build tree relative search path in order to locate the lldb.py module. When
'-i' is present, the test driver relies on the PYTHONPATH environment variable
to locate the lldb.py module.
llvm-svn: 114094
This will remove the confusion experienced when previous test runs left some
files (both intermediate or by-product as a result of the test).
lldbtest.TestBase defines a classmethod tearDownClass(cls) which invokes the
platform-specific cleanup() function as defined by the plugin; after that, it
invokes a subclass-specific function classCleanup(cls) if defined; and, finally,
it restores the old working directory.
An example of classCleanup(cls) is in settings/TestSettings.py:
@classmethod
def classCleanup(cls):
system(["/bin/sh", "-c", "rm output.txt"])
where it deletes the by-product "output.txt" as a result of running a.out.
llvm-svn: 114058
(lldb) settings set process.run-args A B C
(lldb) settings set process.env-vars ["MY_ENV_VAR"]=YES
commands. The main.cpp checks whether A, B, C is passed to main and whether
the $MY_ENV_VAR env variable is defined and outputs the findings to a file.
llvm-svn: 114031
lldb.py module. The priorities to search for are Debug, Release, then
BuildAndIntegration. You can always override this with a valid PYTHONPATH
environment variable before running the test driver.
For example:
$ PYTHONPATH=/Find/My/LLDB/Module/Here ./dotest.py -v .
Python runtime will try to locate the lldb.py module from
/Find/My/LLDB/Module/Here first before trying the Debug, Release, and then
BuildAndIntegration directories.
llvm-svn: 113991
find the hotspots in our code when indexing the DWARF. A combination of
using SmallVector to avoid collection allocations, using fixed form
sizes when possible, and optimizing the hot loops contributed to the
speedup.
llvm-svn: 113961
or a settings prefix, and it will list information about the subset of settings
you requested. Also added tab-completion (now that it takes an optional argument).
llvm-svn: 113952
Added a "bool show_fullpaths" to many more objects that were
previously always dumping full paths.
Fixed a few places where the DWARF was not indexed when we
we needed it to be when making queries. Also fixed an issue
where the DWARF in .o files wasn't searching all .o files
for the types.
Fixed an issue with the output from "image lookup --type <TYPENAME>"
where the name and byte size might not be resolved and might not
display. We now call the accessors so we end up seeing all of the
type info.
llvm-svn: 113951
interpreter from working. The communication read thread could
startup and immediately exit if m_read_thread_enabled was
checked in the thread function before it was set by the
thread that spawns the read thread. Now m_read_thread_enabled is set
to true prior to spawning the read thread to avoid this issue.
Hopefully this will clear up the sporatic failures in our test suite.
llvm-svn: 113947
all types in all compile units. I added a new kind of accelerator table to
the DWARF that allows us to index the DWARF compile units and DIEs in a way
that doesn't require the data to stay loaded. Currently when indexing the
DWARF we check if the compile unit had parsed its DIEs and if it hasn't we
index the data and free all of the DIEs so we can reparse later when we need
to after using one of our complete accelerator tables to determine we need
to reparse some DWARF. If the DIEs had already been parsed we leave them
loaded. The new accelerator table uses the "const char *" pointers from our
ConstString class as the keys, and NameToDIE::Info as the value. This info
contains the compile unit index and the DIE index which means we are pointed
right to the DIE we need unlike the other DWARF accelerator tables that often
just point us to the compile unit we would find our answer in.
llvm-svn: 113933
Added the ability to specify a preference for mangled or demangled to Mangled::GetName.
Changed one place where mangled was prefered in GetName.
The Dynamic loader should look up the target of a stub by mangled name if it exists.
llvm-svn: 113869
as it now passes. Added some extra tests to breakpoint_creation_by_filespec_python().
More clarification for the "os command" output and error as defined in
lldbtest.system() function.
Cleaned up the option processing of the test driver (dotest.py) and fixed the comment
about enabling gdb-remote logging. Example:
$ GDB_REMOTE_LOG=/tmp/log.txt ./dotest.py -v -t enum_types
llvm-svn: 113868
expressions. This involved three main changes:
- In ClangUserExpression::ClangUserExpression(),
we now insert the following lines into the
expression:
#define this ___clang_this
#define self ___clang_self
- In ClangExpressionDeclMap::GetDecls(), we
special-case ___clang_(this|self) and instead
look up "this" or "self"
- In ClangASTSource, we introduce the capability
to generate Decls with a different, overridden,
name from the one that was requested, e.g.
this for ___clang_this.
llvm-svn: 113866
to any inferior process because the code was checking if no run args were
set and then adding and empty string. This was happening for environment
vars as well.
llvm-svn: 113831
to return the correct result.
Fixed "bool Variable::IsInScope (StackFrame *frame)" to return the correct
result when there are no location lists.
Modified the "frame variable" command such that:
- if no arguments are given (dump all frame variables), then we only show
variables that are currently in scope
- if some arguments are given, we show an error if the variable is out of
scope
llvm-svn: 113830
debug map showed that the location lists in the .o files needed some
refactoring in order to work. The case that was failing was where a function
that was in the "__TEXT.__textcoal_nt" in the .o file, and in the
"__TEXT.__text" section in the main executable. This made symbol lookup fail
due to the way we were finding a real address in the debug map which was
by finding the section that the function was in in the .o file and trying to
find this in the main executable. Now the section list supports finding a
linked address in a section or any child sections. After fixing this, we ran
into issue that were due to DWARF and how it represents locations lists.
DWARF makes a list of address ranges and expressions that go along with those
address ranges. The location addresses are expressed in terms of a compile
unit address + offset. This works fine as long as nothing moves around. When
stuff moves around and offsets change between the remapped compile unit base
address and the new function address, then we can run into trouble. To deal
with this, we now store supply a location list slide amount to any location
list expressions that will allow us to make the location list addresses into
zero based offsets from the object that owns the location list (always a
function in our case).
With these fixes we can now re-link random address ranges inside the debugger
for use with our DWARF + debug map, incremental linking, and more.
Another issue that arose when doing the DWARF in the .o files was that GCC
4.2 emits a ".debug_aranges" that only mentions functions that are externally
visible. This makes .debug_aranges useless to us and we now generate a real
address range lookup table in the DWARF parser at the same time as we index
the name tables (that are needed because .debug_pubnames is just as useless).
llvm-gcc doesn't generate a .debug_aranges section, though this could be
fixed, we aren't going to rely upon it.
Renamed a bunch of "UINT_MAX" to "UINT32_MAX".
llvm-svn: 113829
was used to set the selected thread if none was selected. Use a more robust
API to accomplish the task.
Also fixed an error found, while investigating, in CommandObjectThreadSelect::
Execute() where the return status was not properly set if successful.
As a result, both the stl step-in test cases with expectedFailure decorators now
passed.
llvm-svn: 113825
- If you put a semicolon at the end of an expression,
this no longer causes the expression parser to
error out. This was a two-part fix: first,
ClangExpressionDeclMap::Materialize now handles
an empty struct (such as when there is no return
value); second, ASTResultSynthesizer walks backward
from the end of the ASTs until it reaches something
that's not a NullStmt.
- ClangExpressionVariable now properly byte-swaps when
printing itself.
- ClangUtilityFunction now cleans up after itself when
it's done compiling itself.
- Utility functions can now use external functions just
like user expressions.
- If you end your expression with a statement that does
not return a value, the expression now runs correctly
anyway.
Also, added the beginnings of an Objective-C object
validator function, which is neither installed nor used
as yet.
llvm-svn: 113789
the binaries.
If the build* function is passed the compiler argument, for example, 'llvm-gcc',
it is passed as a make variable to the make command. Otherwise, we check the
LLDB_CC environment variable; if it is defined, it is passed as a make variable
to the make command.
If neither the compiler keyword argument nor the LLDB_CC environment variable is
specified, no CC make variable is passed to the make command. The Makefile gets
to define the default CC being used.
--------------------------------------------------------------------------------
Example usage follows:
o Via the keyword argument:
class ArrayTypesTestCase(TestBase):
mydir = "array_types"
@unittest2.skipUnless(sys.platform.startswith("darwin"), "requires Darwin")
def test_with_dsym_and_run_command(self):
"""Test 'frame variable var_name' on some variables with array types."""
self.buildDsym(compiler='llvm-gcc')
self.array_types()
...
o Via LLDB_CC environment variable:
$ LLDB_CC=llvm-gcc ./dotest.py -v -t array_types
----------------------------------------------------------------------
Collected 4 tests
test_with_dsym_and_python_api (TestArrayTypes.ArrayTypesTestCase)
Use Python APIs to inspect variables with array types. ...
os command: [['/bin/sh', '-c', 'make clean; make MAKE_DSYM=YES CC=llvm-gcc']]
output: rm -rf "a.out" "a.out.dSYM" main.o main.d
llvm-gcc -arch x86_64 -gdwarf-2 -O0 -arch x86_64 -gdwarf-2 -O0 -c -o main.o main.c
llvm-gcc -arch x86_64 -gdwarf-2 -O0 main.o -o "a.out"
/usr/bin/dsymutil -o "a.out.dSYM" "a.out"
...
llvm-svn: 113781
no elements so that they at least have 1 element.
Added the ability to show the declaration location of variables to the
"frame variables" with the "--show-declaration" option ("-c" for short).
Changed the "frame variables" command over to use the value object code
so that we use the same code path as the public API does when accessing and
displaying variable values.
llvm-svn: 113733
static class array. It turns out that gcc 4.2 will emit DWARF that correctly
describes the PointType, but it will incorrectly emit debug info for the
"g_points" array where the following things are wrong:
- the DW_TAG_array_type won't have a subrange info
- the DW_TAG_variable for "g_points" won't have a valid byte size, so even
though we know the size of PointType, we can't infer the actual size
of the array by dividing the size of the variable by the number of
elements.
We want to make sure clang and llvm-gcc handle this correctly.
llvm-svn: 113730
union, or class that contained an enumeration type. When I was creating
the clang enumeration decl, I wasn't calling "EnumDecl::setIntegerType (QualType)"
which means that if the enum decl was ever asked to figure out it's bit width
(getTypeInfo()) it would crash. We didn't run into this with enum types that
weren't inside classes because the DWARF already told us how big the type was
and when we printed an enum we would never need to calculate the size, we
would use the pre-cached byte size we got from the DWARF. When the enum was
in a struct/union/class and we tried to layout the struct, the layout code
would attempt to get the type info and segfault.
llvm-svn: 113729
Fixed an issue where LLDB would fail to set a breakpoint by
file and line if the DWARF line table has multiple file entries
in the support files for a source file.
llvm-svn: 113721
we cached remapping information using the old nlist index to the
new symbol index, yet we tried to lookup the symbol stubs that
were for symbols that had been remapped by ID instead of using
the new symbol index. This is now fixed and the mach-o symbol tables
are fixed.
Use the delta between two vector entries to determine the stride
in case any padding is inserted by compilers for bsearch calls
on symbol tables when finding symbols by their original ID.
llvm-svn: 113719
They will now be represented as:
eSymbolTypeFunction: eSymbolTypeCode with IsDebug() == true
eSymbolTypeGlobal: eSymbolTypeData with IsDebug() == true and IsExternal() == true
eSymbolTypeStatic: eSymbolTypeData with IsDebug() == true and IsExternal() == false
This simplifies the logic when dealing with symbols and allows for symbols
to be coalesced into a single symbol most of the time.
Enabled the minimal symbol table for mach-o again after working out all the
kinks. We now get nice concise symbol tables and debugging with DWARF in the
.o files with a debug map in the binary works well again. There were issues
where the SymbolFileDWARFDebugMap symbol file parser was using symbol IDs and
symbol indexes interchangeably. Now that all those issues are resolved
debugging is working nicely.
llvm-svn: 113678
to specify a one-liner (either scripting or lldb command) inline.
Refactored CommandObjectBreakpointCommandAdd::Execute() a little bit and added
some comments.
Sn now, we use:
breakpoint command add -p 1 -o "conditional_break.stop_if_called_from_a()"
to specify a Python one-liner as the callback for breakpoint #1.
llvm-svn: 113672
SBValue to access it. For now this is just the result of ObjC NSPrintForDebugger,
but could be extended. Also store the results of the ObjC Object Printer in a
Stream, not a ConstString.
llvm-svn: 113660
pointed out by Jim Ingham. The convenient one-liner specification should only
apply when there is only one breakpoint id being specified for the time being.
llvm-svn: 113609
up a seciton offset address (SBAddress) within a module that returns a
symbol context (SBSymbolContext). Also added a SBSymbolContextList in
preparation for adding find/lookup APIs that can return multiple results.
Added a lookup example code that shows how to do address lookups.
llvm-svn: 113599
command for a breakpoint, for example:
(lldb) breakpoint command add -p 1 "conditional_break.stop_if_called_from_a()"
The ScriptInterpreter interface has an extra method:
/// Set a one-liner as the callback for the breakpoint command.
virtual void
SetBreakpointCommandCallback (CommandInterpreter &interpreter,
BreakpointOptions *bp_options,
const char *oneliner);
to accomplish the above.
Also added a test case to demonstrate lldb's use of breakpoint callback command
to stop at function c() only when its immediate caller is function a(). The
following session shows the user entering the following commands:
1) command source .lldb (set up executable, breakpoint, and breakpoint command)
2) run (the callback mechanism will skip two breakpoints where c()'s immeidate caller is not a())
3) bt (to see that indeed c()'s immediate caller is a())
4) c (to continue and finish the program)
test/conditional_break $ ../../build/Debug/lldb
(lldb) command source .lldb
Executing commands in '.lldb'.
(lldb) file a.out
Current executable set to 'a.out' (x86_64).
(lldb) breakpoint set -n c
Breakpoint created: 1: name = 'c', locations = 1
(lldb) script import sys, os
(lldb) script sys.path.append(os.path.join(os.getcwd(), os.pardir))
(lldb) script import conditional_break
(lldb) breakpoint command add -p 1 "conditional_break.stop_if_called_from_a()"
(lldb) run
run
Launching '/Volumes/data/lldb/svn/trunk/test/conditional_break/a.out' (x86_64)
(lldb) Checking call frames...
Stack trace for thread id=0x2e03 name=None queue=com.apple.main-thread:
frame #0: a.out`c at main.c:39
frame #1: a.out`b at main.c:34
frame #2: a.out`a at main.c:25
frame #3: a.out`main at main.c:44
frame #4: a.out`start
c called from b
Continuing...
Checking call frames...
Stack trace for thread id=0x2e03 name=None queue=com.apple.main-thread:
frame #0: a.out`c at main.c:39
frame #1: a.out`b at main.c:34
frame #2: a.out`main at main.c:47
frame #3: a.out`start
c called from b
Continuing...
Checking call frames...
Stack trace for thread id=0x2e03 name=None queue=com.apple.main-thread:
frame #0: a.out`c at main.c:39
frame #1: a.out`a at main.c:27
frame #2: a.out`main at main.c:50
frame #3: a.out`start
c called from a
Stopped at c() with immediate caller as a().
a(1) returns 4
b(2) returns 5
Process 20420 Stopped
* thread #1: tid = 0x2e03, 0x0000000100000de8 a.out`c + 7 at main.c:39, stop reason = breakpoint 1.1, queue = com.apple.main-thread
36
37 int c(int val)
38 {
39 -> return val + 3;
40 }
41
42 int main (int argc, char const *argv[])
(lldb) bt
bt
thread #1: tid = 0x2e03, stop reason = breakpoint 1.1, queue = com.apple.main-thread
frame #0: 0x0000000100000de8 a.out`c + 7 at main.c:39
frame #1: 0x0000000100000dbc a.out`a + 44 at main.c:27
frame #2: 0x0000000100000e4b a.out`main + 91 at main.c:50
frame #3: 0x0000000100000d88 a.out`start + 52
(lldb) c
c
Resuming process 20420
Process 20420 Exited
a(3) returns 6
(lldb)
llvm-svn: 113596
The Unwind and RegisterContext subclasses still need
to be finished; none of this code is used by lldb at
this point (unless you call into it by hand).
The ObjectFile class now has an UnwindTable object.
The UnwindTable object has a series of FuncUnwinders
objects (Function Unwinders) -- one for each function
in that ObjectFile we've backtraced through during this
debug session.
The FuncUnwinders object has a few different UnwindPlans.
UnwindPlans are a generic way of describing how to find
the canonical address of a given function's stack frame
(the CFA idea from DWARF/eh_frame) and how to restore the
caller frame's register values, if they have been saved
by this function.
UnwindPlans are created from different sources. One source is the
eh_frame exception handling information generated by the compiler
for unwinding an exception throw. Another source is an assembly
language inspection class (UnwindAssemblyProfiler, uses the Plugin
architecture) which looks at the instructions in the funciton
prologue and describes the stack movements/register saves that are
done.
Two additional types of UnwindPlans that are worth noting are
the "fast" stack UnwindPlan which is useful for making a first
pass over a thread's stack, determining how many stack frames there
are and retrieving the pc and CFA values for each frame (enough
to create StackFrameIDs). Only a minimal set of registers is
recovered during a fast stack walk.
The final UnwindPlan is an architectural default unwind plan.
These are provided by the ArchDefaultUnwindPlan class (which uses
the plugin architecture). When no symbol/function address range can
be found for a given pc value -- when we have no eh_frame information
and when we don't have a start address so we can't examine the assembly
language instrucitons -- we have to make a best guess about how to
unwind. That's when we use the architectural default UnwindPlan.
On x86_64, this would be to assume that rbp is used as a stack pointer
and we can use that to find the caller's frame pointer and pc value.
It's a last-ditch best guess about how to unwind out of a frame.
There are heuristics about when to use one UnwindPlan versues the other --
this will all happen in the still-begin-written UnwindLLDB subclass of
Unwind which runs the UnwindPlans.
llvm-svn: 113581
cleaning up the output of many GetDescription objects that are part of a
symbol context. This fixes an issue where no ranges were being printed out
for functions, blocks and symbols.
llvm-svn: 113571
to be set up the way they are. Comment out code that removes pending
settings for live instances (after the settings are copied over).
llvm-svn: 113519
are always printed immediately after the command, before optional
options; also so that in the detailed descriptions of each command
option, the options and their help are output in alphabetical order
(sorted by the short option) rather in whatever order they happened to
be in the table.
llvm-svn: 113496
Sean Callanan