: Give instrumentation based profiling tests their own directory
These tests are logically related, but they're spread about several
different CodeGen directories. Consolidate them in one place to make
them easier to manage.
Richard Trieu [Tue, 11 Mar 2014 03:11:08 +0000 (03:11 +0000)]
Move the warning about unused relational comparison from -Wunused-value to
-Wunused-comparison. Also, newly warn on unused result from overloaded
relational comparisons, now also in -Wunused-comparison.
Richard Smith [Tue, 11 Mar 2014 03:10:46 +0000 (03:10 +0000)]
If a visibility update record is found for a DeclContext after that Decl has
already been loaded, apply that update record to the Decl immediately, rather
than adding it to a pending list and never applying it.
Richard Smith [Tue, 11 Mar 2014 02:02:47 +0000 (02:02 +0000)]
If a module map is found in a relative -I path, convert the filenames within it
to absolute paths when building the includes file for the module. Without this,
the module build would fail, because the relative paths we were using are not
necessarily relative to a directory in our include path.
Ben Langmuir [Mon, 10 Mar 2014 23:17:14 +0000 (23:17 +0000)]
Add -fmodules-cache-path to test/Headers/c11.c
This started failing for me the last time someone modified the AST file
format. It would be nice if we could just have lit take care of the
module cache used during testing for us, but this helps in the meantime.
Ben Langmuir [Mon, 10 Mar 2014 22:34:23 +0000 (22:34 +0000)]
Turn on the uncovered umbrella header warning by default
This warning (under -Wincomplete-umbrella) diagnoses cases that are
difficult to understand without the compiler's help, since the symptom
is likely to be that you are missing the contents of headers that are
mistakenly omitted from a module. This seems like a good thing to have
on by default (with the usual caveat for -Wsystem-headers).
John McCall [Mon, 10 Mar 2014 22:27:33 +0000 (22:27 +0000)]
RTTI symbols for visible local types may need weak linkage.
Previously, we would always emit them with internal linkage,
but with hidden visibility when the function was hidden, which
is an illegal combination, which could lead LLVM to actually
emit them as strong hidden symbols with hilarious results.
Aaron Ballman [Mon, 10 Mar 2014 17:08:28 +0000 (17:08 +0000)]
[C++11] Replacing DeclBase iterators specific_attr_begin() and specific_attr_end() with iterator_range specific_attrs(). Updating all of the usages of the iterators with range-based for loops.
Samuel Benzaquen [Mon, 10 Mar 2014 15:40:23 +0000 (15:40 +0000)]
Add loc() to the dynamic registry.
Summary:
Add loc() to the dynamic registry.
Other fixes:
- Fix the polymorphic variant value to accept an exact match, even if
there are other possible conversions.
- Fix specifiesTypeLoc() to not crash on an empty
NestedNameSpecifierLoc.