hintonda created this revision.
Quash several warnings of this type using clang's suggested fix:
[2741/3631] Building CXX object
tools/lldb/source/Plugins/Process/gdb-remote/CMakeFiles/lldbPluginProcessGDBRemote.dir/GDBRemoteCommunication.cpp.o
/Users/dhinton/projects/llvm_project/llvm/tools/lld
hintonda created this revision.
These two methods are essentially assignents, but don't check
for self-assignment and use memcpy for member variables.
Since they aren't actually operator=(), it's unclear what should be done:
- check and shortcut
- assert
- allow but s/memcpy/memmove/
https://r
hintonda created this revision.
Add type to FileSpec::PathSyntax enum to decrease size for
FileSpec on systems with 32 bit pointers.
Thanks to @zturner for pointing this out.
https://reviews.llvm.org/D39574
Files:
include/lldb/Utility/FileSpec.h
Index: include/lldb/Utility/FileSpec.h
=
This revision was automatically updated to reflect the committed changes.
Closed by commit rL317277: Remove getCategories mechanism of specifying test
categories (authored by labath).
Repository:
rL LLVM
https://reviews.llvm.org/D39515
Files:
lldb/trunk/packages/Python/lldbsuite/test/exampl
Author: labath
Date: Thu Nov 2 15:13:09 2017
New Revision: 317277
URL: http://llvm.org/viewvc/llvm-project?rev=317277&view=rev
Log:
Remove getCategories mechanism of specifying test categories
Summary:
This mechanism was mostly redundant with the file-based .categories
mechanism, and it was inte
Author: labath
Date: Thu Nov 2 15:12:55 2017
New Revision: 317276
URL: http://llvm.org/viewvc/llvm-project?rev=317276&view=rev
Log:
Remove android watchpoint xfails
Now that the wathpoint tests have their own category, we can easily skip
them on devices which don't have watchpoint support. There
Author: labath
Date: Thu Nov 2 14:35:26 2017
New Revision: 317270
URL: http://llvm.org/viewvc/llvm-project?rev=317270&view=rev
Log:
Fix some warnings found by ToT clang
These fall into two categories:
- unused variables
- (uint8_t *)NULL + X -- changed to reinterpret_cast(X)
Modified:
lldb/
zturner added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto
hintonda updated this revision to Diff 121350.
hintonda added a comment.
I realize this probably isn't an acceptable change, but it was a good
learing experience for me, as I'd never looked at the lldb code before
last Sunday, so I figured I'd go ahead and complete it.
While the basic technique r
labath added a comment.
In https://reviews.llvm.org/D38829#914305, @lawrence_danna wrote:
> @labath
>
> > Couldn't we just change the File::Read/Write functions to call these
> > directly
>
> Like I said to @zturner , this is possible, but it can't work with the
> existing APIs in SBDebugger.h.
lawrence_danna added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+
labath added a comment.
In https://reviews.llvm.org/D38829#914300, @zturner wrote:
> Ok, I wasn't aware of the libedit problem.
>
> I still don't like this approach, because it causes the design of the `File`
> class to be influenced by a limitation of a 3rd party library.
>
> What about adding
clayborg added a comment.
In https://reviews.llvm.org/D38829#914300, @zturner wrote:
> Ok, I wasn't aware of the libedit problem.
>
> I still don't like this approach, because it causes the design of the `File`
> class to be influenced by a limitation of a 3rd party library.
>
> What about addin
lawrence_danna added a comment.
@labath
> (Please add lldb-commits to future reviews, so that people are aware of
> what's going on.)
ok
> Couldn't we just change the File::Read/Write functions to call these directly
Like I said to @zturner , this is possible, but it can't work with the exist
zturner added a comment.
Ok, I wasn't aware of the libedit problem.
I still don't like this approach, because it causes the design of the `File`
class to be influenced by a limitation of a 3rd party library.
What about adding a method to `IOObject` called `FILE *GetFileStream()`. Then
the met
clayborg added a comment.
Another option would be to add a URL version of these functions:
class SBDebugger {
void SetInputURL(const char *url);
void SetOutputURL(const char *url);
void SetErrorURL(const char *url);
};
Then we allow this to trickle down to the IOBase in that way.
labath added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto p
zturner added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto
zturner added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto
labath added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto p
jingham accepted this revision.
jingham added a comment.
Very nice, thanks!
https://reviews.llvm.org/D39515
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
lawrence_danna added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+
zturner added inline comments.
Comment at:
lldb/trunk/source/Plugins/ScriptInterpreter/Python/PythonDataObjects.cpp:1075-1096
+static int readfn(void *ctx, char *buffer, int n)
+{
+ auto state = PyGILState_Ensure();
+ auto *file = (PyObject *) ctx;
+ int result = -1;
+ auto
This revision was automatically updated to reflect the committed changes.
Closed by commit rL317218: Use LLVM version string (authored by sas).
Repository:
rL LLVM
https://reviews.llvm.org/D39429
Files:
lldb/trunk/source/lldb.cpp
Index: lldb/trunk/source/lldb.cpp
==
Author: sas
Date: Thu Nov 2 09:56:52 2017
New Revision: 317219
URL: http://llvm.org/viewvc/llvm-project?rev=317219&view=rev
Log:
Run clang-format on lldb.cpp
Modified:
lldb/trunk/source/lldb.cpp
Modified: lldb/trunk/source/lldb.cpp
URL:
http://llvm.org/viewvc/llvm-project/lldb/trunk/source
Author: sas
Date: Thu Nov 2 09:56:19 2017
New Revision: 317218
URL: http://llvm.org/viewvc/llvm-project?rev=317218&view=rev
Log:
Use LLVM version string
Summary:
macOS builds of LLDB use the bundle version from
`tools/driver/lldb-Info.plist`. That file hasn't been updated since the 4.0
release s
labath added inline comments.
Comment at:
source/Plugins/Process/Linux/NativeRegisterContextLinux_ppc64le.cpp:534
+Status NativeRegisterContextLinux_ppc64le::ReadVMX() {
+ void *buf = GetVMXBuffer();
+ if (!buf)
Isn't this somewhat over-engineered?
I mean, Ge
labath updated this revision to Diff 121287.
labath added a comment.
Herald added a subscriber: ki.stfu.
This removes the getCategories function altogether. I've moved the
filesystem-based category computation (which was implemented as a getCategories
base case) into the consumer, so that it is i
labath added a comment.
I agree that we have too many ways of specifying categories. I think that the
file-based and getCategories() way are redundant, particularly, as we tend to
only have one test file per directory anyway. For very fine grained
annotations, we can use the `@add_test_categori
labath added subscribers: lldb-commits, labath.
labath reopened this revision.
labath added a comment.
This revision is now accepted and ready to land.
(Please add lldb-commits to future reviews, so that people are aware of what's
going on.)
So, the reason why this failed to compile is that andr
30 matches
Mail list logo
