Jordan Rose [Thu, 6 Jun 2013 01:57:24 +0000 (01:57 +0000)]
[analyzer] Look through ExprWithCleanups to see if an expr's consumed.
We based decisions during analysis and during path generation on whether
or not an expression is consumed, so if a top-level expression has
cleanups it's important for us to look through that.
Anna Zaks [Thu, 6 Jun 2013 00:19:36 +0000 (00:19 +0000)]
[analyzer] Fix a crash that occurs when processing an rvalue array.
When processing ArrayToPointerDecay, we expect the array to be a location, not a LazyCompoundVal.
Special case the rvalue arrays by using a location to represent them. This case is handled similarly
elsewhere in the code.
Jordan Rose [Thu, 6 Jun 2013 00:12:41 +0000 (00:12 +0000)]
[analyzer; new edges] Don't crash if the top-level entry edge is missing.
We previously asserted that there was a top-level function entry edge, but
if the function decl's location is invalid (or within a macro) this edge
might not exist. Change the assertion to an actual check, and don't drop
the first path piece if it doesn't match.
Jordan Rose [Thu, 6 Jun 2013 00:12:37 +0000 (00:12 +0000)]
[analyzer; new edges] Ignore self-edges, not all edges with the same location.
The edge optimizer needs to see edges for, say, implicit casts (which have
the same source location as their operand) to uniformly simplify the
entire path. However, we still don't want to produce edges from a statement
to /itself/, which could occur when two nodes in a row have the same
statement location.
This necessitated moving the check for redundant notes to after edge
optimization, since the check relies on notes being adjacent in the path.
If we're splitting the dwarf for fission and don't want to use the
integrated assembler then go ahead and still split the dwarf anyhow.
Add two tests, one to exercise existing behavior of not splitting
when we're just emitting assembly files and the other to test
that we split when we're not in integrated as mode.
David Blaikie [Wed, 5 Jun 2013 18:30:31 +0000 (18:30 +0000)]
PR14763: Debug info for non-trivial record parameters
There seems to have been some erroneous code attempting to describe the
ABI of parameters (non-trivial record parameters are passed by
reference). This would break the type of the function (especially when
it caused a mismatch between the type of a declaration & a definition)
causing PR14763 and PR14645.
Pavel Labath [Wed, 5 Jun 2013 16:23:30 +0000 (16:23 +0000)]
clang-check: Enable specification of additional compiler arguments
Summary:
This adds two command-line parameters: -extra-arg and -extra-arg-before, which
enable the user to pass additional parameters to the compiler command.
Reid Kleckner [Wed, 5 Jun 2013 15:58:29 +0000 (15:58 +0000)]
[ms-cxxabi] Fix vbptr offsets in memptrs when the vbptr is in an nvbase
Also addresses a review comment from John from on r180985 by removing
the "== -1" check, since it's now reusing the correct code which has the
comment.
Summary:
Detect if the file is valid UTF-8, and if this is the case, count code
points instead of just using number of bytes in all (hopefully) places, where
number of columns is needed. In particular, use the new
FormatToken.CodePointCount instead of TokenLength where appropriate.
Changed BreakableToken implementations to respect utf-8 character boundaries
when in utf-8 mode.
Recommit r183298+r183300 'Get rid of VTableContext::ComputeMethodVTableIndices() and VTableContext::getNumVirtualFunctionPointers()'
In r183298, I've used llvm::SmallPtrSet<..., 8> instead of llvm::SmallVector<..., 8> for NewVirtualFunctionsTy by mistake.
This only manifested when a class had more than 8 virtual functions, which wasn't covered by unit-tests
Make clang tools ignore -fcolor-diagnostics and -fdiagnostics-color retrieved from the compilation database.
Summary:
Clang tools' diagnostic output could be force colored when a command
line from the compilation database contains -fcolor-diagnostics or
-fdiagnostics-color. This is not what we want e.g. for vim integration.
Get rid of VTableContext::ComputeMethodVTableIndices() and VTableContext::getNumVirtualFunctionPointers(); also add some tests for the VTable indices output to make sure we don't regress
David Blaikie [Wed, 5 Jun 2013 05:32:23 +0000 (05:32 +0000)]
PR16214: Debug Info: -flimit-debug-info doesn't omit definitions for types used via typedefs
In an effort to make -flimit-debug-info more consistent I over-shot the
mark & made types used via typedefs never produce definitions in the
debug info (even if the type was used in a way that would require a
definition).
The fix for this is to do exactly what I was hoping to do at some point
- plumb the declaration/definition choice through the various layers of
"CreateType" in CGDebugInfo. In this way we can produce declarations
whenever they are sufficient & definitions otherwise - including when
qualifiers are used, for example (discovered in PR14467). This may not
be complete (there may be other types/situations where we need to
propagate the "declaration/definition" choice) but it lays the basic
foundation which we can enhance in future iterations.
This could actually be implemented with the LLVM IR va_arg instruction,
but it doesn't seem to offer any advantages over accessing the va_list
pointer directly.
Using the va_list pointer directly makes it possible to perform type
coercion directly from the argument array, and the va_list updates are
exposed to the optimizers.
Fix CoerceIntOrPtrToIntOrPtr on big-endian targets.
Type coercion for argument passing is equivalent to storing the source
type and loading the destination type from the same pointer. On
big-endian targets, this means that the high bits of integers are
preserved.
This patch fixes the CoerceIntOrPtrToIntOrPtr() function on big-endian
targets by inserting the required shift instructions to preserve the
high bits instead of the low bits.
This is used by SparcABIInfo when passing small structs in the high bits
of registers.
The 'inreg' attribute can also be applied to function return values in
LLVM IR. The SPARC v9 backend is using the flag when returning structs
containing 32-bit floats.
Richard Smith [Wed, 5 Jun 2013 00:46:14 +0000 (00:46 +0000)]
Model temporary lifetime-extension explicitly in the AST. Use this model to
handle temporaries which have been lifetime-extended to static storage duration
within constant expressions. This correctly handles nested lifetime extension
(through reference members of aggregates in aggregate initializers) but
non-constant-expression emission hasn't yet been updated to do the same.
Summary:
There's Lexer::getBufferStart(), and we need getBufferEnd() to access
the whole input buffer in clang::format::reformat. We don't want to
rely on the fact that the Lexer::BufferEnd always points to '\0', as there can
be embedded '\0's as well.
Fix linkage computation for local types in template functions.
Template functions (and member functions of class templates) present the same
problem as inline functions. They need to be uniqued, so we need to assign
VisibleNoLinkage linkage to types defined in them.
The text of this diagnostic was unnecessarily specific to the current ARM
implementation of validateConstraintModifier, and it gave a potentially
confusing suggestion for fixing the problem. The ARM-specific issue is not
a big deal since that is the only target that currently does any checking of
asm operand modifiers, but until my change in 183172 it was still wrong for
output operands in the way that it referred to the value being truncated when
put into a register, since output operands are retrieved from the registers
instead of being put into them. The bigger problem is that its suggestion to
"use a modifier" is wrong and confusing in the case where a "q" modifier is
incorrectly used with an "r" constraint. In that case, the solution might
well be to remove the modifier or perhaps change the constraint. It's better
to just leave the diagnostic message more generic.
David Blaikie [Tue, 4 Jun 2013 00:28:46 +0000 (00:28 +0000)]
Bound member function diagnostic - suggest no-args calls and note overload candidates
Still missing cases for templates, but this is a step in the right
direction. Also omits suggestions that would be ambiguous (eg: void
func(int = 0); + void func(float = 0); func;)
Bob Wilson [Mon, 3 Jun 2013 23:57:13 +0000 (23:57 +0000)]
Do not report -Wasm-operand-widths for ARM output operands. <rdar://14050337>
We're getting reports of this warning getting triggered in cases where it
is not adding any value. There is no asm operand modifier that you can use
to silence it, and there's really nothing wrong with having an LDRB, for
example, with a "char" output.
Jordan Rose [Mon, 3 Jun 2013 23:00:19 +0000 (23:00 +0000)]
[analyzer] Enable the new edge algorithm by default.
...but don't yet migrate over the existing plist tests. Some of these
would be trivial to migrate; others could use a bit of inspection first.
In any case, though, the new edge algorithm seems to have proven itself,
and we'd like more coverage (and more usage) of it going forwards.
Jordan Rose [Mon, 3 Jun 2013 23:00:09 +0000 (23:00 +0000)]
[analyzer; new edges] Omit subexpression back-edges that span multiple lines.
A.1 -> A -> B
becomes
A.1 -> B
This only applies if there's an edge from a subexpression to its parent
expression, and that is immediately followed by another edge from the
parent expression to a subsequent expression. Normally this is useful for
bringing the edges back to the left side of the code, but when the
subexpression is on a different line the backedge ends up looking strange,
and may even obscure code. In these cases, it's better to just continue
to the next top-level statement.
Jordan Rose [Mon, 3 Jun 2013 23:00:05 +0000 (23:00 +0000)]
[analyzer; new edges] Don't eliminate subexpr edge cycles if the line is long.
Specifically, if the line is over 80 characters, or if the top-level
statement spans mulitple lines, we should preserve sub-expression edges
even if they form a simple cycle as described in the last commit, because
it's harder to infer what's going on than it is for shorter lines.
Jordan Rose [Mon, 3 Jun 2013 23:00:00 +0000 (23:00 +0000)]
[analyzer; new edges] Eliminate "cycle edges" for a single subexpression.
Generating context arrows can result in quite a few arrows surrounding a
relatively simple expression, often containing only a single path note.
|
1 +--2---+
v/ v
auto m = new m // 3 (the path note)
|\ |
5 +--4---+
v
Note also that 5 and 1 are two ends of the "same" arrow, i.e. they go from
event to event. 3 is not an arrow but the path note itself.
Now, if we see a pair of edges like 2 and 4---where 4 is the reverse of 2
and there is optionally a single path note between them---we will
eliminate /both/ edges. Anything more complicated will be left as is
(more edges involved, an inlined call, etc).
The next commit will refine this to preserve the arrows in a larger
expression, so that we don't lose all context.
Jordan Rose [Mon, 3 Jun 2013 22:59:53 +0000 (22:59 +0000)]
[analyzer; new edges] Improve enclosing contexts for logical expressions.
The old edge builder didn't have a notion of nested statement contexts,
so there was no special treatment of a logical operator inside an if
(or inside another logical operator). The new edge builder always tries
to establish the full context up to the top-level statement, so it's
important to know how much context has been established already rather
than just checking the innermost context.
This restores some of the old behavior for the old edge generation:
the context of a logical operator's non-controlling expression is the
subexpression in the old edge algorithm, but the entire operator
expression in the new algorithm.
Jordan Rose [Mon, 3 Jun 2013 22:59:48 +0000 (22:59 +0000)]
[analyzer; new edges] Include context for edges to sub-expressions.
The current edge-generation algorithm sometimes creates edges from a
top-level statement A to a sub-expression B.1 that's not at the start of B.
This creates a "swoosh" effect where the arrow is drawn on top of the
text at the start of B. In these cases, the results are clearer if we see
an edge from A to B, then another one from B to B.1.
Admittedly, this does create a /lot/ of arrows, some of which merely hop
into a subexpression and then out again for a single note. The next commit
will eliminate these if the subexpression is simple enough.
This updates and reuses some of the infrastructure from the old edge-
generation algorithm to find the "enclosing statement" context for a
given expression. One change in particular marks the context of the
LHS or RHS of a logical binary operator (&&, ||) as the entire operator
expression, rather than the subexpression itself. This matches our behavior
for ?:, and allows us to handle nested context information.
Jordan Rose [Mon, 3 Jun 2013 22:59:41 +0000 (22:59 +0000)]
CFG: In a DeclStmt, skip anything that's not a VarDecl.
Neither the compiler nor the analyzer are doing anything with non-VarDecl
decls in the CFG, and having them there creates extra nodes in the
analyzer's path diagnostics. Simplify the CFG (and the path edges) by
simply leaving them out. We can always add interesting decls back in when
they become relevant.
Note that this only affects decls declared in a DeclStmt, and then only
those that appear within a function body.
David Blaikie [Mon, 3 Jun 2013 19:46:44 +0000 (19:46 +0000)]
UnresolvedMemberExpr should have an expr location of the member (not the base)
This matches the behavior of MemberExpr and makes diagnostics such as
"reference to non-static member function must be called" more legible in
the case that the base & member are split over multiple lines (prior to
this change the diagnostic would point to the base, not the member -
making it very unclear in chained multi-line builder-style calls)
Richard Smith [Mon, 3 Jun 2013 07:13:35 +0000 (07:13 +0000)]
Do not walk through member-accesses on bitfields when looking for the object
which is lifetime-extended by a reference binding. An additional temporary is
created for such a bitfield access (although we have no explicit AST
representation for it).
Richard Smith [Mon, 3 Jun 2013 05:03:02 +0000 (05:03 +0000)]
Refactor constant expression evaluation to associate the complete object of a
materialized temporary with the corresponding MaterializeTemporaryExpr. This is
groundwork for providing C++11's guaranteed static initialization for global
references bound to lifetime-extended temporaries (if the initialization is a
constant expression).
In passing, fix a couple of bugs where some evaluation failures didn't trigger
diagnostics, and a rejects-valid where potential constant expression testing
would assume that it knew the dynamic type of *this and would reject programs
which relied on it being some derived type.
David Majnemer [Sun, 2 Jun 2013 08:40:42 +0000 (08:40 +0000)]
Allow paren casted throw statements inside of ternary expressions
clang would incorrectly not allow the following:
int x = true ? (throw 1) : 2;
The problem exists because we don't see beyond the parens.
This, in turn, causes us to believe that we are choosing between void
and int which we diagnose as an error.
Instead, allow clang to see the 'throw' inside the parens.
Richard Smith [Sun, 2 Jun 2013 00:09:52 +0000 (00:09 +0000)]
PR12848: When emitting a local variable declared 'constexpr', always initialize it with a store or a memcpy, not by emitting the initializer expression. This is not required for correctness, but more closely aligns with people's expectations, and is cheap (since we've already evaluated the initializer).
Anna Zaks [Fri, 31 May 2013 23:47:32 +0000 (23:47 +0000)]
[analyzer] Malloc checker should only escape the receiver when “[O init..]” is called.
Jordan has pointed out that it is valuable to warn in cases when the arguments to init escape.
For example, NSData initWithBytes id not going to free the memory.
Ed Schouten [Fri, 31 May 2013 19:27:59 +0000 (19:27 +0000)]
Add support for optimized (non-generic) atomic libcalls.
For integer types of sizes 1, 2, 4 and 8, libcompiler-rt (and libgcc)
provide atomic functions that pass parameters by value and return
results directly.
libgcc and libcompiler-rt only provide optimized libcalls for
__atomic_fetch_*, as generic libcalls on non-integer types would make
little sense. This means that we can finally make __atomic_fetch_* work
on architectures for which we don't provide these operations as builtins
(e.g. ARM).
This should fix the dreaded "cannot compile this atomic library call
yet" error that would pop up once every while.
Daniel Jasper [Fri, 31 May 2013 14:56:20 +0000 (14:56 +0000)]
Make formatting of empty blocks more consistent.
With this patch, the simplified rule is:
If the block is part of a declaration (class, namespace, function,
enum, ..), merge an empty block onto a single line. Otherwise
(specifically for the compound statements of if, for, while, ...),
keep the braces on two separate lines.
The reasons are:
- Mostly the formatting of empty blocks does not matter much.
- Empty compound statements are really rare and are usually just
inserted while still working on the code. If they are on two lines,
inserting code is easier. Also, overlooking the "{}" of an
"if (...) {}" can be really bad.
- Empty declarations are not uncommon, e.g. empty constructors. Putting
them on one line saves vertical space at no loss of readability.
Richard Smith [Fri, 31 May 2013 02:56:17 +0000 (02:56 +0000)]
Fix handling of braced-init-list as reference initializer within aggregate
initialization. Previously we would incorrectly require an extra set of braces
around such initializers.
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!).
projects / clang / log