Richard Smith [Thu, 30 May 2013 22:40:16 +0000 (22:40 +0000)]
Walk over MaterializeTemporaryExpr when reverting an initializer to its
syntactic form in template instantiation. Previously, this blocked the
reversion and we ended up losing inner CXXBindTemporaryExprs (and thus
forgetting to call destructors!).
Objective-C: Implements gcc's -Wselector option
which diagnoses type mismatches of identical
selectors declared in classes throughout.
// rdar://14007194
Jordan Rose [Thu, 30 May 2013 21:30:17 +0000 (21:30 +0000)]
Revert "[analyzer; alternate edges] don't add an edge incoming from the start of a function"
...and make this work correctly in the current codebase.
After living on this for a while, it turns out to look very strange for
inlined functions that have only a single statement, and somewhat strange
for inlined functions in general (since they are still conceptually in the
middle of the path, and there is a function-entry path note).
It's worth noting that this only affects inlined functions; in the new
arrow generation algorithm, the top-level function still starts at the
first real statement in the function body, not the enclosing CompoundStmt.
Rafael Espindola [Thu, 30 May 2013 21:23:15 +0000 (21:23 +0000)]
Fix PR16060.
The testcase in PR16060 points out that while template arguments can
show that a type is not externally visible, the standards still says
they have external linkage.
In terms of our implementation, it means that we should merge just the
isExternallyVisible bit, not the formal linkage.
Fix potential infinite loop when iterating over redeclarations of an ObjMethodDecl, resulting from invalid code.
Check for invalid decls in ObjCMethodDecl::getNextRedeclaration(); otherwise if we start from an invalid redeclaration
of an @implementation we would move to the @interface and not reach the original declaration again.
Jordan Rose [Thu, 30 May 2013 18:14:27 +0000 (18:14 +0000)]
[analyzer] Don't crash if a block's signature just has the return type.
It is okay to declare a block without an argument list: ^ {} or ^void {}.
In these cases, the BlockDecl's signature-as-written will just contain
the return type, rather than the entire function type. It is unclear if
this is intentional, but the analyzer shouldn't crash because of it.
Adrian Prantl [Thu, 30 May 2013 18:12:23 +0000 (18:12 +0000)]
Do not reuse the debug location of the return value's store if there is autorelease code to be emitted between store and return instructions. This is analoguous to what we do for lexical scope cleanups.
Aaron Ballman [Thu, 30 May 2013 16:20:00 +0000 (16:20 +0000)]
Microsoft has a language extension which allows union members to be
references. What's more, they use this language extension in their
ATL header files (which come as part of MFC and the Win32 SDK). This patch implements support for the Microsoft extension, and addresses PR13737.
Aaron Ballman [Thu, 30 May 2013 01:55:39 +0000 (01:55 +0000)]
Add support to fallback on operator new when a placement operator new[] is called for which there is no valid declaration. This fallback only happens in Microsoft compatibility mode. This patch addresses PR13164, and improves support for the WDK.
Jordan Rose [Thu, 30 May 2013 01:05:58 +0000 (01:05 +0000)]
[analyzer; new edges] In for(;;), use the ForStmt itself for loop notes.
Most loop notes (like "entering loop body") are attached to the condition
expression guarding a loop or its equivalent. For loops may not have a
condition expression, though. Rather than crashing, just use the entire
ForStmt as the location. This is probably the best we can do.
Manuel Klimek [Wed, 29 May 2013 22:06:18 +0000 (22:06 +0000)]
Fixes error when splitting block comments.
When trying to fall back to search from the end onwards, we
would still find leading whitespace if the leading whitespace
went on after the end of the line.
Ted Kremenek [Wed, 29 May 2013 21:50:46 +0000 (21:50 +0000)]
Split off casts to void* for -Wint-to-pointer-cast to subgroup -Wint-to-void-pointer-cast.
This change is motivated from user feedback that some APIs use
void* as an opaque "context" object that may not really be a pointer.
Such users want an ability to turn off the warning for casts
to void* while preserving the warning for other cases.
Jordan Rose [Wed, 29 May 2013 20:50:34 +0000 (20:50 +0000)]
[analyzer] Accept references to variables declared "extern void" (C only).
In C, 'void' is treated like any other incomplete type, and though it is
never completed, you can cast the address of a void-typed variable to do
something useful. (In C++ it's illegal to declare a variable with void type.)
Previously we asserted on this code; now we just treat it like any other
incomplete type.
And speaking of incomplete types, we don't know their extent. Actually
check that in TypedValueRegion::getExtent, though that's not being used
by any checkers that are on by default.
Edwin Vane [Wed, 29 May 2013 16:01:10 +0000 (16:01 +0000)]
Tooling: Call back for both begin and end of sources
newFrontendActionFactory() took a pointer to a callback to call when a source
file was done being processed by an action. This revision updates the callback
to include an ante-processing callback as well.
Callback-providing class renamed and callback functions themselves renamed.
Functions are no longer pure-virtual so users aren't forced to implement both
callbacks if one isn't needed.
Manuel Klimek [Wed, 29 May 2013 15:10:11 +0000 (15:10 +0000)]
Use a non-recursive implementation to reconstruct line breaks.
Now that the TokenAnnotator does not require stack space anymore,
reconstructing the lines has become the limiting factor. This patch
fixes that problem, allowing large files with multiple megabytes of
single unwrapped lines to be formatted.
Manuel Klimek [Wed, 29 May 2013 14:47:47 +0000 (14:47 +0000)]
The second step in the token refactoring.
Gets rid of AnnotatedToken, putting everything into FormatToken.
FormatTokens are created once, and only referenced by pointer. This
enables multiple future features, like having tokens shared between
multiple UnwrappedLines (while there's still work to do to fully enable
that).
Rafael Espindola [Wed, 29 May 2013 04:55:30 +0000 (04:55 +0000)]
Don't compute the visibility unless we really have to.
This brings the number of linkage computations in "clang -cc1" in SemaExpr.ii
from 58426 to 43134. With -emit-llvm the number goes from 161045 to 145461.
Anna Zaks [Tue, 28 May 2013 23:24:01 +0000 (23:24 +0000)]
[analyzer] Use the expression’s type instead of region’s type in ArrayToPointer decay evaluation
This gives slightly better precision, specifically, in cases where a non-typed region represents the array
or when the type is a non-array type, which can happen when an array is a result of a reinterpret_cast.
Anna Zaks [Tue, 28 May 2013 22:32:08 +0000 (22:32 +0000)]
[analyzer] Re-enable reasoning about CK_LValueBitCast
It’s important for us to reason about the cast as it is used in std::addressof. The reason we did not
handle the cast previously was a crash on a test case (see commit r157478). The crash was in
processing array to pointer decay when the region type was not an array. Address the issue, by
just returning an unknown in that case.
Rafael Espindola [Tue, 28 May 2013 19:43:11 +0000 (19:43 +0000)]
Check the linkage cache at every recursive step.
Before this patch the linkage cache was only used by the entry level function
(getLinkage). The function that does the actual computation (getLVForDecl),
never looked at it.
This means that we would not reuse an entry in the cache when getLVForDecl did
a recursive call. This patch fixes that by adding another computation enum
value for when we don't care about the linkage at all and having getLVForDecl
check the cache in that case.
When running "clang -cc1" over SemaExpr.ii this brings the number of linkage
computations from 93749 to 58426. When running "clang -cc1 -emit-llvm -O3" it
goes from 198708 to 161444.
For SemaExpr.ii at least linkage computation is a small enough percentage of
the work that the time difference was in the noise.
When asserts are enabled this patch also causes clang to check the linkage
cache even on recursive calls.
Daniel Jasper [Tue, 28 May 2013 18:50:02 +0000 (18:50 +0000)]
Support uniform inits in braced lists.
This made it necessary to remove an error detection which would let us
bail out of braced lists in certain situations of missing "}". However,
as we always entirely escape from the braced list on finding ";", this
should not be a big problem.
With this, we can no format braced lists with uniformat inits:
Anna Zaks [Tue, 28 May 2013 17:31:43 +0000 (17:31 +0000)]
[analyzer] Use a more generic MemRegion.getAsOffset to evaluate bin operators on MemRegions
In addition to enabling more code reuse, this suppresses some false positives by allowing us to
compare an element region to its base. See the ptr-arith.cpp test cases for an example.
Manuel Klimek [Tue, 28 May 2013 11:55:06 +0000 (11:55 +0000)]
A first step towards giving format tokens pointer identity.
With this patch, we create all tokens in one go before parsing and pass
an ArrayRef<FormatToken*> to the UnwrappedLineParser. The
UnwrappedLineParser is switched to use pointer-to-token internally.
The UnwrappedLineParser still copies the tokens into the UnwrappedLines.
This will be fixed in an upcoming patch.
Manuel Klimek [Tue, 28 May 2013 10:01:59 +0000 (10:01 +0000)]
Disable tab expansion when counting the columns in block comments.
To fully support this, we also need to expand tabs in the text before
the block comment. This patch breaks indentation when there was a
non-standard mixture of spaces and tabs used for indentation, but
fixes a regression in the simple case:
{
/*
* Comment.
*/
int i;
}
Is now formatted correctly, if there were tabs used for indentation
before.
Manuel Klimek [Tue, 28 May 2013 08:55:01 +0000 (08:55 +0000)]
Fixes indentation of empty lines in block comments.
Block comment indentation of empty lines regressed, as we did not
have a test for it.
/* Comment with...
*
* empty line. */
is now formatted correctly again.
Daniel Jasper [Tue, 28 May 2013 07:42:44 +0000 (07:42 +0000)]
Fix formatting of expressions containing ">>".
This gets turned into two ">" operators at the beginning in order to
simplify template parameter handling. Thus, we need a special case to
handle those two binary operators correctly.
With this patch, clang-format can now correctly handle cases like:
aaaaaa = aaaaaaa(aaaaaaa, // break
aaaaaa) >>
bbbbbb;
1. Pad structs to a multiple of 64 bits, so they are passed
'left-aligned' in registers.
2. Expose aligned floating point elements as first-level elements, so
the code generator knows to pass them in floating point registers.
We also compute the InReg flag which indicates that the struct contains
aligned 32-bit floats. This flag is used by the code generator to pick
the right registers.
Add a SparcV9ABIInfo class for handling the standard SPARC v9 ABI.
- All integer arguments smaller than 64 bits are extended.
- Large structs are passed indirectly, not using 'byval'.
- Structs up to 32 bytes in size are returned in registers.
Some things are not implemented yet:
- EmitVAArg can be implemented in terms of the va_arg instruction.
- When structs are passed in registers, float members require special
handling because they are passed in the floating point registers.
- Structs are left-aligned when passed in registers. This may require
padding.
Manuel Klimek [Mon, 27 May 2013 15:23:34 +0000 (15:23 +0000)]
Major refactoring of BreakableToken.
Unify handling of whitespace when breaking protruding tokens with other
whitespace replacements.
As a side effect, the BreakableToken structure changed significantly:
- have a common base class for single-line breakable tokens, as they are
much more similar
- revamp handling of multi-line comments; we now calculate the
information about lines in multi-line comments similar to normal
tokens, and always issue replacements
As a result, we were able to get rid of special casing of trailing
whitespace deletion for comments in the whitespace manager and the
BreakableToken and fixed bugs related to tab handling and escaped
newlines.
as they tend to make code really hard to read (how would you even indent the
next line?). Previously we have implemented this in a hacky way, which has now
shown to lead to problems. This fixes a few weird looking formattings, such as:
Sergey Matveev [Mon, 27 May 2013 11:17:01 +0000 (11:17 +0000)]
Add -fsanitize=leak to driver options.
If -fsanitize=leak is specified, link the program with the
LeakSanitizer runtime. Ignore this option when -fsanitize=address is specified,
because AddressSanitizer has this functionality built in.
':'s in dictionary literals (and the corresponding {}s) are now marked as
TT_ObjCDictLiteral too, which makes further improvements to dict literal
layout possible.
Rafael Espindola [Sat, 25 May 2013 17:16:20 +0000 (17:16 +0000)]
Fix linkage computation for derived types in inline functions.
John noticed that the fix for pr15930 (r181981) didn't handle indirect
uses of local types. For example, a pointer to local struct, or a
function that returns it.
One way to implement this would be to recursively look for local
types. This would look a lot like the linkage computation itself for
types.
To avoid code duplication and utilize the existing linkage cache, this
patch just makes the computation of "type with no linkage but
externally visible because it is from an inline function" part of the
linkage computation itself.
[Preprocessor] Prevent expansion of y in x ## y when x is empty
When x is empty, x ## is suppressed, and when y gets expanded, the fact that it follows ## is not
available in the macro expansion result. The macro definition can be checked instead, the ## will
be available there regardless of what x expands to.
Jordan Rose [Fri, 24 May 2013 21:43:11 +0000 (21:43 +0000)]
[analyzer] Treat analyzer-synthesized function bodies like implicit bodies.
When generating path notes, implicit function bodies are shown at the call
site, so that, say, copying a POD type in C++ doesn't jump you to a header
file. This is especially important when the synthesized function itself
calls another function (or block), in which case we should try to jump the
user around as little as possible.
By checking whether a called function has a body in the AST, we can tell
if the analyzer synthesized the body, and if we should therefore collapse
the call down to the call site like a true implicitly-defined function.
Jordan Rose [Fri, 24 May 2013 21:43:05 +0000 (21:43 +0000)]
[analyzer; new edges] Properly set location after exiting an inlined call.
The new edge algorithm would keep track of the previous location in each
location context, so that it could draw arrows coming in and out of each
inlined call. However, it tried to access the location of the call before
it was actually set (at the CallEnter node). This only affected
unterminated calls at the end of a path; calls with visible exit nodes
already had a valid location.
This patch ditches the location context map, since we're processing the
nodes in order anyway, and just unconditionally updates the PrevLoc
variable after popping out of an inlined call.
David Blaikie [Fri, 24 May 2013 21:33:22 +0000 (21:33 +0000)]
DebugInfo: Rename CreatePointerType to getOrCreateTypeDeclaration
To make this more consistent with 'getOrCreateType' & clarify the
distinction between the two. The only thing I couldn't quite communicate
in the name is that getOrCreateTypeDeclaration may actually produce a
full definition (in -fno-limit-debug-info) but the point is to call it
whenever only a declaration is needed & the implementation can choose
whether to provide a declaration or definition.
(also, unfortunately, getOrCreateType can produce declarations too - we
should sure this up by making it not do that - any caller that can
tolerate a declaration should be calling getOrCreateTypeDeclaration
instead)
David Blaikie [Fri, 24 May 2013 21:24:35 +0000 (21:24 +0000)]
PR16091: Error when attempting to emit debug info for undeduced auto return types
Perhaps we should just suppress this, rather than erroring, but since we
have the infrastructure for it I figured I'd use it - if this is
determined to be not the right thing we should probably remove that
infrastructure entirely. I guess it's lying around from the early days
of implementing debug info support.
Diego Novillo [Fri, 24 May 2013 20:18:15 +0000 (20:18 +0000)]
[PATCH] Generate cold attribute for functions marked __atribute__((cold))
This removes a FIXME in CodeGenModule::SetLLVMFunctionAttributesForDefinition.
When a function is declared cold we can now generate the IR attribute in
addition to marking the function to be optimized for size.
I tried adding a separate CHECK in the existing test, but it was
failing. I suppose CHECK matches one line exactly once? This would be
a problem if the attributes are listed in a different order, though they
seem to be sorted.
projects / clang / log