Summary:
The main difference here is that in the WINLOG macros you can specify
log categories per call, whereas here you have to go the usual lldb
route of getting a Log* variable first. While this means you have to
write at least two statements, it usually means that each statement will
fit on a single line, whereas fitting the WINLOG invocation on a single
line was almost impossible. So the total size of code does not increase
even in functions with a single log statement, and functions with more
logging get shorter.
The downside here is reduced flexibility in specifying the log
categories, which a couple of functions used quite heavily (e.g.
RefreshStateAfterStop). For these I chose a single category used most
prominently and put everything into that, although a solution with
multiple log variables is definitely possible.
Reviewers: zturner, amccarth
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D30172
llvm-svn: 295822
Summary:
The code was attempting to copy the shared pointer member in order to
guarantee atomicity, but this is not enough. Instead, protect the
pointer with a proper read-write mutex.
This bug was present here for a long time, but my recent refactors must
have altered the timings slightly, such that now this fails fairly often
when running the tests: the test runner runs the "log disable" command
just as the thread monitoring the lldb-server child is about to report
that the server has exited.
I add a test case for this. It's not possible to reproduce the race
deterministically in normal circumstances, but I have verified that
before the fix, the test failed when run under tsan, and was running
fine afterwards.
Reviewers: clayborg, zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D30168
llvm-svn: 295712
Summary:
I originally set out to move the NameMatches closer to the relevant
function and add some unit tests. However, in the process I've found a
couple of bugs in the implementation:
- the early exits where not always correct:
- (test==pattern) does not mean the match will always suceed because
of regular expressions
- pattern.empty() does not mean the match will fail because the "" is
a valid prefix of any string
So I cleaned up those and added some tests. The only tricky part here
was that regcomp() implementation on darwin did not recognise the empty
string as a regular expression and returned an REG_EMPTY error instead.
The simples fix here seemed to be to replace the empty expression with
an equivalent non-empty one.
Reviewers: clayborg, zturner
Subscribers: mgorny, lldb-commits
Differential Revision: https://reviews.llvm.org/D30094
llvm-svn: 295651
Use both LLDB- and LLVM-specific tool/library directories when LLDB is
being built stand-alone. This ensures that the freshly-built tools
(and libraries) are used correctly.
Without this patch, the test suite uses LLVM_TOOLS_DIR and LLVM_LIBS_DIR
to locate lldb, and set PATH and LD_LIBRARY_PATH. When doing
a stand-alone build, these variables represent the installed LLVM.
As a result, tests either fail due to missing lldb executable
or use an earlier installed LLDB version rather than the one being
built.
To solve this, additional LLDB_TOOLS_DIR and LLDB_LIBS_DIR variables
are added and populated using LLVM_*_OUTPUT_INTDIR. Those variables
contain directories used to output built executables and libraries.
In stand-alone builds, they represent the build-tree directories
used by LLDB. In integrated builds, they have the same values as
LLVM_*_DIR and therefore using them does not harm.
The new variables are prepended to PATH and LD_LIBRARY_PATH to ensure
that freshly built binaries are preferred over potentially earlier
installed ones. Furthermore, paths used to locate various tools are
updated to match appropriate locations.
Differential Revision: https://reviews.llvm.org/D29985
llvm-svn: 295621
unittest/Utility to unittest/Target so the unit
tests can be run from xcode again.
The diff is enormous - I think zachary might have
put windows line endings or something with his last
commit? didn't look too closely but his commit & this
commit have every line being different.
llvm-svn: 295530
The testsuite's results formatter maintains a result_status_counts
structure solely for the purpose of setting the return status code
after the testsuite has run. This data is redundant with the
result_events structure that contains the results of individual
tests.
There are subtle bugs arising from this redundancy that make some
builds report no errors but a nonzero status. Rather than try to
make sure these two are always in agreement, I've just rewritten
the code that used to use the counts to now use the per-test
results.
<rdar://problem/30496966>
llvm-svn: 295522
Summary:
There have been a few new values added to a few LLVM enums
this change makes sure that LLDB code handles them correctly.
Reviewers: labath
Reviewed By: labath
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D30005
Author: Eugene Zemtsov <ezemtsov@google.com>
llvm-svn: 295445
The way of injecting an error into the printf call was not working on
darwin - the C library still happily format the character. It only
returns an error after we use a wide character that does not fit into a
single byte, so switch the test to use that.
llvm-svn: 295443
Changes wrt. previous version:
- add #include <atomic>: fix build on windows
- add extra {} around the string literals used to initialize
llvm::StringLiteral: fix gcc build
llvm-svn: 295442
In the case we are stepping over the thread creation instruction, we
will end up calling Thread::SingleStep back-to-back twice (because of
the intermediate PTRACE_EVENT_CLONE stop). This will cause the cpu mask
to be set inappropriately (because the old SingleStepCheck object will
be destroyed after we create the new one), and the single-step will
fail.
Before the refactor the code was still incorrect in this case, but in a
different way (the thread was left with the incorrect mask after the
stepping was complete), so this was not easy to spot.
This fixes TestCreateDuringInstructionStep on the affected devices.
llvm-svn: 295440
While refactoring the code in r293046 I made a very basic error -
relying on destructor side-effects of a copyable object. Fix that and
make the object non-copyable.
This fixes the tests on the platforms that need this workaround, but
unfortunately we don't have a way to make a more platform-agnostic test
right now.
llvm-svn: 295345
TSan now has the ability to report races on "external" object, i.e. any library class/object that has read-shared write-exclusive threading semantics. The detection and reporting work almost out of the box, but TSan can now provide the type of the object (as a string). This patch implements this into LLDB.
Differential Revision: https://reviews.llvm.org/D30024
llvm-svn: 295342
effects was passed as an expression to assert() calls. If lldb is
built without asserts, the expression was eliminated and we lost
the side effects -- these methods stopped working.
<rdar://problem/30342959>
llvm-svn: 295271
Summary:
We currently have two log channel registration mechanisms. One uses a
set of function pointers and the other one is based on the
PluginManager.
The PluginManager dependency is unfortunate, as logging
is also used in lldb-server, and the PluginManager pulls in a lot of
classes which are not used in lldb-server.
Both approach have the problem that they leave too much to do for the
user, and so the individual log channels end up reimplementing command
line argument parsing, category listing, etc.
Here, I replace the PluginManager-based approach with a one. The new API
is more declarative, so the user only needs to specify the list of list
of channels, their descriptions, etc., and all the common tasks like
enabling/disabling categories are hadled by common code. I migrate the
LogChannelDWARF (only user of the PluginManager method) to the new API.
In the follow-up commits I'll replace the other channels with something
similar.
Reviewers: clayborg, zturner, beanz
Subscribers: aprantl, lldb-commits
Differential Revision: https://reviews.llvm.org/D29895
llvm-svn: 295190
Summary:
It turns out listing each library twice is not enough to resolve all
references in a debug build on linux - a number of executables fails to
link with random symbols missing. Increasing the number to three seems
to be enough. The choice of lldbCore to set the multiplicity on is
somewhat arbitrary, but it seems fitting, as it is the biggest layering
transgressor.
Reviewers: beanz
Subscribers: mgorny, lldb-commits
Differential Revision: https://reviews.llvm.org/D29888
llvm-svn: 295189
Summary:
GCC emits also symbols for the __PRETTY_FUNCTION__ virtual variable,
which we accidentaly pick up when looking for functions for with
"unique_function_name" in the name. This makes the target.FindFunctions
call fail, as that symbol is not a function.
I also strenghten the test a bit to make sure we actually find all the
functions we are interested in. I've put a check that we find at least 6
functions, but maybe this should be *exactly* 6 ?
Reviewers: clayborg
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D29932
llvm-svn: 295170
With this patch, the only dependency left is from Utility
to Host. After this is broken, Utility will finally be
standalone.
Differential Revision: https://reviews.llvm.org/D29909
llvm-svn: 295088
Summary:
We've had two ways to print a "debug" log message.
- Log::GetDebug() was testing a Stream flag which was never set.
- Log::Debug() was checking for the presence of "log enable --debug"
flag.
Given that these two were used very rarely and we already have a
different way to specify "I want a more verbose log", I propose to remove
these two functions and migrate the callers to LLDB_LOGV. This commit
does that.
Reviewers: clayborg, zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D29823
llvm-svn: 294939