Author: ki.stfu
Date: Mon Mar 23 07:44:54 2015
New Revision: 232946
URL: http://llvm.org/viewvc/llvm-project?rev=232946view=rev
Log:
Fix ObjCDataFormatterTestCase.test_nsdate_* tests
The time/date strings (like 6pm April 10, 1985) are interpreted as a local
time but CFDateGetAbsoluteTime()
Hi,
what is the exact case this patch is trying to fix? Could you give
steps to reproduce? Or even better, a test case?
cheers,
pl
http://reviews.llvm.org/D8541
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
abandoning as this is fixed already
http://reviews.llvm.org/D8482
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
In http://reviews.llvm.org/D8541#145060, @labath wrote:
Hi,
what is the exact case this patch is trying to fix? Could you give
steps to reproduce? Or even better, a test case?
cheers,
pl
Sure. See it in TEST PLAN section.
http://reviews.llvm.org/D8541
EMAIL PREFERENCES
Hi jingham, clayborg,
This patch fixes Target::Launch in case of synchronous execution.
http://reviews.llvm.org/D8541
Files:
source/Target/Target.cpp
Index: source/Target/Target.cpp
===
--- source/Target/Target.cpp
+++
Could you upload the diffs with full context please? (i.e., `git diff -U99`
or `svn diff --diff-cmd=diff -x -U99`)
http://reviews.llvm.org/D8538
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits
wow, that's quite a test plan. :)
Sorry about that, I didn't see anything in the email, and I did not
know that phab does not send out the test plans O:)
pl
http://reviews.llvm.org/D8541
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Hi rengolin, vharron, tberghammer,
This patch adds support aarch64-linux-gnu (SysV) abi in lldb. Code is similar
to what has been implemented for MacOS and I have gone through abi
specification this code will work without any changes though I have not seen
any improvements in lldb-server
Hi rengolin, vharron, tberghammer,
This patch is first in a series of upcoming patches that add support for
arm-linux-lldb.
This patch will be usable once once support for Register Context will be added
later this week.
http://reviews.llvm.org/D8539
Files:
cmake/LLDBDependencies.cmake
In http://reviews.llvm.org/D8444#145131, @abidh wrote:
MI part looks ok to me. Please wait for Greg to approve the rest.
Thanks for review!
@clayborg, take a look please.
http://reviews.llvm.org/D8444
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Author: ki.stfu
Date: Mon Mar 23 09:22:27 2015
New Revision: 232951
URL: http://llvm.org/viewvc/llvm-project?rev=232951view=rev
Log:
Fix ObjCDataFormatterTestCase.test_nsdate_* tests (ver. 2)
This commit reverts r232946 because it caused an another error with absolute
time.
Modified:
Looks good with a few minor suggestions
Comment at: source/Plugins/ABI/SysV-arm/ABISysV_arm.cpp:224
@@ +223,3 @@
+reg_value.SetUInt32(*ai);
+if
(!reg_ctx-WriteRegister(reg_ctx-GetRegisterInfoByName(reg_names[i]),
reg_value))
+return false;
The comment says 'aarch64-linux-gnu' ABI support -- this should be AArch64
ABI or AArch64 SysV ABI, no?
You need to handle aarch64 in the other case in `CreateRegisterContextForFrame`
as well.
Comment at: source/Plugins/Process/elf-core/ThreadElfCore.cpp:129-131
@@ -127,2
In http://reviews.llvm.org/D8541#145077, @labath wrote:
Sorry about that, I didn't see anything in the email, and I did not
know that phab does not send out the test plans O:)
It's a surprise for me too :)
http://reviews.llvm.org/D8541
EMAIL PREFERENCES
I added a few inline comments but overall it looks good. I think you haven't
seen any improvement in the test results because this class isn't used by any
of the tests (not sure who else use this class except ThreadElfCore).
P.S.: Next time please upload the diff with full context to make it
Author: ki.stfu
Date: Mon Mar 23 09:00:54 2015
New Revision: 232950
URL: http://llvm.org/viewvc/llvm-project?rev=232950view=rev
Log:
Convert CRLF to LF in CODE_OWNERS.txt
Modified:
lldb/trunk/CODE_OWNERS.txt
Modified: lldb/trunk/CODE_OWNERS.txt
URL:
Hi ovyalov, clayborg,
Fetch module specification from remote process also
Previously the remote module specification was fetched only from the
remote platform. With this CL if we have a remote process then we ask it
if it have any information from a given module. It is required because
on
Author: tberghammer
Date: Mon Mar 23 10:50:03 2015
New Revision: 232952
URL: http://llvm.org/viewvc/llvm-project?rev=232952view=rev
Log:
Remove virtual and add override to all virtual functions in Process/gdb-remote.
Modified:
Author: ki.stfu
Date: Mon Mar 23 15:46:10 2015
New Revision: 233017
URL: http://llvm.org/viewvc/llvm-project?rev=233017view=rev
Log:
Fix/Add comments in CMICmdCmdGdbSet (MI)
Modified:
lldb/trunk/tools/lldb-mi/MICmdCmdGdbSet.cpp
lldb/trunk/tools/lldb-mi/MICmdCmdGdbSet.h
Modified:
It added a new warning.
/lldb/trunk/source/Plugins/Platform/Android/PlatformAndroidRemoteGDBServer.cpp:41
```
[4/13] Building CXX object
tools/lldb/source/Plugins/Platform/Android/C...iles/lldbPluginPlatformAndroid.dir/PlatformAndroidRemoteGDBServer.cpp.o
Why are we putting logic for a specific language runtime in generic code?
This doesn't seem right to me.
On Mon, Mar 23, 2015 at 11:38 AM Greg Clayton clayb...@gmail.com wrote:
lgtm
http://reviews.llvm.org/D8558
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
I agree with Zachary. That wasn't what the FIXME was about, and moving more
runtime specific code into target is not what is desired. This should go into
the Runtime, and get called by the dynamic loader. I don't think this change
is a change in the right direction.
Jim
On Mar 23, 2015,
How do you get a thread that has NO frame 0? That seems weird, we always have
a register context to make frame 0...
Anyway, if you're going to do this, add it to the thread == NULL checks
earlier on in the function so we can get out of the command with an appropriate
error message. If we
This is a patch that I've been keeping in my branches for a while. I
hit a null pointer dereference there. Looks like Ed had the same
issue. I'll move the check to the beginning of the function as you
suggested though.
On Mon, Mar 23, 2015 at 12:01 PM, Jim Ingham jing...@apple.com wrote:
How do
I encountered an issue like this early on while bringing up a new target
(FreeBSD/mips64 I think), while the work was incomplete. Of course once I
finished we always have frame 0, which is why I didn't commit it at the time.
But it (or, perhaps just an assert) presumably will be handy when
Hi clayborg,
If a struct type S has a member T that has a member that is a function that
returns a typedef of S* the respective field would be duplicated, which caused
an assert down the line in RecordLayoutBuilder. This patch adds a check that
removes the possibility of trying to resolve the
I'd like to add some unit tests for this patch but I don't know how to create
one that has two separate source files (which is required to repro this bug).
Does the unit test framework support that?
http://reviews.llvm.org/D8561
EMAIL PREFERENCES
How do you get a thread that has NO frame 0? That seems weird, we always have
a register context to make frame 0...
Anyway, if you're going to do this, add it to the thread == NULL checks
earlier on in the function so we can get out of the command with an appropriate
error message. If we
Comment at:
test/functionalities/data-formatter/data-formatter-stl/libstdcpp/vector/Makefile:12
@@ +11,3 @@
+ifneq (,$(findstring clang,$(CC)))
+ CFLAGS_EXTRAS += -fstandalone-debug
+endif
I think you'll need to replace this with the equivalent
Comment at:
test/functionalities/data-formatter/data-formatter-stl/libstdcpp/vector/Makefile:9
@@ +8,3 @@
+# clang-3.5+ outputs FullDebugInfo by default for Darwin/FreeBSD
+# # targets. Other targets do not, which causes this test to fail.
+# # This flag enables FullDebugInfo
Its mainly an efficiency thing. If you have a std::string you might need to
allocate data on the heap to store the string just to make the call. Since this
argument is a path, it is more likely to have to allocate on the heap. Most
std::string implementations store the first 20 some bytes in
See inline comments.
You can easily make a test case with two files, there are many that do it.
Comment at: source/Symbol/ClangASTImporter.cpp:289-296
@@ -288,2 +288,10 @@
{
-return CompleteTagDecl(tag_type-getDecl());
+auto tag_decl = tag_type-getDecl();
+
Author: ovyalov
Date: Mon Mar 23 16:03:02 2015
New Revision: 233021
URL: http://llvm.org/viewvc/llvm-project?rev=233021view=rev
Log:
Move ADB communications to AdbClient class - to make it accessible by other
components.
http://reviews.llvm.org/D8535
Added:
Hi abidh, clayborg,
This path adds -gdb-show command with 1 option: target-async.
Also it adds tests for -gdb-set and -gdb-show commands.
All tests pass on OS X.
http://reviews.llvm.org/D8566
Files:
test/tools/lldb-mi/TestMiGdbSetShow.py
tools/lldb-mi/CMakeLists.txt
Author: ki.stfu
Date: Mon Mar 23 16:16:25 2015
New Revision: 233022
URL: http://llvm.org/viewvc/llvm-project?rev=233022view=rev
Log:
Fix Target::Launch in case of synchronous execution
Summary: This patch fixes Target::Launch in case of synchronous execution.
Test Plan:
# Create file with
No, don't change any existing API. Unfortunately by the time we caught this
one, it was out in the wild and we had programs linking against it. Just make
sure to not add const to any new API.
http://reviews.llvm.org/D8444
EMAIL PREFERENCES
Author: ki.stfu
Date: Mon Mar 23 15:49:51 2015
New Revision: 233018
URL: http://llvm.org/viewvc/llvm-project?rev=233018view=rev
Log:
Add target-async option in -gdb-set command (MI)
Modified:
lldb/trunk/tools/lldb-mi/MICmdCmdGdbSet.cpp
lldb/trunk/tools/lldb-mi/MICmdCmdGdbSet.h
In http://reviews.llvm.org/D8444#145654, @clayborg wrote:
No, don't change any existing API. Unfortunately by the time we caught this
one, it was out in the wild and we had programs linking against it. Just make
sure to not add const to any new API.
ok.
http://reviews.llvm.org/D8444
Comment at: include/lldb/API/SBCommandInterpreter.h:27-67
@@ -26,43 +26,43 @@
bool
GetStopOnContinue () const;
void
SetStopOnContinue (bool);
bool
GetStopOnError () const;
void
SetStopOnError (bool);
bool
GetStopOnCrash
Please abandon it as http://reviews.llvm.org/D8547 will fix this issue
http://reviews.llvm.org/D8338
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
Stéphane,
please add me as a reviewer for this patch. I have an account, spyffe.
Sean
On Mar 23, 2015, at 12:48 PM, Stephane Sezer s...@cd80.net wrote:
I'd like to add some unit tests for this patch but I don't know how to create
one that has two separate source files (which is required
Stéphane,
please add me as a reviewer for this patch. I have an account, spyffe.
Sean
http://reviews.llvm.org/D8561
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
Add GetPromptOnQuit/SetPromptOnQuit to SBCommandInterpreter.i
http://reviews.llvm.org/D8444
Files:
include/lldb/API/SBCommandInterpreter.h
include/lldb/Interpreter/CommandInterpreter.h
scripts/Python/interface/SBCommandInterpreter.i
source/API/SBCommandInterpreter.cpp
Author: ki.stfu
Date: Mon Mar 23 16:50:21 2015
New Revision: 233029
URL: http://llvm.org/viewvc/llvm-project?rev=233029view=rev
Log:
Fix missing interfaces to scripts/Python/interface after r229110, r231858 and
r232891
This includes:
# SBCommandInterpreter::EventIsCommandInterpreterEvent
#
Comment at: include/lldb/API/SBCommandInterpreter.h:223
@@ +222,3 @@
+bool
+GetPromptOnQuit() const;
+
clayborg wrote:
You can actually take the const off of this as it does nothing when your only
member is an std::shared_ptr, std::auto_ptr or just a
Remove const as per inline comments.
Comment at: include/lldb/API/SBCommandInterpreter.h:223
@@ +222,3 @@
+bool
+GetPromptOnQuit() const;
+
You can actually take the const off of this as it does nothing when your only
member is an std::shared_ptr,
@clayborg, take a look again please.
Comment at: include/lldb/API/SBCommandInterpreter.h:219-223
@@ -218,1 +218,7 @@
+bool
+GetPromptOnQuit() const;
+
+void
+SetPromptOnQuit(bool b);
+
clayborg wrote:
You also need to add these to the
Looks good.
http://reviews.llvm.org/D8541
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
Add the new API to the .i files as mentioned in the inline comments.
Comment at: include/lldb/API/SBCommandInterpreter.h:219-223
@@ -218,1 +218,7 @@
+bool
+GetPromptOnQuit() const;
+
+void
+SetPromptOnQuit(bool b);
+
You also need to add these
Author: sas
Date: Mon Mar 23 12:05:41 2015
New Revision: 232966
URL: http://llvm.org/viewvc/llvm-project?rev=232966view=rev
Log:
Do not assert on POSIXDYLD double-eAdd.
Summary:
This has been discovered while experimenting with the gecko linker on android.
In general, assert()'ing on user input
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8495
Files:
lldb/trunk/source/Plugins/DynamicLoader/POSIX-DYLD/DYLDRendezvous.cpp
Index: lldb/trunk/source/Plugins/DynamicLoader/POSIX-DYLD/DYLDRendezvous.cpp
===
---
Hi clayborg,
http://reviews.llvm.org/D8554
Files:
source/Commands/CommandObjectThread.cpp
Index: source/Commands/CommandObjectThread.cpp
===
--- source/Commands/CommandObjectThread.cpp
+++ source/Commands/CommandObjectThread.cpp
We really do not want this change. This means anytime we play with a class it
will complete it right away, regardless of needing a full definition or not.
This fix should be done in clang so that clang naturally knows how to say I
need a full definition and I need to know to complete a class
LGTM, I think I had this same change in a local branch at one point.
http://reviews.llvm.org/D8554
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
Hi clayborg,
To run tests against a different target platform many extra compiler flags are
needed to specify sysroot, include dirs, etc. The environment variable
CFLAGS_EXTRAS seems suited for this purpose except that several Makefiles
clobber the current flags. This change modifies all of
Fix the std::string arguments as noted in inline comments and this is good to
go.
Comment at: include/lldb/Host/common/NativeProcessProtocol.h:291
@@ +290,3 @@
+virtual lldb_private::FileSpec
+GetLoadedModuleFileSpec(const std::string module_path) = 0;
+
Hi clayborg,
Saw this while reading some code in DynamicLoader classes. Looks like this has
been a FIXME since 2011 at least.
http://reviews.llvm.org/D8558
Files:
source/Plugins/DynamicLoader/MacOSX-DYLD/DynamicLoaderMacOSXDYLD.cpp
source/Target/Target.cpp
Index:
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8554
Files:
lldb/trunk/source/Commands/CommandObjectThread.cpp
Index: lldb/trunk/source/Commands/CommandObjectThread.cpp
===
--- lldb/trunk/source/Commands/CommandObjectThread.cpp
+++
Comment at: source/Plugins/Platform/Android/AdbClient.cpp:71-75
@@ +70,7 @@
+response.split (devices, \n);
+if (devices.empty ())
+{
+error.SetErrorString (Wrong number of devices returned from ADB);
+return error;
+}
+
tberghammer
Author: sas
Date: Mon Mar 23 13:36:54 2015
New Revision: 232983
URL: http://llvm.org/viewvc/llvm-project?rev=232983view=rev
Log:
Initialize ObjC runtime at the right location.
Summary:
Saw this while reading some code in DynamicLoader classes. Looks like this has
been a FIXME since 2011 at
Author: emaste
Date: Mon Mar 23 12:52:10 2015
New Revision: 232969
URL: http://llvm.org/viewvc/llvm-project?rev=232969view=rev
Log:
Fix whitespace in finishSwigWrapperClasses.py
- replace hard tabs with 4-space indents
- delete EOL whitespace
Modified:
Author: emaste
Date: Mon Mar 23 12:52:38 2015
New Revision: 232970
URL: http://llvm.org/viewvc/llvm-project?rev=232970view=rev
Log:
Use .so library extension by default if platform is not Windows or Darwin
Modified:
lldb/trunk/scripts/Python/finishSwigPythonLLDB.py
Modified:
Author: emaste
Date: Mon Mar 23 12:57:42 2015
New Revision: 232973
URL: http://llvm.org/viewvc/llvm-project?rev=232973view=rev
Log:
Remove non-standard Environment header section
Modified:
lldb/trunk/scripts/Python/buildSwigPython.py
lldb/trunk/scripts/Python/finishSwigPythonLLDB.py
Author: zturner
Date: Mon Mar 23 13:00:07 2015
New Revision: 232974
URL: http://llvm.org/viewvc/llvm-project?rev=232974view=rev
Log:
Rely on Makefile.rules in lldb/test/driver/batch_mode.
This was hardcoding some make rules instead of relying on
Makefile.rules, which was causing some of the
lgtm
http://reviews.llvm.org/D8363
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
lgtm
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8559
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
lgtm
http://reviews.llvm.org/D8558
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
Author: sas
Date: Mon Mar 23 13:15:07 2015
New Revision: 232979
URL: http://llvm.org/viewvc/llvm-project?rev=232979view=rev
Log:
Add a missing null pointer check in CommandObjectThread.cpp.
Reviewers: clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D8554
Fixed accordingly to suggestions.
Please take another look.
Thank you.
http://reviews.llvm.org/D8535
Files:
source/Plugins/Platform/Android/AdbClient.cpp
source/Plugins/Platform/Android/AdbClient.h
source/Plugins/Platform/Android/CMakeLists.txt
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8558
Files:
lldb/trunk/source/Plugins/DynamicLoader/MacOSX-DYLD/DynamicLoaderMacOSXDYLD.cpp
lldb/trunk/source/Target/Target.cpp
Index: lldb/trunk/source/Target/Target.cpp
===
---
I don't see how this does multiple connections? Seems like the do/while loop
would need to be in a thread?
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8452
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
Comment at: source/Target/Target.cpp:2605-2606
@@ -2604,4 +2610,1 @@
{
-if (!synchronous_execution)
-m_process_sp-RestoreProcessEvents ();
-
Can we use a mixed code (for sync and async execution) to
What's the status on this CL?
http://reviews.llvm.org/D6910
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
Looks good.
http://reviews.llvm.org/D8444
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
Author: ki.stfu
Date: Mon Mar 23 17:45:13 2015
New Revision: 233034
URL: http://llvm.org/viewvc/llvm-project?rev=233034view=rev
Log:
Turn off 'quit' confirmation in lldb-mi
Summary:
# Turn off interpreter.prompt-on-quit on startup (MI)
# Add CommandInterpreter::SetPromptOnQuit
# Add
Hi clayborg,
This commit adds this alternate route only when parsing variable dies
corresponding to global or static variables. The motivation for this is that GCC
does not emit linkage names for functions and variables declared/defined in
anonymous namespaces. Having this alternate route fixes
Just fix the language compare so we aren't looking for a string c++ and use
the enum and this should be good.
Comment at: source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp:7407
@@ +7406,3 @@
+{
+std::string
Comment at: source/Host/linux/HostInfoLinux.cpp:225
@@ -224,3 +224,3 @@
bool
-HostInfoLinux::ComputeSharedLibraryDirectory(FileSpec file_spec)
{
Do we need to keep this method around since
HostInfoPosix::ComputeSupportExeDirectory uses
Should be fixed in http://reviews.llvm.org/rL233041
USERS
ovyalov (Author)
ki.stfu (Auditor)
http://reviews.llvm.org/rL233021
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
Do we want the have one thread that listens for incoming connections, and once
a connection is made, make a new GDBRemoteCommunicationServerPlatform shared
pointer and launch a new thread that would communicate over the newly bound
port? Then we could have multiple connections to the same
Comment at: source/Target/Target.cpp:2605-2606
@@ -2604,4 +2610,1 @@
{
-if (!synchronous_execution)
-m_process_sp-RestoreProcessEvents ();
-
zturner wrote:
ki.stfu wrote:
Can we use a mixed code (for
Hi Greg, What do you think about this? supporting multiple connections to
lldb-platform will allow us to trivially run the test suite in parallel which
is going to be incredibly important to us testing against 6 remote Linux
architectures and 6 remote Android architectures.
F436486: archs.png
Not worth changing the patch but fwiw you could have added eFlagRequiresFrame
to the CommandObjectParsed flags for this command to do the same thing. My
main reaction echoes Jim's - I don't know what it means to have a thread and no
frame 0. But given that two lldb devs have hit it, I guess
Yes, I think there are a lot of places around in lldb where we assume that if
you have a valid thread you have a valid frame 0. You're not going to get very
far if you don't get that right.
Jim
REPOSITORY
rL LLVM
http://reviews.llvm.org/D8554
EMAIL PREFERENCES
Good change as long as it accomplishes the same thing as your null check
against GetStackFrameAtIndex(0). eFlagRequiresFrame is slightly different,
checking that the ExecutionContext got a frame but I suppose the ECX got it
from the same place -- GetStackFrameAtIndex(0) so it should be fine.
lgtm
http://reviews.llvm.org/D8570
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
lldb-commits mailing list
lldb-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
86 matches
Mail list logo
