summaryrefslogtreecommitdiffstats
path: root/clang
Commit message (Collapse)AuthorAgeFilesLines
...
* Add a helper to get a CXXConstructExpr from a CXXNewExpr.Matt Beaumont-Gay2012-03-091-0/+5
| | | | llvm-svn: 152433
* Improve our semantic error recovery.Argyrios Kyrtzidis2012-03-094-15/+47
| | | | | | | | | | | | | | | | | When an error made a record member invalid, the record would stay as "isBeingDefined" and not "completeDefinition". Even easily recoverable errors ended up propagating records in such "beingDefined" state, for example: struct A { ~A() const; // expected-error {{'const' qualifier is not allowed on a destructor}} }; struct B : A {}; // A & B would stay as "not complete definition" and "being defined". This weird state was impending lookups in the records and hitting assertion in the ASTWriter. Part of rdar://11007039 llvm-svn: 152432
* [AST] Mark Expr::Ignore*() functions as LLVM_READONLY.Daniel Dunbar2012-03-091-15/+17
| | | | llvm-svn: 152431
* [AST] Mark ASTContext::getTypeInfo and some friends as LLVM_READONLY.Daniel Dunbar2012-03-091-4/+6
| | | | llvm-svn: 152430
* [AST] Mark Expr::getExprLoc() as LLVM_READONLY.Daniel Dunbar2012-03-092-10/+14
| | | | llvm-svn: 152429
* [Sema] Remove dead getSourceRange() call, caught by Clang after markingDaniel Dunbar2012-03-091-2/+1
| | | | | | LLVM_READONLY. llvm-svn: 152428
* [AST/etc] Mark {getSourceRange(),getStartLoc(),getEndLoc()} as LLVM_READONLY.Daniel Dunbar2012-03-0921-211/+237
| | | | | | | | | | | | | | | | - The theory here is that we have these functions sprinkled in all over the place. This should allow the optimizer to at least realize it can still do load CSE across these calls. - I blindly marked all instances as such, even though the optimizer can infer this attribute in some instances (some of the inline ones) as that was easier and also, when given the choice between thinking and not thinking, I prefer the latter. You might think this is mere frivolity, but actually this is good for a .7 - 1.1% speedup on 403.gcc/combine.c, JSC/Interpreter.cpp, OGF/NSBezierPath-OAExtensions.m. llvm-svn: 152426
* Add user-defined literals to release notes.Richard Smith2012-03-091-0/+1
| | | | llvm-svn: 152424
* lldb support: under debugger support flag, when sending messageFariborz Jahanian2012-03-093-0/+27
| | | | | | | to forward class, and assigning to an 'id' type var, message sends default to 'id'. // rdar"//10988847 llvm-svn: 152420
* [AST/Sema/libclang] Replace getSourceRange().getBegin() with getLocStart().Daniel Dunbar2012-03-0911-162/+162
| | | | | | | | | - getSourceRange().getBegin() is about as awesome a pattern as .copy().size(). I already killed the hot paths so this doesn't seem to impact performance on my tests-of-the-day, but it is a much more sensible (and shorter) pattern. llvm-svn: 152419
* [AST/Sema] Add ↵Daniel Dunbar2012-03-093-0/+11
| | | | | | {CXXBaseSpecifier,Declarator,DeclSpec,TypeLoc,UnqualifiedId}::getLoc{Start,End}. llvm-svn: 152418
* [AST] Define a few more key getLocStart() implementations.Daniel Dunbar2012-03-093-0/+23
| | | | | | | - This cuts the # of getSourceRange calls by 60% on OGF/NSBezierPath-OAExtensions.m. llvm-svn: 152412
* [AST] Reimplement Stmt::getLoc{Start,End} to dispatch to subclass overloads.Daniel Dunbar2012-03-093-3/+71
| | | | | | | | - getSourceRange() can be very expensive, we should try to avoid it if at all possible. In conjunction with the previous commit I measured a ~2% speedup on 403.gcc/combine.c and a 3% speedup on OmniGroupFrameworks/NSBezierPath-OAExtensions.m. llvm-svn: 152411
* [AST] Add {DeclRefExpr,MemberExpr,ImplicitCastExpr}::{getLocStart,getLocEnd} ↵Daniel Dunbar2012-03-092-14/+36
| | | | | | | | methods. - There are probably a lot more of these worth adding, but these are a start at hitting some of the exprs for which getSourceRange().getBegin() is a poor substitute for getLocStart(). llvm-svn: 152410
* [AST] Add DeclaratioName::getLoc{Start,End} for consistency.Daniel Dunbar2012-03-091-0/+7
| | | | llvm-svn: 152409
* [Basic] Rename LangOptions::NoInline to NoInlineDefine, to make things a bitDaniel Dunbar2012-03-093-3/+3
| | | | | | more obvious. llvm-svn: 152408
* Harden test for 32 bit platforms, where size_t is unsigned int (not unsigned ↵Benjamin Kramer2012-03-091-5/+5
| | | | | | long). llvm-svn: 152404
* -Wformat-non-iso: warn about positional arguments (pr12017)Hans Wennborg2012-03-0910-41/+68
| | | | | | | | | This renames the -Wformat-non-standard flag to -Wformat-non-iso, rewords the current warnings a bit (pointing out that a format string is not supported by ISO C rather than being "non standard"), and adds a warning about positional arguments. llvm-svn: 152403
* Fix statement printing for raw and template user-defined literals.Richard Smith2012-03-092-6/+28
| | | | llvm-svn: 152401
* User-defined literals are done.Richard Smith2012-03-094-4/+13
| | | | llvm-svn: 152396
* Literal operator suffixes and regular names live in separate namespaces.Richard Smith2012-03-093-3/+23
| | | | llvm-svn: 152395
* Literal operators can't have default arguments.Richard Smith2012-03-094-7/+16
| | | | llvm-svn: 152394
* Support for raw and template forms of numeric user-defined literals,Richard Smith2012-03-0921-128/+562
| | | | | | and lots of tidying up. llvm-svn: 152392
* Remove accidental binary commit.John McCall2012-03-091-0/+0
| | | | llvm-svn: 152378
* [AST] Memoize ASTContext::getTypeInfo().Daniel Dunbar2012-03-092-4/+21
| | | | | | | | | | | | | | | | | | - On -emit-llvm-only of 403.gcc/combine.c, for example, we make 160k calls to getTypeInfo but only ever deal with 680 some distinct types. I saw these speedups (user time): 403.gcc/combine.c -- 3.1% OmniGroupFrameworks/NSBezierPath-OAExtensions.m -- 3.6% JavaScriptCore/Interpreter.cpp -- 1.4% which seems pretty sweet. I ran some histograms on those compiles and we end up doing a ton of getTypeInfo() on 'char' and 'int'. I tried splitting out a fast path for builtin types, but this wasn't a win. Still kinda seems like we could be doing better here. llvm-svn: 152377
* Perform l2r conversions on delete operands before doingJohn McCall2012-03-093-3/+24
| | | | | | | type-analysis; otherwise, we just completely do the wrong thing for placeholders. llvm-svn: 152375
* Make sure we update the static local decl address map when we are forced to ↵Eli Friedman2012-03-093-6/+24
| | | | | | rebuild a global because of the initializer. <rdar://problem/10957867>. llvm-svn: 152372
* Make sure constant emission handles initializer lists with strings ↵Eli Friedman2012-03-092-1/+7
| | | | | | correctly. Part of <rdar://problem/10957867>. llvm-svn: 152370
* [AST] NestedNameSpecifier's ctor/dtor are trivial and should be inlined.Daniel Dunbar2012-03-092-10/+6
| | | | llvm-svn: 152365
* [AST] Type::isVoidType() is trivial and should be inlined.Daniel Dunbar2012-03-092-7/+6
| | | | llvm-svn: 152364
* [AST] Reduce Decl::getASTContext() calls.Daniel Dunbar2012-03-098-33/+50
| | | | | | | - This function is not at all free; pass it around along some hot paths instead of recomputing it deep inside various VarDecl methods. llvm-svn: 152363
* [analyzer] Implement basic path diagnostic pruning based on "interesting" ↵Ted Kremenek2012-03-0925-747/+944
| | | | | | | | | | | | | | | | | symbols and regions. Essentially, a bug centers around a story for various symbols and regions. We should only include the path diagnostic events that relate to those symbols and regions. The pruning is done by associating a set of interesting symbols and regions with a BugReporter, which can be modified at BugReport creation or by BugReporterVisitors. This patch reduces the diagnostics emitted in several of our test cases. I've vetted these as having desired behavior. The only regression is a missing null check diagnostic for the return value of realloc() in test/Analysis/malloc-plist.c. This will require some investigation to fix, and I have added a FIXME to the test case. llvm-svn: 152361
* [analyzer] Use call graph to determine order in which functions areAnna Zaks2012-03-081-7/+88
| | | | | | | | | | | | | | analyzed. The CallGraph is used when inlining is on, which is the current default. This alone does not bring any performance improvement. It's a stepping stone for the upcoming optimization in which we do not re-analyze a function that has already been analyzed while inlined in other functions. Using the call graph makes it easier to play with the order of functions to minimize redundant analyzes. llvm-svn: 152352
* [analyzer] Rework inlining related command line options.Anna Zaks2012-03-0819-32/+149
| | | | | | | | | - Remove -analyzer-inline-call. - Add -analyzer-ipa=[none|inlining] - Add -analyzer-inlining-mode to allow experimentation for different performance tuning methods. llvm-svn: 152351
* Call Graph: Only the root node is allowed to have an invalid Decl*.Anna Zaks2012-03-081-1/+4
| | | | llvm-svn: 152350
* Remove a test FIXME for a case which is already fixed.Richard Smith2012-03-081-3/+2
| | | | llvm-svn: 152349
* Fix a couple of issues with literal-operator-id parsing, and provide recoveryRichard Smith2012-03-084-11/+88
| | | | | | | | | | | | for a few kinds of error. Specifically: Since we're after translation phase 6, the "" token might be formed by multiple source-level string literals. Checking the token width is not a correct way of detecting empty string literals, due to escaped newlines. Diagnose and recover from a missing space between "" and suffix, and from string literals other than "", which are followed by a suffix. llvm-svn: 152348
* Don't crash when a statement in a block is ill-formed butJohn McCall2012-03-082-0/+15
| | | | | | introduces cleanups anyway. llvm-svn: 152345
* When checking the encoding of an 8-bit string literal, don't just check theRichard Smith2012-03-085-19/+51
| | | | | | | first codepoint! Also, don't reject empty raw string literals for spurious "encoding" issues. Also, don't rely on undefined behavior in ConvertUTF.c. llvm-svn: 152344
* Turn explicit construction of temporaries using initializer list syntax into ↵Sebastian Redl2012-03-082-4/+26
| | | | | | CXXTemporaryObjectExprs, not just CXXConstructExprs, which have a worrying tendency to vanish. Fixes PR12167. llvm-svn: 152340
* Untangle getUnderlyingDeclImpl, no functionality change.Benjamin Kramer2012-03-081-9/+7
| | | | llvm-svn: 152339
* [utils/FuzzTest] Add '--stop-on-fail' to stop the script on the first failureArgyrios Kyrtzidis2012-03-081-1/+11
| | | | | | without reverting the changes. llvm-svn: 152333
* [AST] Clean up NamedDecl::getUnderlyingDecl() change a bit.Daniel Dunbar2012-03-081-3/+7
| | | | llvm-svn: 152332
* Remove stray semi-colon.Daniel Dunbar2012-03-081-1/+1
| | | | llvm-svn: 152331
* [AST] APValue: Split the fast path of MakeUninit to be inline.Daniel Dunbar2012-03-082-2/+6
| | | | | | | - This change seems to be a tiny loss on 403.gcc/combine.c (.2%), but I think it is the right thing to do. llvm-svn: 152330
* [utils/FuzzTest] The help message says that if you don't specify --max-testsArgyrios Kyrtzidis2012-03-081-1/+1
| | | | | | "the script will run forever continually picking new tests to run", make it so. llvm-svn: 152327
* [utils/FuzzTest] For --filelist the option type is string, not integer.Argyrios Kyrtzidis2012-03-081-1/+1
| | | | llvm-svn: 152324
* [Basic] PartialDiagnostic: Manual split out the cold part of freeStorage(), theDaniel Dunbar2012-03-081-0/+10
| | | | | | | hot path will typically profitably get inlined (based on what I see from the users of PartialDiagnostic). llvm-svn: 152322
* [AST] Change NamedDecl::getUnderlyingDecl() to inline the fast (and ↵Daniel Dunbar2012-03-082-2/+8
| | | | | | incredibly common) path. llvm-svn: 152321
* Replace MarkVarRequired with a more genericRafael Espindola2012-03-088-21/+22
| | | | | | HandleCXXStaticMemberVarInstantiation. Suggested by Argyrios. llvm-svn: 152320
OpenPOWER on IntegriCloud