If you pass an invalid compiler/debugger path on the cmdline to `dotest.py` this is what you get.
Traceback (most recent call last):
[...]
File "dotest.py", line 7, in <module>
lldbsuite.test.run_suite()
[...]
And with the patch applied:
/home/davide/work/build-lldb/bin/clandasfasg is not a valid path, exiting
Differential Revision: https://reviews.llvm.org/D39199
llvm-svn: 316393
Since the ipv6 patch, we've experienced occasional flakyness in
lldb-server tests. This was due to the fact that lldb-server was trying
to listen on both v4 and v6 localhost sockets (and consider it a success
if at least one of them succeeded), while the test framework was only
trying to connect to the v4 one.
This change makes sure lldb-server only listens on the v4 socket.
llvm-svn: 316391
The core of this change is the new CommandInterpreter::m_command_state,
which models the state transitions for interactive commands, including
an "interrupted" state transition.
In general, command interruption requires cooperation from the code
executing the command, which needs to poll for interruption requests
through CommandInterpreter::WasInterrupted().
CommandInterpreter::PrintCommandOutput() implements an optionally
interruptible printing of the command output, which for large outputs
was likely the longest blocking part.
(ex. target modules dump symtab on a complex binary could take 10+ minutes)
Differential Revision: https://reviews.llvm.org/D37923
llvm-svn: 315037
running watchos. These tests cannot run on normal customer devices,
but I hope to some day have a public facing bot running against a
device.
llvm-svn: 314355
in TestDataFormatterSkipSummary.py - I'm building this test
with the default c++ library.
Skip TestMTCSimple.py when running for i386.
llvm-svn: 314155
Normal customer devices won't be able to run these tests, we're hoping to get
a public facing bot set up at some point. Both devices pass the testsuite without
any errors or failures.
I have seen some instability with the armv7 test runs, I may submit additional patches
to address this. arm64 looks good.
I'll be watching the bots for the rest of today; if any problems are introduced by
this patch I'll revert it - if anyone sees a problem with their bot that I don't
see, please do the same. I know it's a rather large patch.
One change I had to make specifically for iOS devices was that debugserver can't
create files. There were several tests that launch the inferior process redirecting
its output to a file, then they retrieve the file. They were not trying to test
file redirection in these tests, so I rewrote those to write their output to a file
directly.
llvm-svn: 314132
Normal customer devices won't be able to run these devices, we're hoping to get
a public facing bot set up at some point. Both devices pass the testsuite without
any errors or failures.
I have seen some instability with the armv7 test runs, I may submit additional patches
to address this. arm64 looks good.
I'll be watching the bots for the rest of today; if any problems are introduced by
this patch I'll revert it - if anyone sees a problem with their bot that I don't
see, please do the same. I know it's a rather large patch.
One change I had to make specifically for iOS devices was that debugserver can't
create files. There were several tests that launch the inferior process redirecting
its output to a file, then they retrieve the file. They were not trying to test
file redirection in these tests, so I rewrote those to write their output to a file
directly.
llvm-svn: 314038
Normal customer devices won't be able to run these devices, we're hoping to get
a public facing bot set up at some point.
There will be some smaller follow-on patches. The changes to tools/lldb-server are
verbose and I'm not thrilled with having to skip all of these tests manually.
There are a few places where I'm making the assumption that "armv7", "armv7k", "arm64"
means it's an ios device, and I need to review & clean these up with an OS check
as well. (Android will show up as "arm" and "aarch64" so by pure luck they shouldn't
cause problems, but it's not an assumption I want to rely on).
I'll be watching the bots for the rest of today; if any problems are introduced by
this patch I'll revert it - if anyone sees a problem with their bot that I don't
see, please do the same. I know it's a rather large patch.
One change I had to make specifically for iOS devices was that debugserver can't
create files. There were several tests that launch the inferior process redirecting
its output to a file, then they retrieve the file. They were not trying to test
file redirection in these tests, so I rewrote those to write their output to a file
directly.
llvm-svn: 313932
The main change is to avoid setting the process state as running when
debugging core/minidumps (details in the bug). Also included a few small,
related fixes around how the errors propagate in this case.
Fixed the FreeBSD/Windows break: the intention was to keep
Process::WillResume() and Process::DoResume() "in-sync", but this had the
unfortunate consequence of breaking Process sub-classes which don't override
WillResume().
The safer approach is to keep Process::WillResume() untouched and only
override it in the minidump and core implementations.
patch by lemo
Bug: https://bugs.llvm.org/show_bug.cgi?id=34532
Differential Revision: https://reviews.llvm.org/D37651
llvm-svn: 313655
When introduced, breakpoint names were just tags that you could
apply to breakpoints that would allow you to refer to a breakpoint
when you couldn't capture the ID, or to refer to a collection of
breakpoints.
This change makes the names independent holders of breakpoint options
that you can then apply to breakpoints when you add the name to the
breakpoint. It adds the "breakpoint name configure" command to set
up or reconfigure breakpoint names. There is also full support for
then in the SB API, including a new SBBreakpointName class.
The connection between the name and the breakpoints
sharing the name remains live, so if you reconfigure the name, all the
breakpoint options all change as well. This allows a quick way
to share complex breakpoint behavior among a bunch of breakpoints, and
a convenient way to iterate on the set.
You can also create a name from a breakpoint, allowing a quick way
to copy options from one breakpoint to another.
I also added the ability to make hidden and delete/disable protected
names. When applied to a breakpoint, you will only be able to list,
delete or disable that breakpoint if you refer to it explicitly by ID.
This feature will allow GUI's that need to use breakpoints for their
own purposes to keep their breakpoints from getting accidentally
disabled or deleted.
<rdar://problem/22094452>
llvm-svn: 313292
The auto-continue test was using the new (better) name
for providing commands (-C) but I haven't checked in that change
yet. Put the test back to the old way for now.
llvm-svn: 313221
The main change is to avoid setting the process state as running when
debugging core/minidumps (details in the bug).
Also included a few small, related fixes around how the errors propagate in
this case.
patch by lemo
Bug: https://bugs.llvm.org/show_bug.cgi?id=34532
Differential Revision: https://reviews.llvm.org/D37651
llvm-svn: 313210
On Windows a process can't delete its own current direcotry, that's why the test
needs to return to the original direcotry before removing newdir.
llvm-svn: 313113
Summary:
Test was skipped because -data-evaluate-expression was thought
to not work on globals. This is not the case - the issue was clang
removes debug info for globals in cpp files that are not used.
Add a reference to the globals in question, and fix memory patter in
test to match memory pattern in testcase.
Reviewers: ki.stfu, abidh
Reviewed By: ki.stfu
Subscribers: aprantl, lldb-commits
Differential Revision: https://reviews.llvm.org/D37533
llvm-svn: 312726
The FXSAVE member `ftw` (FPU Tag Word) was given the wrong size (8-bit)
instead of the correct width (16-bit) as per the x87 Programmer's
Manual. Adjust this to ensure that we print out the complete value for
the register.
llvm-svn: 311579
The Linux xfail decorator was removed in r272326 with the claim that the
test "runs reliably on the linux x86 buildbot." It also runs reliably on
FreeBSD for me.
llvm.org/pr25925
llvm-svn: 310644
This test is consistently reporting unexpected pass for me on FreeBSD
10 and 12. It was failing on the old FreeBSD buildbot which has now been
retired for some time. Will investigate further if this fails once a new
buildbot is configured and running tests.
llvm.org/pr17807
llvm-svn: 310626
This is the FreeBSD equivalent of r238549.
This serves 2 purposes:
* LLDB should handle inferior process signals SIGSEGV/SIGILL/SIGBUS/
SIGFPE the way it is suppose to be handled. Prior to this fix these
signals will neither create a coredump, nor exit from the debugger
or work for signal handling scenario.
* eInvalidCrashReason need not report "unknown crash reason" if we have
a valid si_signo
llvm.org/pr23699
Patch by Karnajit Wangkhem
Differential Revision: https://reviews.llvm.org/D35223
llvm-svn: 310591
Summary:
The available platform list was previously only accessible via the
`platform list` command, this patch makes it possible to access that
list via the SBDebugger API. The active platform list has likewise
been exposed via the SBDebugger API.
Differential Revision: https://reviews.llvm.org/D35760
llvm-svn: 310452
You can get a breakpoint to auto-continue by adding "continue"
as a command, but that has the disadvantage that if you hit two
breakpoints simultaneously, the continue will force the process
to continue, and maybe even forstalling the commands on the other.
The auto-continue flag means the breakpoints can negotiate about
whether to stop.
Writing tests, I wanted to supply some commands when I made the
breakpoints, so I also added that ability.
llvm-svn: 309969
When an option was set at on a location, I was just copying the whole option set
to the location, and letting it shadow the breakpoint options. That was wrong since
it meant changes to unrelated options on the breakpoint would no longer take on this
location. I added a mask of set options and use that for option propagation.
I also added a "location" property to breakpoints, and added SBBreakpointLocation.{G,S}etCommandLineCommands
since I wanted to use them to write some more test cases.
<rdar://problem/24397798>
llvm-svn: 309772
Summary:
Implement SBProcessInfo to wrap lldb_private::ProcessInstanceInfo,
and add SBProcess::GetProcessInfo() to retrieve info like parent ID,
group ID, user ID etc. from a live process.
Differential Revision: https://reviews.llvm.org/D35881
llvm-svn: 309664
Now -shlib flag can be provided alongside with names of symbols files:
(lldb) target symbols add --shlib stripper-lib.so unstripper-lib.so
This is helpful when default matching mechanisms by name and UUID
can't find a module, and the user needs to explicitly specify
which module the given symbol file belongs to.
Differential Revision: https://reviews.llvm.org/D35607
llvm-svn: 308933
This reapplies https://reviews.llvm.org/D35740 with a tweak to find
the section by name rather than type. Section types don't distinguish
between regular sections and their DWO counterparts.
llvm-svn: 308905
The DWO handling code can get confused by clang modules which also use
skeleton CUs to point to the object file with the full debug
info. This patch detects whether an object is a "real" DWO or a clang
module and prevents LLDB from interpreting clang modules as DWO. This
fixes the regression in TestWithModuleDebugging.
http://llvm.org/bugs/show_bug.cgi?id=33875
Differential Revision: https://reviews.llvm.org/D35740
llvm-svn: 308850
Summary:
SBBreakpointLocation exposed the ignore count, but didn't expose
the hit count. Both values were exposed by SBBreakpoint and
SBWatchpoint, so this makes things a bit more consistent.
Reviewers: lldb-commits
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D31283
llvm-svn: 308480
In NativeProcessLinux::MonitorSIGTRAP we were asserting that the si_code
value is one of the codes we know about. However, that list was very
incomplete -- for example, we were not handling SI_TKILL/SI_USER,
generated by raise(SIGTRAP). A cursory examination show there are at
least a dozen codes like these that an app can generate, and more can be
added at any point.
So, instead of trying to play catchup, I change the default behavior to
treat an unknown si_code like an ordinary signal. The only reason we
needed to inspect si_code in the first place is because
watchpoint/breakpoints are notified as SIGTRAP, but we already know
about those, and us starting to use a new debug event is far less likely
than somebody introducing a new non-debug event.
I add a test case to TestRaise to verify we are handling raise(SIGTRAP)
in an application properly.
llvm-svn: 307644
Starting with android ndk r15, clang much more tests are affected by the
-mstackrealign bugl (now nearly all functions are affected, and not just
the ones requiring 16-byte alignment). Due to their numbers, Xfailing
all of them is not a viable option, so we will just have to declare this
configuration unsupported, and wait until ndk ships a clang version that
has this bug fixed.
llvm-svn: 307252
to make a target, set a source regex breakpoint, run to
the breakpoint and find the thread that hit the breakpoint.
Start the process of replacing the boiler plate with this
routine.
llvm-svn: 307234
This adds a simple testcase for MainThreadCheckerRuntime. The tool (Main Thread Checker) is only available on Darwin, so the test also detects the presence of libMainThreadChecker.dylib and is skipped if the tool is not available.
llvm-svn: 307170
All android builds systems have switched to -mstackrealign for building
x86 binaries, so follow their cue with our mini build system.
This presently breaks just one test (TestReturnValue), and this is due
to a compiler bug, which has already been fixed in clang, but it hasn't
made it yet into the official NDK compiler. While I'm touching that
test, I also remove an android-specific XFAIL, which is not relevant
anymore.
llvm-svn: 306683
components to not depend on "." characters in the fileanme
(e.g. "Foundation.framework") but instead to just use path
separators. The names of the files themselves may have dots
in them ("com.apple.sbd") which would break the old scheme.
Also add a test case for this (macosx/find-dsym/bundle-with-dot-in-filename)
as well as a test case for r304520 (macosx/find-dsym/deep-bundle)
which needed a similar setup to test correctly on a single machine.
(both of these are really testing remote debug session situations
where the binary can't be found on the system where lldb is running,
complicating the test case a bit.)
<rdar://problem/31825940>
llvm-svn: 305441
NSString is loaded from the DWARF, which doesn't have the concept of protocols.
When this is used with the NSMutableDictionary type from Objective-C modules,
this produces errors of the form
error: cannot initialize a parameter of type 'id<NSCopying> _Nonnull' with an rvalue of type 'NSString *'
We're aware of these problems and have an internal bug report filed
(<rdar://problem/32777981>)
llvm-svn: 305424
The step count depends on code generated by compiler (GCC/Clang).
It will also vary for different MIPS arch variant. Hence skipping these test for MIPS.
Subscribers: jaydeep, bhushan, lldb-commits, slthakur
llvm-svn: 305383
Clang does not accept regparm attribute on these platforms.
Fortunately, the default calling convention passes arguments
in registers any way
Subscribers: jaydeep, bhushan, lldb-commits, slthakur
llvm-svn: 305378
This test started being flaky since r303848 (RunThreadPlan: Fix halting
logic in IgnoreBreakpoints = false). I am not reverting that, as I am
confident that actually fixed a problem. A more likely explanation is
that there is still one corner case that is not handled correctly there.
Marking the test as flaky until I get a chance to investigate. I also
mark the test as no-debug-info-dependend -- it stresses expression
evaluation, as far as debug info goes, the test if extremely simple.
llvm-svn: 305286
This patch introduces a new thread backtrace command "unique".
The command is based off of "thread backtrace all" but will instead
find all threads which share matching call stacks and de-duplicate
their output, listing call stack and all the threads which share it.
This is especially useful for apps which use thread/task pools
sitting around waiting for work and cause excessive duplicate output.
I needed this behavior recently when debugging a core with 700+ threads.
Differential Revision: https://reviews.llvm.org/D33426
Reviewers: clayborg, jingham
Patch by Brian Gianforcaro <b.gianfo@gmail.com>
llvm-svn: 305197
I was over-eager to unable this test in r304976. It still fails in this
combination, at there does not seem to be anything we can do about it,
as the generated code does not preserve the link register.
llvm-svn: 305062
Summary:
When a call instruction is the last instruction in a function, the
backtrace PC will point past the end of the function. We already had
special code to handle that, but we did not handle the case where the PC
ends up outside of the bounds of the module containing the function,
which is a situation that occured in TestNoreturnUnwind on android for
some arch/compiler combinations.
I fix this by adding an argument to Address resolution code which states
that we are ok with addresses pointing to the end of a module/section to
resolve to that module/section.
I create a reproducible test case for this situation by hand-crafting an
executable which has a noreturn function at the end of a module.
Reviewers: jasonmolenda, jingham
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D32022
llvm-svn: 304976
When parsing types originating in modules, it is possible to encounter AttributedTypes
(such as the type generated for NSString *_Nonnull). Some of LLDB's ClangASTContext
methods deal with them; others do not. In particular, one function that did not was
GetTypeInfo, causing TestObjCNewSyntax to fail.
This fixes that, treating AttributedType as essentially transparent and getting the
information for the modified type.
In addition, however, TestObjCNewSyntax is a monolithic test that verifies a bunch of
different things, all of which can break independently of one another. I broke it
apart into smaller tests so that we get more precise failures when something (like
this) breaks.
Differential Revision: https://reviews.llvm.org/D33812
llvm-svn: 304510
This works on SVN but is a bit fragile on the Swift branch.
I'm adding the test to both, so we have this path covered.
<rdar://problem/32372372>
llvm-svn: 304314
Summary:
- Added API to access data types
-- integer, double, array, string, boolean and dictionary data types
-- Earlier user had to parse through the string output to get these
values
- Added Test cases for API testing
- Added new StructuredDataType enum in public include file
-- Replaced locally-defined enum in StructuredData.h with this new
one
-- Modified other internal files using this locally-defined enum
Signed-off-by: Abhishek Aggarwal <abhishek.a.aggarwal@intel.com>
Reviewers: clayborg, lldb-commits
Reviewed By: clayborg
Subscribers: labath
Differential Revision: https://reviews.llvm.org/D33434
llvm-svn: 304138
This is a resubmit of r303732, which was reverted due to a regression.
The original patch caused a regression in TestLoadUnload, which has only showed
up when running the remote test suite. The problem there was that we interrupted
the target just as it has hit the rendezvous breakpoint in the dlopen call. This
meant that the stop reason was set to "breakpoint" even though the event would
not have been broadcast if we had not stopped the process. I fix this by
checking StopInfo->ShouldNotify() before stopping.
I also add a new test for the handling of conditional breakpoints in
expressions, which I noticed to be broken (pr33164)
Differential Revision: https://reviews.llvm.org/D33283
llvm-svn: 303848
Summary:
The function had logic to handle the case when the expression terminated
while we were trying to halt the process, but it failed to take into
account the possibility that the expression stopped because it hit a
breakpoint. This was caused by the fact that the handling of the stopped
events was duplicated for the "halting" and regular cases (the regular
case handled this situation correctly). I've tried to merge these two
cases into one to make sure they stay in sync.
I should call out that the two cases were checking whether the thread
plan has completed in slightly different ways. I am not sure what is the
difference between them, but I think the check should be the same in
both cases, whatever it is, so I just took the one from the regular
case, as that is probably more tested.
For the test, I modified TestUnwindExpression to run the expression with
a smaller timeout (this is how I found this bug originally). With a 1ms
one thread timeout, the test failed consistently without this patch.
Reviewers: jingham
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D33283
llvm-svn: 303732
For remote targets we need to call registerSharedLibrariesWithTarget to
make sure they are installed alongside main executable. This also
required a small fixup in the the mentioned function as in this case
"One" was both a directory name and a library name template. I fixed it
to make sure it checks that the string refers to a file before it
assumed it was a full library path.
llvm-svn: 303248
When it resolves symbol-only variables, the expression parser
currently looks only in the global module list. It should prefer
the current module.
I've fixed that behavior by making it search the current module
first, and only search globally if it finds nothing. I've also
added a test case.
After review, I moved the core of the lookup algorithm into
SymbolContext for use by other code that needs it.
Thanks to Greg Clayton and Pavel Labath for their help.
Differential Revision: https://reviews.llvm.org/D33083
llvm-svn: 303223
ptr_refs exposed a problem in ClangASTContext's implementation: it
uses an accessor to downcast a QualType to an
ObjCObjectPointerType, but the accessor is not fully general.
getAs() is the safer way to go.
I've added a test case that uses ptr_refs in a way that would
crash before the fix.
<rdar://problem/31363513>
llvm-svn: 303110
as described in pr33042, we cannot reliably retrieve the return value on
arm64 in cases it is returned via x8 pointer. I tried to do this as
surgically as possible and disabled it only on targets I know to be
affected, as the code is still useful, even though it can only work on
best-effort basis.
llvm-svn: 303076
Templates can end in parameter packs, like this
template <class T...> struct MyStruct
{ /*...*/ };
LLDB does not currently support these parameter packs;
it does not emit them into the template argument list
at all. This causes problems when you specialize, e.g.:
template <> struct MyStruct<int>
{ /*...*/ };
template <> struct MyStruct<int, int> : MyStruct<int>
{ /*...*/ };
LLDB generates two template specializations, each with
no template arguments, and then when they are imported
by the ASTImporter into a parser's AST context we get a
single specialization that inherits from itself,
causing Clang's record layout mechanism to smash its
stack.
This patch fixes the problem for classes and adds
tests. The tests for functions fail because Clang's
ASTImporter can't import them at the moment, so I've
xfailed that test.
Differential Revision: https://reviews.llvm.org/D33025
llvm-svn: 302833
Summary:
'arch' is a valid qHostInfo key, but the unit
test for qHostInfo did not include it in the set of possible keys.
Reviewers: tfiala, labath
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D32711
llvm-svn: 302260
Summary:
I have found a way to segfault lldb in 7 keystrokes! Steps to reproduce:
1) Launch lldb
2) Type `print` and hit enter. lldb will now prompt you to type a list of
expressions, followed by an empty line.
3) Hit enter, indicating the end of your input.
4) Segfault!
After some investigation, I've found the issue in Host/common/Editline.cpp.
Editline::MoveCursor() relies on m_input_lines not being empty when the `to`
argument is CursorPosition::BlockEnd. This scenario, as far as I can tell,
occurs in one specific instance: In Editline::EndOrAddLineCommand() when the
list of lines being processed contains exactly one string (""). Meeting this
condition is fairly simple, I have posted steps to reproduce above.
Reviewers: krytarowski, zturner, labath
Reviewed By: labath
Subscribers: scott.smith, lldb-commits
Differential Revision: https://reviews.llvm.org/D32421
Patch by Alex Langford.
llvm-svn: 302225
Summary:
Arm64 Procedure Call Standard specifies than only vectors up to 16 bytes
are stored in v0 (which makes sense, as that's the size of the
register). 32-byte vector types are passed as regular structs via x8
pointer. Treat them as such.
This fixes TestReturnValue for arm64-clang. I also split the test case
into two so I can avoid the if(gcc) line, and annotate each test
instead. (It seems the vector type tests fail with gcc only when
targetting x86 arches).
Reviewers: tberghammer, eugene
Subscribers: aemerson, omjavaid, rengolin, srhines, lldb-commits
Differential Revision: https://reviews.llvm.org/D32813
llvm-svn: 302220
Summary:
The existing cpp-level checks using PR_MPX_ENABLE_MANAGEMENT aren't sufficient,
as this isn't defined for linux kernel versions below 3.19.
Reviewers: valentinagiusti, zturner, labath
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D32719
llvm-svn: 302027
gnuwin32 rm does not like wildcards that match nothing even if we
specify -f (probably because the wildcard expansion happens in-process
there). We could use make $(wildcard) here, but it seems safer to
explicitly list the files here, just like the normal Makefile.rules
does.
llvm-svn: 302013
Loading a shared library can require a large amount of work; rather than do that serially for each library,
this patch will allow parallelization of the symbols and debug info name indexes.
From scott.smith@purestorage.comhttps://reviews.llvm.org/D32598
llvm-svn: 301609
Many times a user wants to access a type when there's a variable of
the same name, or a variable when there's a type of the same name.
Depending on the precise context, currently the expression parser
can fail to resolve one or the other.
This is because ClangExpressionDeclMap has logic to limit the
amount of information it searches, and that logic sometimes cuts
down the search prematurely. This patch removes some of those early
exits.
In that sense, this patch trades performance (early exit is faster)
for correctness.
I've also included two new test cases showing examples of this
behavior – as well as modifying an existing test case that gets it
wrong.
llvm-svn: 301273
LLDB uses clang::DeclContexts for lookups, and variables get put into
the DeclContext for their abstract origin. (The abstract origin is a
DWARF pointer that indicates the unique definition of inlined code.)
When the expression parser is looking for variables, it locates the
DeclContext for the current context. This needs to be done carefully,
though, e.g.:
__attribute__ ((always_inline)) void f(int a) {
{
int b = a * 2;
}
}
void g() {
f(3);
}
Here, if we're stopped in the inlined copy of f, we have to find the
DeclContext corresponding to the definition of f – its abstract
origin. Clang doesn't allow multiple functions with the same name and
arguments to exist. It also means that any variables we see must be
placed in the appropriate DeclContext.
[Bug 1]: When stopped in an inline block, the function
GetDeclContextDIEContainingDIE for that block doesn't properly
construct a DeclContext for the abstract origin for inlined
subroutines. That means we get duplicated function DeclContexts, but
function arguments only get put in the abstract origin's DeclContext,
and as a result when we try to look for them in nested contexts they
aren't found.
[Bug 2]: When stopped in an inline block, the DWARF (for space
reasons) doesn't explicitly point to the abstract origin for that
block. This means that the function GetClangDeclContextForDIE returns
a different DeclContext for each place the block is inlined. However,
any variables defined in the block have abstract origins, so they
will only get placed in the DeclContext for their abstract origin.
In this fix, I've introduced a test covering both of these issues,
and fixed them.
Bug 1 could be resolved simply by making sure we look up the abstract
origin for inlined functions when looking up their DeclContexts on
behalf of nested blocks.
For Bug 2, I've implemented an algorithm that makes the DeclContext
for a block be the containing DeclContext for the closest entity we
would find during lookup that has an abstract origin pointer. That
means that in the following situation:
{ // block 1
int a;
{ // block 2
int b;
}
}
if we looked up the DeclContext for block 2, we'd find the block
containing the abstract origin of b, and lookup would proceed
correctly because we'd see b and a. However, in the situation
{ // block 1
int a;
{ // block 2
}
}
since there isn't anything to look up in block 2, we can't determine
its abstract origin (and there is no such pointer in the DWARF for
blocks). However, we can walk up the parent chain and find a, and its
abstract origin lives in the abstract origin of block 1. So we simply
say that the DeclContext for block 2 is the same as the DeclContext
for block 1, which contains a. Lookups will return the same results.
Thanks to Jim Ingham for review and suggestions.
Differential revision: https://reviews.llvm.org/D32375
llvm-svn: 301263
Davide Italiano