pointer that is passed down through the APIs, and make
FileSystemStatCache::get be the one that filters out
directory lookups that hit files. This also paves the
way to have stat queries be able to return opened files.
llvm-svn: 120060
two copies, since they are fundamentally different
operations and the StringRef one should go away
(it shouldn't be part of FileManager at least).
Remove some dead arguments.
llvm-svn: 120013
FileSystemOpts through a ton of apis, simplifying a lot of code.
This also fixes a latent bug in ASTUnit where it would invoke
methods on FileManager without creating one in some code paths
in cindextext.
llvm-svn: 120010
This patch completely defeated the "passing in a prestat'd size
to MemoryBuffer" optimization, leading to an extra fstat call for
every buffer opened, in order to find out if the datestamp and size
of the file on disk matches what is in the stat cache.
I fully admit that I don't completely understand what is going on here:
why punish code when a stat cache isn't in use? what is the point of a
stat cache if you have to turn around and stat stuff to validate it?
To resolve both these issues, just drop the modtime check and check the
file size, which is the important thing anyway. This should also resolve
PR6812, because presumably windows is stable when it comes to file sizes.
If the modtime is actually important, we should get it and keep it on the
first stat.
This eliminates 833 fstat syscalls when processing Cocoa.h, speeding up
system time on -Eonly Cocoa.h from 0.041 to 0.038s.
llvm-svn: 120001
-Move the stuff of Diagnostic related to creating/querying diagnostic IDs into a new DiagnosticIDs class.
-DiagnosticIDs can be shared among multiple Diagnostics for multiple translation units.
-The rest of the state in Diagnostic object is considered related and tied to one translation unit.
-Have Diagnostic point to the SourceManager that is related with. Diagnostic can now accept just a
SourceLocation instead of a FullSourceLoc.
-Reflect the changes to various interfaces.
llvm-svn: 119730
When -working-directory is passed in command line, file paths are resolved relative to the specified directory.
This helps both when using libclang (where we can't require the user to actually change the working directory)
and to help reproduce test cases when the reproduction work comes along.
--FileSystemOptions is introduced which controls how file system operations are performed (currently it just contains
the working directory value if set).
--FileSystemOptions are passed around to various interfaces that perform file operations.
--Opening & reading the content of files should be done only through FileManager. This is useful in general since
file operations will be abstracted in the future for the reproduction mechanism.
FileSystemOptions is independent of FileManager so that we can have multiple translation units sharing the same
FileManager but with different FileSystemOptions.
Addresses rdar://8583824.
llvm-svn: 118203
in asm's. PR 8501, 8602988.
I don't like including Type.h where it is; the idea was
to get references to X86_MMXTy out of the common code.
Maybe there's a better way?
llvm-svn: 117736
This adds an option to set the _MSC_VER macro without
recompiling. This is very useful when testing compatibility
with the Windows SDK and c++stdlib headers.
-fmsc-version=<version> (defaults to VS2003 (1300))
llvm-svn: 116999
identifiers to determine good typo-correction candidates. Once we've
identified those candidates, we perform name lookup on each of them
and the consider the results.
This optimization makes typo correction > 2x faster on a benchmark
example using a single typo (NSstring) in a tiny file that includes
Cocoa.h from a precompiled header, since we are deserializing far less
information now during typo correction.
There is a semantic change here, which is interesting. The presence of
a similarly-named entity that is not visible can now affect typo
correction. This is both good (you won't get weird corrections if the
thing you wanted isn't in scope) and bad (you won't get good
corrections if there is a similarly-named-but-completely-unrelated
thing). Time will tell whether it was a good choice or not.
llvm-svn: 116528
emitting diagnostics in a binary form to be consumed by libclang,
since libclang no longer does any of its work out-of-process, making
this code dead. Besides, this stuff never worked at 100% anyway.
llvm-svn: 116250
- I don't like returning StringRef's ever, unless it is actually important for
performance, which it isn't here.
- Also, stop validating getClangRevision to be an integer, I don't see a good
reason to do this.
llvm-svn: 115071
#pragma once wasn't working on win32 if the header file was included
using a different case.
I tracked down the problem to the fact that clang::FileManager was
caching files using case sensitive string (UniqueFiles) on Windows.
I changed FileManager to cache filename in lower case only.
Doesn't affect UNIX because UNIX uses Inode to uniquely identify files.
unix doesn't use this codepath.
Analysis and patch by Francois Pichet!
llvm-svn: 111866
-There are 2 instances that change the TokenID for GNU libstdc++ 4.2 compatibility.
To handler those cases introduce a RevertedTokenID bitfield, RevertTokenIDToIdentifier() and hasRevertedTokenIDToIdentifier() methods.
Store the bitfield in PCH.
llvm-svn: 110868
reparsing an ASTUnit. When saving a preamble, create a buffer larger
than the actual file we're working with but fill everything from the
end of the preamble to the end of the file with spaces (so the lexer
will quickly skip them). When we load the file, create a buffer of the
same size, filling it with the file and then spaces. Then, instruct
the lexer to start lexing after the preamble, therefore continuing the
parse from the spot where the preamble left off.
It's now possible to perform a simple preamble build + parse (+
reparse) with ASTUnit. However, one has to disable a bunch of checking
in the PCH reader to do so. That part isn't committed; it will likely
be handled with some other kind of flag (e.g., -fno-validate-pch).
As part of this, fix some issues with null termination of the memory
buffers created for the preamble; we were trying to explicitly
NULL-terminate them, even though they were also getting implicitly
NULL terminated, leading to excess warnings about NULL characters in
source files.
llvm-svn: 109445
reparses an already-parsed translation unit. At the moment it's just a
convenience function, but we hope to use it for performance
optimizations.
llvm-svn: 108756
represent builtins that have the "scanf" attribution (via the format attribute) just
like we do with printf functions. Follow-up work is needed to add similar support
for fscanf et al.
This is to support format-string checking for scanf functions.
llvm-svn: 108499
whether to use objc_msgSend_fpret; the choice is target dependent, not Obj-C ABI
dependent.
- <rdar://problem/8139758> arm objc _objc_msgSend_fpret bug
llvm-svn: 108379
to use them instead of SourceRange. CharSourceRange is just a SourceRange
plus a bool that indicates whether the range has the end character resolved
or whether the end location is the start of the end token. While most of
the compiler wants to think of ranges that have ends that are the start of
the end token, the printf diagnostic stuff wants to highlight ranges within
tokens.
This is transparent to the diagnostic stuff. To start taking advantage of
the new capabilities, you can do something like this:
Diag(..) << CharSourceRange::getCharRange(Begin,End)
llvm-svn: 106338
the x86-64 __va_list_tag with this attribute. The attribute causes the
affected type to behave like a fundamental type when considered by ADL.
(x86-64 is the only target we currently provide with a struct-based
__builtin_va_list)
Fixes PR6762.
llvm-svn: 104941
than 127 groups so this was already failing given -fsigned-char. A subsequent
to commit to TableGen will generate shorts for the arrays themselves.
llvm-svn: 103703
to be algorithmically faster and avoid an std::map. This routine
basically boils down to finding the nearest common ancestor in a
tree, and we (implicitly) have information about nesting depth,
use it!
This wraps up rdar://7948633 - SourceManager::isBeforeInTranslationUnit has poor performance
llvm-svn: 103239
method to be correct. Right now it correctly computes the cache, then
goes ahead and computes the result the hard way, then asserts that they
match. Next I'll actually turn it on.
llvm-svn: 103231
method will sometimes return different results for the same input SourceLocations. I haven't
unraveled this method completely yet, so this truly is a workaround until a better fix comes
along.
llvm-svn: 103143
print the diagnostic category number in the [] at the end
of the line. For example:
$ cat t.c
#include <stdio.h>
void foo() {
printf("%s", 4);
}
$ clang t.c -fsyntax-only -fdiagnostics-print-source-range-info
t.c:3:11:{3:10-3:12}{3:15-3:16}: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat,1]
printf("%s", 4);
~^ ~
1 warning generated.
Clients that want category information can now pick the number
out of the output, rdar://7928231.
More coming.
llvm-svn: 103053
and diagnostic groups. This allows the compiler to group
diagnostics together (e.g. "Logic Warning",
"Format String Warning", etc) like the static analyzer does.
This is not exposed through anything in the compiler yet.
llvm-svn: 103051
- This fixes the last known ABI issues with ARM/APCS.
- I've run the first 1k ABITests with '--no-unsigned --no-vector --no-complex'
on {armv6, armv7} x {-mno-thumb, -mthumb}, and the first 10k tests for armv7
-mthumb, for both function return types and single argument calls. These all
pass now (they failed horribly before without --no-bitfield).
llvm-svn: 102070
about it instead of producing tons of garbage from the lexer.
It would be even better for sourcemgr to dynamically transcode (e.g.
from UTF16 -> UTF8).
llvm-svn: 101924
we will print with each error that occurs during template
instantiation. When the backtrace is longer than that, we will print
N/2 of the innermost backtrace entries and N/2 of the outermost
backtrace entries, then skip the middle entries with a note such as:
note: suppressed 2 template instantiation contexts; use
-ftemplate-backtrace-limit=N to change the number of template
instantiation entries shown
This should eliminate some excessively long backtraces that aren't
providing any value.
llvm-svn: 101882
implemented precisely the same as GCC, but the distinction GCC makes isn't
useful to represent. This allows parsing code which uses GCC-specific keywords
('asm', etc.) without parsing in a fully GNU mode.
llvm-svn: 101667
- Used to determine whether the alignment of the type in a bit-field is
respected when laying out structures. The default is true, targets can
override this as needed.
- This is designed to correspond to the PCC_BITFIELD_TYPE_MATTERS macro in
gcc. The AST/Sema implementation only affects one line, unless I have
forgotten something. I'd appreciate further review.
- IRgen still needs to be updated to fully support this (which is effectively
PR5591).
llvm-svn: 101356
separate count of "suppressed" errors. This way, semantic analysis
bits that depend on the error count to determine whether problems
occured (e.g., some template argument deduction failures, jump-scope
checking) will not get confused.
The actual problem here is that a missing #include (which is a fatal
error) could cause the jump-scope checker to run on invalid code,
which it is not prepared to do. Trivial fix for both
<rdar://problem/7775941> and <rdar://problem/7775709>.
llvm-svn: 101297
actually turned it on. If a diag is produced by a warning which
is an extension but defaults to on, and has no warning group, don't
print any option info.
llvm-svn: 101071
precompiled headers and/or when reading the contents of the file into
memory. These checks seem to be causing spurious regression-test
failures on Windows.
llvm-svn: 100866
of errors and warnings. This allows us to emit something like this:
2 warnings and 1 error generated.
instead of:
3 diagnostics generated.
This also stops counting 'notes' because they are just follow-on information
about the previous diag, not a diagnostic in themselves.
llvm-svn: 100675
cache of PartialDiagnostic::Storage objects into an allocator within
the ASTContext. This eliminates a significant amount of malloc
traffic, for a 10% performance improvement in -fsyntax-only wall-clock
time with 403.gcc's combine.c.
Also, eliminate the RequireNonAbstractType hack I put in earlier,
which was but a symptom of this larger problem.
Fixes <rdar://problem/7806091>.
llvm-svn: 99849
how to handle a diagnostic during template argument deduction, which
may be "substitution failure", "suppress", or "report". This keeps us
from, e.g., emitting warnings while performing template argument
deduction.
llvm-svn: 99560
Diagnostic subsystem, which is used in the rare case where we find a
serious problem (i.e., an inconsistency in the file system) while
we're busy formatting another diagnostic. In this case, the delayed
diagnostic will be emitted after we're done with the other
diagnostic. This is only to be used for fatal conditions detected at
very inconvenient times, where we can neither stop the current
diagnostic in flight nor can we suppress the second error.
llvm-svn: 99175
deserialization of precompiled headers, where the deserialization of
the source location entry for a buffer (e.g., macro instantiation
scratch space) would overwrite a one-element FileID cache in the
source manager. When tickled at the wrong time, we would return the
wrong decomposed source location and eventually cause c-index-test to
crash.
Found by dumb luck. It's amazing this hasn't shown up before.
llvm-svn: 98940
buffer was invalid when it was created, and use that bit to always set
the "Invalid" flag according to whether the buffer is invalid. This
ensures that all accesses to an invalid buffer are marked invalid,
improving recovery.
llvm-svn: 98690
SourceManager's getBuffer() and, therefore, could fail, along with
Preprocessor::getSpelling(). Use the Invalid parameters in the literal
parsers (string, floating point, integral, character) to make them
robust against errors that stem from, e.g., PCH files that are not
consistent with the underlying file system.
I still need to audit every use caller to all of these routines, to
determine which ones need specific handling of error conditions.
llvm-svn: 98608
SourceManager's getBuffer() (and similar) operations. This abstract
can be used to force callers to cope with errors in getBuffer(), such
as missing files and changed files. Fix a bunch of callers to use the
new interface.
Add some very basic checks for file consistency (file size,
modification time) into ContentCache::getBuffer(), although these
checks don't help much until we've updated the main callers (e.g.,
SourceManager::getSpelling()).
llvm-svn: 98585
(even if it is defined). This fixes the issue of this function
returning '0' when SVN_VERSION is defined to be "".
Fixes: <rdar://problem/7663667>
llvm-svn: 97620
end-of-line source location when given a column number beyond the
length of the line, or an end-of-file source location when given a
line number beyond the length of the file. Previously, we would return
an invalid location.
llvm-svn: 97299
we attach diagnostics to translation units and code-completion
results, so they can be queried at any time.
To facilitate this, the new StoredDiagnostic class stores a diagnostic
in a serializable/deserializable form, and ASTUnit knows how to
capture diagnostics in this stored form. CIndex's CXDiagnostic is a
thin wrapper around StoredDiagnostic, providing a C interface to
stored or de-serialized diagnostics.
I've XFAIL'd one test case temporarily, because currently we end up
storing diagnostics in an ASTUnit that's never returned to the user
(because it contains errors). I'll introduce a temporary fix for this
soon; the real fix will be to allow us to return and query invalid ASTs.
llvm-svn: 96592
we'd add an offset from the spelling location space to the
instantiation location, which doesn't make sense and would
lead up to the text diagnostics crashing when presented with
non-sensical locations.
This fixes rdar://7597492, a crash on 255.vortex.
llvm-svn: 96004
their spelling location. This prevents warnings from being swallowed just
because the caret is on the first parenthesis in, say, NULL.
This is an experiment; the risk is that there might be a substantial number
of system headers which #define symbols to expressions which inherently cause
warnings. My theory is that that's rare enough that it can be worked
around case-by-case, and that producing useful warnings around NULL is worth
it. But I'm willing to accept that I might be empirically wrong.
llvm-svn: 95870
by setting the section of the generated global. This is an
optimization done by the code generator, and the code being
removed didn't handle the case when the string contained an
embedded nul (which the code generator does correctly
handle). This is rdar://7589850
llvm-svn: 95003
so that CIndex can report diagnostics through the normal mechanisms
even when executing Clang in a separate process. This applies both
when performing code completion and when using ASTs as an intermediary
for clang_createTranslationUnitFromSourceFile().
The serialized format is not perfect at the moment, because it does
not encapsulate macro-instantiation information. Instead, it maps all
source locations back to the instantiation location. However, it does
maintain source-range and fix-it information. To get perfect fidelity
from the serialized format would require serializing a large chunk of
the source manager; at present, it isn't clear if this code will live
long enough for that to matter.
llvm-svn: 94740
have it return a StringRef instead of an integer (to be more VCS
agnostic).
(2) Add getClangFullRepositoryVersion(), which contains an
amalgamation of the repository name and the revision.
(3) Change PCH to only emit the string returned by
getClangFullRepositoryVersion() instead of also emitting the value
of getClangSubversionRevision() (which has been removed). This is
functionally equivalent.
More cleanup to version string generation pending...
llvm-svn: 94231
- Correctly is in quotes, because we are following what I interpreted as GCC's
intent (which diverges from practice, naturally).
- Also, fix the arch define for arm1136jf-s.
llvm-svn: 91855
- In particular, it can claim features for itself instead of always passing them on to LLVM.
- This allows using the target features as a generic mechanism for passing target specific options to the TargetInfo instance, which may need them for initializing preprocessor defines, etc.
llvm-svn: 91753
inconsistent situations if we do, and they are not important for PCH performance
(which currently only needs the stats to construct the initial FileManager
entries).
- No test case, sorry, the machinations are too involved.
This occurs when, for example, the build makes a PCH file and has a header map
or a -I for a directory that does not yet exist. It is possible we will cache
the negative stat on that directory, and then in the build we will never find
header files inside that dir.
For PCH we don't need these stats anyway for performance, so this also makes PCH
files smaller w/ no loss. I hope to eventually eliminate the stat cache
entirely.
llvm-svn: 91082
files.
- The issue is that PCH uses a stat cache, which may reference files which have
been deleted or moved. In such cases ContentCache::getBuffer was returning 0
but most clients are incapable of dealing with this (i.e., they don't).
For the time being, resolve this issue by just making up some invalid file
contents and. Eventually we should detect that we are in an inconsistent
situation and error out with a nice message that the PCH is out of date.
llvm-svn: 90699
http://llvm.org/viewvc/llvm-project?view=rev&revision=71086
Note - This commit only includes the fix for:
<rdar://problem/6309338> slightly different error message format for Visual Studio.
The fix for <rdar://problem/6845623> from protocol to template. is separate/forthcoming.
llvm-svn: 90642
file. This is accomplished by introducing the notion of a "virtual"
file into the file manager, which provides a FileEntry* for a named
file whose size and modification time are known but which may not
exist on disk.
Added a cute little test that remaps both a .c file and a .h file it
includes to alternative files.
llvm-svn: 90329
files with the contents of an arbitrary memory buffer. Use this new
functionality to drastically clean up the way in which we handle file
truncation for code-completion: all of the truncation/completion logic
is now encapsulated in the preprocessor where it belongs
(<rdar://problem/7434737>).
llvm-svn: 90300
The later assumption is patently false, but this was already broken -- this situation is conceptually impossible, my feeling is we should fix SourceManager and friends to make it impossible in practice as well. However, we need to fix PR5662 and perhaps some other things involving memory buffers first. In the short term I'm pretty sure this is reliable.
Chris, Argiris, is this going to break anything that wasn't already broken?
llvm-svn: 90280
in diagnostics when we fail to open a file. This allows us to
report things like:
$ clang test.c -I.
test.c:2:10: fatal error: error opening file './foo.h': Permission denied
#include "foo.h"
^
llvm-svn: 90276
pass them down into the ArgToStringFn implementation. This allows
redundancy across operands to a diagnostic to be eliminated.
This isn't used yet, so no functionality change.
llvm-svn: 84602
only supporting a single stat cache. The immediate benefit of this
change is that we can now generate a PCH/AST file when including
another PCH file; in the future, the chain of stat caches will likely
be useful with multiple levels of PCH files.
llvm-svn: 84263
1) -fwritable-string does affect the non-utf16 version of cfstrings
just not the utf16 ones.
2) utf16 strings should always be marked constant, as the __TEXT segment
is readonly.
3) The name of the global doesn't matter, remove it from TargetInfo.
4) Trust the asmprinter to drop cstrings into the right section, like llvmgcc does now.
This fixes rdar://7115750
llvm-svn: 84077
what we found when we looked into <blah>", where <blah> is a
DeclContext*. We can now format DeclContext*'s in nice ways, e.g.,
"namespace N", "the global namespace", "'class Foo'".
This is part of PR3990, but we're not quite there yet.
llvm-svn: 84028
-code-completion-at=filename:line:column
which performs code completion at the specified location by truncating
the file at that position and enabling code completion. This approach
makes it possible to run multiple tests from a single test file, and
gives a more natural command-line interface.
llvm-svn: 82571
essence, code completion is triggered by a magic "code completion"
token produced by the lexer [*], which the parser recognizes at
certain points in the grammar. The parser then calls into the Action
object with the appropriate CodeCompletionXXX action.
Sema implements the CodeCompletionXXX callbacks by performing minimal
translation, then forwarding them to a CodeCompletionConsumer
subclass, which uses the results of semantic analysis to provide
code-completion results. At present, only a single, "printing" code
completion consumer is available, for regression testing and
debugging. However, the design is meant to permit other
code-completion consumers.
This initial commit contains two code-completion actions: one for
member access, e.g., "x." or "p->", and one for
nested-name-specifiers, e.g., "std::". More code-completion actions
will follow, along with improved gathering of code-completion results
for the various contexts.
[*] In the current -code-completion-dump testing/debugging mode, the
file is truncated at the completion point and EOF is translated into
"code completion".
llvm-svn: 82166
- Change TargetData string to match llvm-gcc.
- Some -target-abi support for 'apcs-gnu', most importantly the alignment of double and long long changes.
llvm-svn: 81732
qualified name does not actually refer into a class/class
template/class template partial specialization.
Improve printing of nested-name-specifiers to eliminate redudant
qualifiers. Also, make it possible to output a nested-name-specifier
through a DiagnosticBuilder, although there are relatively few places
that will use this leeway.
llvm-svn: 80056
1. Passing something that isn't a string used to cause:
"argument to annotate attribute was not a string literal"
make it say "section attribute" instead.
2. Fix the location of the above message to point to the
bad argument instead of the section token.
3. Implement rdar://4341926, by diagnosing invalid section
specifiers in the frontend rather than letting them slip all
the way to the assembler (a QoI win).
An example of #3 is that we used to produce something like this:
/var/folders/n7/n7Yno9ihEm894640nJdSQU+++TI/-Tmp-//ccFPFGtT.s:2:Expected comma after segment-name
/var/folders/n7/n7Yno9ihEm894640nJdSQU+++TI/-Tmp-//ccFPFGtT.s:2:Rest of line ignored. 1st junk character valued 46 (.).
Daniel improved clang to use llvm_report_error, so now we got:
$ clang t.c -c
fatal error: error in backend: Global variable 'x' has an invalid section specifier 'sadf': mach-o section specifier
requires a segment and section separated by a comma.
with no loc info. Now we get:
$ clang t.c -fsyntax-only
t.c:4:30: error: argument to 'section' attribute is not valid for this target: mach-o section specifier requires a segment
and section separated by a comma
int x __attribute__((section("sadf")));
^
which is nice :)
llvm-svn: 78586
Note that I'm guessing that *BSD and Solaris do the same thing as Linux
here, but it's quite possible I'm wrong; if the following testcase
gives an error on x86-64 with gcc for any of those operating systems, please
tell me:
#include <stdint.h>
int64_t x; long x;
llvm-svn: 74583
This unifies all the targets supported by an OS into a template.
It also cleans up the differences between the darwin targets.
Also __LP64__ wasn't needed for *BSD, since x86-64 target defines it anyway.
llvm-svn: 74532
This is simple enough, but then I thought it would be nice to make PrintingPolicy
get a LangOptions so that various things can key off "bool" and "C++" independently.
This spiraled out of control. There are many fixme's, but I think things are slightly
better than they were before.
One thing that can be improved: CFG should probably have an ASTContext pointer in it,
which would simplify its clients.
llvm-svn: 74493
function attributes. There are predefined macros that are defined when stack
protectors are used: __SSP__=1 with -fstack-protector and __SSP_ALL__=2 with
-fstack-protector-all.
llvm-svn: 74405
registered when PCH wasn't being used. We should always install (in BuiltinInfo)
information about target-specific builtins, but we shouldn't register any builtin
identifier infos. This fixes the build of apps that use PCH and target specific
builtins together.
llvm-svn: 73492
Implement support for C++ Substitution Failure Is Not An Error
(SFINAE), which says that errors that occur during template argument
deduction do *not* produce diagnostics and do not necessarily make a
program ill-formed. Instead, template argument deduction silently
fails. This is currently implemented for template argument deduction
during matching of class template partial specializations, although
the mechanism will also apply to template argument deduction for
function templates. The scheme is simple:
- If we are in a template argument deduction context, any diagnostic
that is considered a SFINAE error (or warning) will be
suppressed. The error will be propagated up the call stack via the
normal means.
- By default, all warnings and errors are SFINAE errors. Add the
NoSFINAE class to a diagnostic in the .td file to make it a hard
error (e.g., for access-control violations).
Note that, to make this fully work, every place in Sema that emits an
error *and then immediately recovers* will need to check
Sema::isSFINAEContext() to determine whether it must immediately
return an error rather than recovering.
llvm-svn: 73332
preprocessor and initialize it early in clang-cc. This
ensures that __has_builtin works in all modes, not just
when ASTContext is around.
llvm-svn: 73319
- Apologies for the extremely gross code duplication, I want to get
this working and then decide how to get this information out of the
back end.
- This replaces -m[no-]sse4[12] by -m[no-]sse4, it appears gcc
doesn't distinguish them?
- -msse, etc. now properly disable/enable related features.
- Don't always define __SSE3__...
- The main missing functionality bit here is that we don't initialize
the features based on the CPU for all -march options.
llvm-svn: 71117
- Default to yonah on Darwin (to get SSE3).
- Default to Pentium4 (32-bit) and x86-64 (64-bit) on
non-Darwin. Welcome to the 21st century.
llvm-svn: 71069
- This is a WIP...
- This adds -march= handling to the driver, and fixes the defaulting
of -mcpu on Darwin (which was using the wrong test).
Instead of handling -m{sse, ...} in the driver, pass them to clang-cc as
-target-feature [+-]name
In clang-cc, communicate with the (clang) target to discover the legal
features of a target, and the features which are enabled based on
-mcpu. This is currently hardcoded just enough to not be a feature
regression, we need to get this information from the backend's
TableGen information somehow.
This is used to construct the full list of features which are being
used, which is in turn used to initialize the predefines.
llvm-svn: 71061
them with -Werror. Custom diags cannot be mapped, and this makes
-Werror cause a determinstic crash for the checker and other
clients of the custom diagnostics machinery. rdar://6816191
llvm-svn: 70639
scheme to be more useful.
The new scheme introduces a set of categories that should be more
readable, and also reflects what we want to consider as an extension
more accurately. Specifically, it makes the "what is a keyword"
determination accurately reflect whether the keyword is a GNU or
Microsoft extension.
I also introduced separate flags for keyword aliases; this is useful
because the classification of the aliases is mostly unrelated to the
classification of the original keyword.
This patch treats anything that's in the implementation
namespace (prefixed with "__", or "_X" where "X" is any upper-case
letter) as a keyword without marking it as an extension. This is
consistent with the standards in that an implementation is allowed to define
arbitrary extensions in the implementation namespace without violating
the standard. This gets rid of all the nasty "extension used" warnings
for stuff like __attribute__ in -pedantic mode. We still warn for
extensions outside of the the implementation namespace, like typeof.
If someone wants to implement -Wextensions or something like that, we
could add additional information to the keyword table.
This also removes processing for the unused "Boolean" language option;
such an extension isn't supported on any other C implementation, so I
don't see any point to adding it.
The changes to test/CodeGen/inline.c are required because previously, we
weren't actually disabling the "inline" keyword in -std=c89 mode.
I'll remove Boolean and NoExtensions from LangOptions in a follow-up
commit.
llvm-svn: 70281
essentially the same thing we do with pretokenized headers. stat()
caching improves performance of the Cocoa-prefixed "Hello, World" by
45%.
llvm-svn: 70223
file. In particular, only eagerly load source location entries for
files and for the predefines buffer. Other buffers and
macro-instantiation source location entries are loaded lazily.
With the Cocoa-prefixed "Hello, World", we only load 815/26555 source
location entities. This halves the amount of user time we spend in
this "Hello, World" program with -fsyntax-only (down to .007s).
This optimization is part 1 of 2 for the source manager. This
eliminates most of the user time in loading a PCH file. We still spend
too much time initialize File structures (especially in the calls to
stat), so we need to either make the loading of source location
entries for files lazy or import the stat cache from the PTH
implementation.
llvm-svn: 70196
- Deal with the Receiver/ClassInfo shared storage in ObjCMessageExpr
- Implement PCH support for ImplicitParamDecl
- Fix the handling of the body of an ObjCMethodDecl
- Several cast -> cast_or_null fixes
- Make Selector::getIdentifierInfoForSlot work for 1-argument, NULL
selectors.
- Make Selector::getAsString() work with NULL selectors.
- Fix the names of VisitObjCAtCatchStmt and VisitObjCAtFinallyStmt
in the PCH reader and writer; these were never getting called.
At this point, all of the pch-test tests pass for C and Objective-C.
llvm-svn: 70163
PCH file. In the Cocoa-prefixed "Hello, World" benchmark, this takes
us from reading 503 identifiers down to 37 and from 470 macros down to
4. It also results in an 8% performance improvement.
llvm-svn: 70094
Let me know if I messed up for some target. Note that for Windows, we
should be able to support it (MSVC supports "__declspec(thread)"), but
I'm pretty sure LLVM doesn't know how to generate the correct code.
llvm-svn: 69552
functions in glibc header files that use FP Stack inline asm which the
backend can't deal with (PR879).
This "fixes" PR3970 for linux. Other affected systems should do similar
things. Maybe this should just go to the general i386/x86-64 sections?
llvm-svn: 69527
with other diagnostic mapping. In the new scheme, -Wfoo or -Wno-foo or
-Werror=foo all override the -pedantic options, and __extension__
robustly silences all extension diagnostics in their scope.
An added bonus of this change is that MAP_DEFAULT goes away, meaning that
per-diagnostic mapping information can now be stored in 2 bits, doubling
the density of the Diagnostic::DiagMapping array. This also
substantially simplifies Diagnostic::getDiagnosticLevel.
OTOH, this temporarily introduces some "macro intensive" code in
Diagnostic.cpp. This will be addressed in a later patch.
llvm-svn: 69154
- Patch by Shantonu Sen (with a minor tweak to split out
getDarwin{OSX,IPhoneOS}Defines)!
- <rdar://problem/6776277> Need clang-cc/ccc-analyzer support for
-miphoneos-version-min
llvm-svn: 68815
and are even set in C mode. As such, move them to Targets.cpp.
__OBJC_GC__ is also darwin specific, but seems reasonable to always
define it when in objc-gc mode.
This fixes rdar://6761450
llvm-svn: 68494
#ifdef'd out the 5 conversion routines that we don't currently need.
Still need a bit more work in GetAddrOfConstantCFString(). Added a FIXME to indicate this.
Expect to remove the FIXME today...
llvm-svn: 68208
- Notably, set section on cfstring literal string data (for now, this
is done everywhere because it matches what we were already doing
for the CFString data itself)
- <rdar://problem/6599098> [irgen] linker requires objc string data
to go into cstring
llvm-svn: 68160
diagnostics following those notes.
Make exceeding the template instantiation depth a fatal error.
Thanks to Daniel for pointing out the problem!
llvm-svn: 67320
diagnostics. This builds on the patch that Sebastian committed and
then revert. Major differences are:
- We don't remove or use the current ".def" files. Instead, for now,
we just make sure that we're building the ".inc" files.
- Fixed CMake makefiles to run TableGen and build the ".inc" files
when needed. Tested with both the Xcode and Makefile generators
provided by CMake, so it should be solid.
- Fixed normal makefiles to handle out-of-source builds that involve
the ".inc" files.
I'll send a separate patch to the list with Sebastian's changes that
eliminate the use of the .def files.
llvm-svn: 67058
to a diagnostic that will be invoked after the diagnostic (if it is
not suppressed). The hooks are allowed to produce additional
diagnostics (typically notes) that provide more information. We should
be able to use this to help diagnostic clients link notes back to the
diagnostic they clarify. Comments welcome; I'll write up documentation
and convert other clients (e.g., overload resolution failures) if
there are no screams of protest.
As the first client of post-diagnostic hooks, we now produce a
template instantiation backtrace when a failure occurs during template
instantiation. There's still more work to do to make this output
pretty, if that's even possible.
llvm-svn: 66557
- Make Selector::getAsIdentifierInfo() private. Using IdentifierInfo* in
Selector is an implementation detail that clients shouldn't think about.
- Modify diagnostic emission in Sema::ProcessPropertyDecl to not use
Selector::getAsIdentifierInfo() (which could crash when IdentifierInfo* is
null) and instead use Selector::getAsString().
- Tidy up Selector::getAsString() implementation.
llvm-svn: 66313
parser. For example, we now print out:
0. t.c:5:10: in compound statement {}
1. t.c:3:12: in compound statement {}
2. clang t.c -fsyntax-only
llvm-svn: 66108
This has two advantages 1) no more leaking them, 2) fewer calls to malloc.
This changes us from calling malloc 3685/1390/883/2974/1185 times respectively on
16/20/24/28/32 byte objects when parsing cocoa.h with pth and -disable-free to
calling it 2816/1020/702/2903/1168 times each respectively.
llvm-svn: 66017
and defining target-specific macros based on them (like __SSE3__ and
friends). After extensive discussion with Daniel, this work will need
driver support, which will translate things like -msse3 into a -mattr
feature. Until this work is done, the code in clang.cpp is disabled and
the X86TargetInfo ctor still defaults to SSE2. With these two things
changed, this code will work. PR3634
llvm-svn: 65966
know how to recover from an error, we can attach a hint to the
diagnostic that states how to modify the code, which can be one of:
- Insert some new code (a text string) at a particular source
location
- Remove the code within a given range
- Replace the code within a given range with some new code (a text
string)
Right now, we use these hints to annotate diagnostic information. For
example, if one uses the '>>' in a template argument in C++98, as in
this code:
template<int I> class B { };
B<1000 >> 2> *b1;
we'll warn that the behavior will change in C++0x. The fix is to
insert parenthese, so we use code insertion annotations to illustrate
where the parentheses go:
test.cpp:10:10: warning: use of right-shift operator ('>>') in template
argument will require parentheses in C++0x
B<1000 >> 2> *b1;
^
( )
Use of these annotations is partially implemented for HTML
diagnostics, but it's not (yet) producing valid HTML, which may be
related to PR2386, so it has been #if 0'd out.
In this future, we could consider hooking this mechanism up to the
rewriter to actually try to fix these problems during compilation (or,
after a compilation whose only errors have fixes). For now, however, I
suggest that we use these code modification hints whenever we can, so
that we get better diagnostics now and will have better coverage when
we find better ways to use this information.
This also fixes PR3410 by placing the complaint about missing tokens
just after the previous token (rather than at the location of the next
token).
llvm-svn: 65570
information about types. We often print diagnostics where we say
"foo_t" is bad, but the user doesn't know how foo_t is declared
(because it is a typedef). Fix this by expanding sugar when present
in a diagnostic (and not one of a few special cases, like vectors).
Before:
t.m:5:2: error: invalid operands to binary expression ('typeof(P)' and 'typeof(F)')
MAX(P, F);
^~~~~~~~~
t.m:1:78: note: instantiated from:
#define MAX(A,B) ({ __typeof__(A) __a = (A); __typeof__(B) __b = (B); __a < __b ? __b : __a; })
^
After:
t.m:5:2: error: invalid operands to binary expression ('typeof(P)' (aka 'struct mystruct') and 'typeof(F)' (aka 'float'))
MAX(P, F);
^~~~~~~~~
t.m:1:78: note: instantiated from:
#define MAX(A,B) ({ __typeof__(A) __a = (A); __typeof__(B) __b = (B); __a < __b ? __b : __a; })
^
llvm-svn: 65081
as the last non-note diagnostic that preceeded them. This ensures
that diagnostics in main files which have notes with locations in
system headers get all the bits and pieces emitted or not in a
unit. This fixes PR3215.
llvm-svn: 64746
Now instead of just tracking the expansion history, also track the full
range of the macro that got replaced. For object-like macros, this doesn't
change anything. For _Pragma and function-like macros, this means we track
the locations of the ')'.
This is required for PR3579 because apparently GCC uses the line of the ')'
of a function-like macro as the location to expand __LINE__ to.
llvm-svn: 64601
a target.
Make Preprocessor.cpp define a new __INTPTR_TYPE__ macro based on this.
On linux/32, set intptr_t to int, instead of long. This fixes PR3563.
llvm-svn: 64495
- set the 'StatSysCallCache' object using a setter method instead of
FileManager's constructor. This allows the cache to be installed after the
FileManager object is created.
- Add 'file mode' to FileEntry (useful for stat caching)
llvm-svn: 64351
for use by FileManager. FileManager now takes a StatSysCallCache* in its
constructor (which defaults to NULL). This will be used for evaluating whether
or not caching 'stat' system calls in PTH is a performance win. This shim adds
no observable performance impact in the case where the 'StatSysCallCache*' is
null.
llvm-svn: 64345
- Changes Lookup*Name functions to return NamedDecls, instead of
Decls. Unfortunately my recent statement that it will simplify lot of
code, was not quite right, but it simplifies some...
- Makes MergeLookupResult SmallPtrSet instead of vector, following
Douglas suggestions.
- Adds %qN format for printing qualified names to Diagnostic.
- Avoids searching for using-directives in Scopes, which are not
DeclScope, during unqualified name lookup.
llvm-svn: 63739
line markers, including maintenance of the virtual include stack.
For something like this:
# 42 "bar.c" 1
# 142 "bar2.c" 1
#warning zappa
# 92 "bar.c" 2
#warning gonzo
# 102 "foo.c" 2
#warning bonkta
we now produce these three warnings:
#1:
In file included from foo.c:3:
In file included from bar.c:42:
bar2.c:143:2: warning: #warning zappa
#warning zappa
^
#2:
In file included from foo.c:3:
bar.c:92:2: warning: #warning gonzo
#warning gonzo
^
#3:
foo.c:102:2: warning: #warning bonkta
#warning bonkta
^
llvm-svn: 63722
location below it report as coming from the #line location. For example,
with:
#line 92 "foo.h"
#warning blarg!
#warning blarg!
we now emit:
foo.h:92:2: warning: #warning blarg!
#warning blarg!
^
foo.h:92:2: warning: #warning blarg!
#warning blarg!
^
llvm-svn: 63709
makes it clear to clients that they have to pick an instantiation
or spelling location before calling it and allows optimization based
on that.
llvm-svn: 63698
ContentCache objects to using a densemap and list, and allocating
the ContentCache objects from a bump pointer. This does not speed
up or slow down things substantially, but gives us control over
their alignment.
llvm-svn: 63628
.def file for each library. This means that adding a diagnostic
to sema doesn't require all the other libraries to be rebuilt.
Patch by Anders Johnsen!
llvm-svn: 63111
as reported to the user and as manipulated by #line. This is what __FILE__,
__INCLUDE_LEVEL__, diagnostics and other things should follow (but not
dependency generation!).
This patch also includes several cleanups along the way:
- SourceLocation now has a dump method, and several other places
that did similar things now use it.
- I cleaned up some code in AnalysisConsumer, but it should probably be
simplified further now that NamedDecl is better.
- TextDiagnosticPrinter is now simplified and cleaned up a bit.
This patch is a prerequisite for #line, but does not actually provide
any #line functionality.
llvm-svn: 63098
address space we used up. Some interesting data:
For c99-intconst-1.c:
6912762 SLocEntry's allocated, 25592386B of Sloc address space used.
For cocoa.h:
26469 SLocEntry's allocated, 10278752B of Sloc address space used.
For carbon.h:
27364 SLocEntry's allocated, 12398141B of Sloc address space used.
Clearly 2G of sloc address space should be enough for anyone?!
llvm-svn: 63093
ground work for implementing #line, and fixes the "out of macro ID's"
problem.
There is nothing particularly tricky about the code, other than the
very performance sensitive SourceManager::getFileID() method.
llvm-svn: 62978
Implemented DiagCollector::IncludeInDiagnosticCounts() to return 'false' so that the batching of diagnostics for use with BugReporter doesn't mess up the count of real diagnostics.
llvm-svn: 62873
tells us whether Preprocessor::HandleIdentifier needs to be called.
Because this method is only rarely needed, this saves a call and a
bunch of random checks. This drops the time in HandleIdentifier
from 3.52ms to .98ms on cocoa.h on my machine.
llvm-svn: 62675
Changes to IdentifierTable:
- High-level summary: StringMap never owns IdentifierInfos. It just
references them.
- The string map now has StringMapEntry<IdentifierInfo*> instead of
StringMapEntry<IdentifierInfo>. The IdentifierInfo object is
allocated using the same bump pointer allocator as used by the
StringMap.
Changes to IdentifierInfo:
- Added an extra pointer to point to the
StringMapEntry<IdentifierInfo*> in the string map. This pointer
will be null if the IdentifierInfo* is *only* used by the PTHLexer
(that is it isn't in the StringMap).
Algorithmic changes:
- Non-PTH case:
IdentifierInfo::get() will always consult the StringMap first to
see if we have an IdentifierInfo object. If that StringMapEntry
references a null pointer, we allocate a new one from the BumpPtrAllocator
and update the reference in the StringMapEntry.
- PTH case:
We do the same lookup as with the non-PTH case, but if we don't get
a hit in the StringMap we do a secondary lookup in the PTHManager for
the IdentifierInfo. If we don't find an IdentifierInfo we create a
new one as in the non-PTH case. If we do find and IdentifierInfo
in the PTHManager, we update the StringMapEntry to refer to it so
that the IdentifierInfo will be found on the next StringMap lookup.
This way we only do a binary search in the PTH file at most once
for a given IdentifierInfo. This greatly speeds things up for source
files containing a non-trivial amount of code.
Performance impact:
While these changes do add some extra indirection in
IdentifierTable to access an IdentifierInfo*, I saw speedups even
in the non-PTH case as well.
Non-PTH: For -fsyntax-only on Cocoa.h, we see a 6% speedup.
PTH (with Cocoa.h in token cache): 11% speedup.
I also did an experiment where we did -fsyntax-only on a source file
including a large header and Cocoa.h, but the token cache did not
contain the larger header. For this file, we were seeing a performance
*regression* when using PTH of 3% over non-PTH. Now we are seeing
a performance improvement of 9%!
Tests:
The serialization tests are now failing. I looked at this extensively,
and I my belief is that this change is unmasking a bug rather than
introducing a new one. I have disabled the serialization tests for now.
llvm-svn: 62636
Chris Lattner