Author: gclayton
Date: Mon Oct 19 19:14:20 2015
New Revision: 250768
URL: http://llvm.org/viewvc/llvm-project?rev=250768&view=rev
Log:
Make sure we restore the process events so they aren't hijacked when using the
async attach when attaching to a process by name and with waitfor.
Modified:
Author: gclayton
Date: Mon Oct 19 19:17:39 2015
New Revision: 250770
URL: http://llvm.org/viewvc/llvm-project?rev=250770&view=rev
Log:
When calling FileSpec::AppendPathComponent() we don't need to include "." in
the path if m_filename is set to exactly '.'. Previously this would cause a
FileSpec
Author: gclayton
Date: Mon Oct 19 19:22:50 2015
New Revision: 250772
URL: http://llvm.org/viewvc/llvm-project?rev=250772&view=rev
Log:
Fixed PlatformDarwin to locate Xcode by using HostInfo::GetProgramFileSpec().
This function returns the FileSpec to the program that is running the
LLDB.framewor
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Minor nit over ordering of #include, but other than that it looks good.
Comment at:
source/Plugins/Process/gdb-remote/GDBRemoteCommunicationServerPlatform.cpp:36
@@ -31,1
clayborg added a comment.
So you are saying to have a setting that allows this plug-in to be enabled and
have it set to false by default? That sounds good to me. It probably should be
false on all platforms and anyone wanting JIT support should enable it?
http://reviews.llvm.org/D12890
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
Is there a list of error codes and what they mean for the GDB remote protocol?
debugserver returns random errors and doesn't abide by any specific error
codes. Quote from some GD
clayborg added inline comments.
Comment at:
lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp:2087-2088
@@ +2086,4 @@
+results.emplace_back(TaskPool::AddTask(parser_fn, cu_idx));
+for (auto& f : results)
+f.wait();
+
S
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good to me. Wait for Enrico to OK as well.
http://reviews.llvm.org/D13836
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
htt
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good. Going forward there is no real need to review any new commands that
only exist in the renderscript source files or plug-ins. If you change any code
outside of the the renderscr
clayborg added a comment.
See inlined comments.
Comment at:
lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp:2087-2088
@@ +2086,4 @@
+results.emplace_back(TaskPool::AddTask(parser_fn, cu_idx));
+for (auto& f : results)
+f.wait();
+
clayborg accepted this revision.
clayborg added a comment.
Accepted as long as Enrico is OK with this.
http://reviews.llvm.org/D13778
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commi
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks fine. Might want to make "h" const in places noted by inlined comments,
but other than that it looks good.
Comment at: source/Core/ConstString.cpp:39
@@ -38,2 +38,3
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
"thread_local" isn't well supported and has performance issues. See inlined
comments.
Comment at: include/lldb/Core/Timer.h:89
@@ +88,3 @@
+
+static thread_
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
This is really the fault of AppleObjCRuntimeV1::CreateInstance() and
AppleObjCRuntimeV2::CreateInstance() not checking the triple to begin with, but
this is a simple solution that will eff
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good.
http://reviews.llvm.org/D13942
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
clayborg added a comment.
Yes, please use any llvm system utilities when possible. And scratch the need
for LLDB_THREAD_LOCAL at any point in the future and just use llvm::ThreadLocal
all the time as thread local variables have already been abstracted by llvm.
http://reviews.llvm.org/D13940
Author: gclayton
Date: Wed Oct 21 16:55:16 2015
New Revision: 250952
URL: http://llvm.org/viewvc/llvm-project?rev=250952&view=rev
Log:
Fix typo so that "./dotest.py --results-formatter=curses_results.Curses
--results-file=/dev/stdout" works again.
Modified:
lldb/trunk/test/lldbcurses.py
Mo
clayborg accepted this revision.
clayborg added a comment.
Looks good. At some point we should add these to the PluginMananger and have
new instances found by having them register URL prefixes which match up with
CreateInstance methods that are given the URL.
http://reviews.llvm.org/D13970
clayborg accepted this revision.
clayborg added a comment.
Looks good. I would venture to say just checkin any clang tidy things that
aren't changing code functionality without need for review.
Repository:
rL LLVM
http://reviews.llvm.org/D13602
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good.
http://reviews.llvm.org/D13940
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
clayborg added a comment.
I would venture to say we should optimize for C++ since those libraries tend to
be larger, but I will leave the decision to you.
http://reviews.llvm.org/D13662
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http
clayborg added a comment.
In http://reviews.llvm.org/D13970#273267, @ovyalov wrote:
> In http://reviews.llvm.org/D13970#273217, @clayborg wrote:
>
> > Looks good. At some point we should add these to the PluginMananger and
> > have new instances found by having them register URL prefixes which m
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good as long as we always want to send the broadcast even if the continue
packet fails to send.
Comment at:
source/Plugins/Process/gdb-remote/GDBRemoteCommunicatio
Author: gclayton
Date: Mon Oct 26 12:16:50 2015
New Revision: 251310
URL: http://llvm.org/viewvc/llvm-project?rev=251310&view=rev
Log:
Re-use prologue parsing code that was already written instead of having two
copies of code that parse line table prologues.
Also since we always read in the DWAR
Author: gclayton
Date: Mon Oct 26 12:52:16 2015
New Revision: 251315
URL: http://llvm.org/viewvc/llvm-project?rev=251315&view=rev
Log:
Fixed the test suite on MacOSX so that
"test/api/multithreaded/TestMultithreaded.py" works without errors.
The problem was that the @skipIfNoSBHeaders on darwin
clayborg added a comment.
The plug-in interface I was thinking of was the Connection class. This would
allow different platforms to support different Connection subclasses for things
like network, serial, USB, Firewire, shared memory, IPC and others. Not sure if
socket is stand alone enough to
clayborg added a comment.
In http://reviews.llvm.org/D14085#275507, @zturner wrote:
> I think most of the time they are used in a Connection class, but I don't
> think it's necessarily guaranteed they will always be that way.
>
> I can think of at least one use case in the future where we will n
clayborg resigned from this revision.
clayborg edited reviewers, added: spyffe, jingham; removed: clayborg.
clayborg added a comment.
Lets let Sean Callanan and Jim Ingham make sure this is good as they are the
expression parser masters.
Repository:
rL LLVM
http://reviews.llvm.org/D14111
clayborg resigned from this revision.
clayborg removed a reviewer: clayborg.
clayborg added a comment.
I don't know anything about shared library loading on Linux.
http://reviews.llvm.org/D14118
___
lldb-commits mailing list
lldb-commits@lists.llvm.
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
Need to make a copy of socket_scheme in GDBRemoteCommunicationServerPlatform
(save it to a std::string).
Comment at:
source/Plugins/Process/gdb-remote/GDBRemot
Author: gclayton
Date: Wed Oct 28 13:04:38 2015
New Revision: 251537
URL: http://llvm.org/viewvc/llvm-project?rev=251537&view=rev
Log:
Make core files not crash when you load a core file into LLDB with just "lldb
-c core".
To do this I added a few new ways to determine the OS from PT_NOTE notes
Author: gclayton
Date: Wed Oct 28 15:49:34 2015
New Revision: 251555
URL: http://llvm.org/viewvc/llvm-project?rev=251555&view=rev
Log:
Don't crash when opening a fuzzed mach-o file that has bad dyld trie data.
Modified:
lldb/trunk/source/Plugins/ObjectFile/Mach-O/ObjectFileMachO.cpp
Modif
Author: gclayton
Date: Wed Oct 28 18:26:59 2015
New Revision: 251579
URL: http://llvm.org/viewvc/llvm-project?rev=251579&view=rev
Log:
Make sure we don't over specify an architecture when we connect to KDP and use
the CPU type and subtype to fill out an architecture. We do this by letting the
ve
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Much better.
Repository:
rL LLVM
http://reviews.llvm.org/D13859
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llv
clayborg resigned from this revision.
clayborg removed a reviewer: clayborg.
clayborg added a comment.
I will defer to Jason Molenda.
http://reviews.llvm.org/D14226
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-
clayborg resigned from this revision.
clayborg edited reviewers, added: tfiala; removed: clayborg.
clayborg added a comment.
I don't build cmake either, I am probably not a great person to ask. Todd Fiala
might be a good person.
http://reviews.llvm.org/D13819
Author: gclayton
Date: Tue Nov 3 16:40:07 2015
New Revision: 252001
URL: http://llvm.org/viewvc/llvm-project?rev=252001&view=rev
Log:
With the new modules debugging, we have seen cases where clang is not emitting
full definitions for types that are member variables of classes. If we try to
make
Author: gclayton
Date: Tue Nov 3 16:42:04 2015
New Revision: 252005
URL: http://llvm.org/viewvc/llvm-project?rev=252005&view=rev
Log:
Fix a deadlock when connecting to a remote GDB server that might not support
all packets that lldb-server or debugserver supports. The issue was the
m_last_stop_
Author: gclayton
Date: Tue Nov 3 17:23:59 2015
New Revision: 252013
URL: http://llvm.org/viewvc/llvm-project?rev=252013&view=rev
Log:
All instance variables start with "m_". Fix "options" to be "m_options".
Modified:
lldb/trunk/include/lldb/DataFormatters/ValueObjectPrinter.h
lldb/trunk
Author: gclayton
Date: Tue Nov 3 17:23:22 2015
New Revision: 252012
URL: http://llvm.org/viewvc/llvm-project?rev=252012&view=rev
Log:
Fix __fp16 types so we can display them and use them in expressions.
I am not adding a test case for this since I don't know how portable the __fp16
type is betw
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
You should populate the m_address_class_map in ObjectFileELF::ParseSymbols() as
this is where the m_address_class_map is already being populated. For ARM
symbols, just check the
clayborg added a comment.
See Inlined comment.
Comment at: source/Plugins/ObjectFile/ELF/ObjectFileELF.cpp:975-997
@@ -974,5 +974,25 @@
{
-// No entry in the address class map before the address. Return
-// default address class for an address in a code sect
Author: gclayton
Date: Tue Nov 10 11:47:04 2015
New Revision: 252622
URL: http://llvm.org/viewvc/llvm-project?rev=252622&view=rev
Log:
Fixed TypeMemberFunctionImpl to not use clang types directly but use the new
CompilerDecl class to do the job in an abstract way.
Fixed a crash that would happen
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
Symbols are the #1 most expensive item memory wise in LLDB right now. We remove
many symbols in Mach-O and we set the m_uid of each symbol to the original
symbol table index. The
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good!
http://reviews.llvm.org/D14549
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
This is not the right fix. Compilers that emit addresses with DW_FORM_data*
forms are saying that this address is an offset from the DW_AT_low_pc so this
function will not return
clayborg requested changes to this revision.
clayborg added a comment.
I agree with tberghammer.
Repository:
rL LLVM
http://reviews.llvm.org/D14633
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
clayborg requested changes to this revision.
This revision now requires changes to proceed.
Comment at:
source/Plugins/Process/Linux/NativeRegisterContextLinux_mips64.cpp:1378
@@ -1377,2 +1377,3 @@
GPR_linux_mips regs;
+lldb_private::ArchSpec arch;
::memset(®s, 0, s
clayborg added a comment.
So yes this is a compiler bug where the compiler is producing invalid DWARF.
Repository:
rL LLVM
http://reviews.llvm.org/D14631
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mail
clayborg added a comment.
GDB remote protocol specifies that register values are sent in target byte
order. We shouldn't change this. A big endian system should not send things as
little endian. That being said, the current register context assumes you have a
buffer that can contain all registe
clayborg added a comment.
I would prefix the first fix:
> Add a new function to llvm::APInt what returns the pointer to the actual data
> in case of both endian (possibly the best option)
Using llvm::APInt gets us unlimited integer size abilities (support 128 bit
ints and higher) along with a
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
See inlined comments.
Comment at: include/lldb/API/SBProcess.h:344-346
@@ -343,1 +343,5 @@
+// Save the state of the process in a core file (or mini dump o
clayborg added a comment.
Code correction and added that we should make sure the process is alive and
stopped.
Comment at: source/API/SBProcess.cpp:1433-1437
@@ +1432,7 @@
+{
+ProcessSP process_sp(GetSP());
+lldb::SBFileSpec core_file(file_name);
+lldb::SBError erro
clayborg added a comment.
This patch was accepted on November 13th. So yes this is ready to land. Not
sure what you are waiting for?
Repository:
rL LLVM
http://reviews.llvm.org/D14530
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
htt
clayborg added a comment.
Yes, I agree Zach, the caller of SaveCore should verify the process is in a
good state first. The only downside of that is if the windows mini-dumper can
actually make a core file on a running process, we might want to allow this and
make actual SaveCore implementation
clayborg resigned from this revision.
clayborg edited reviewers, added: jingham, spyffe; removed: clayborg.
clayborg added a comment.
Resigning as I will let Jim Ingham and Sean Callanan review.
http://reviews.llvm.org/D13350
___
lldb-commits mailin
clayborg added a comment.
OK, then callers of SaveCore should make sure the process is in a good state
before calling the API.
http://reviews.llvm.org/D14793
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/ma
Author: gclayton
Date: Thu Nov 19 17:04:09 2015
New Revision: 253616
URL: http://llvm.org/viewvc/llvm-project?rev=253616&view=rev
Log:
Fixed Xcode build by adding the PlatformNetBSD code.
Modified:
lldb/trunk/lldb.xcodeproj/project.pbxproj
Modified: lldb/trunk/lldb.xcodeproj/project.pbxproj
Author: gclayton
Date: Thu Nov 19 17:10:45 2015
New Revision: 253618
URL: http://llvm.org/viewvc/llvm-project?rev=253618&view=rev
Log:
Fix a crasher in SymbolContext::SortTypeList() where something that was
iterating over a std::multimap was actually mutating the list.
Modified:
lldb/trun
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
Fix the caps on the error message, remove the new files and inline the code and
this will be ready.
Comment at: source/API/SBProcess.cpp:1445
@@ +1444,3 @@
+
clayborg added a comment.
Ok, so fix the caps and rename the file to be WindowsMiniDump.cpp/.h. I would
rather it not have the "ObjectFilePECOFF" prefix because it would indicated
that it is a new subclass of ObjectFile.
http://reviews.llvm.org/D14793
___
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks fine.
Repository:
rL LLVM
http://reviews.llvm.org/D14876
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good. Thanks for all the fixes.
http://reviews.llvm.org/D14793
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.l
Author: gclayton
Date: Fri Nov 20 17:53:53 2015
New Revision: 253739
URL: http://llvm.org/viewvc/llvm-project?rev=253739&view=rev
Log:
Fix Xcode project after recent Windows core file changes.
Modified:
lldb/trunk/lldb.xcodeproj/project.pbxproj
Modified: lldb/trunk/lldb.xcodeproj/project.pb
clayborg added a comment.
Yep, I took care of that shortly after your commit.
Repository:
rL LLVM
http://reviews.llvm.org/D14793
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
If this is indeed a very similar test, we should fix the original test to do
the right thing.
Repository:
rL LLVM
http://reviews.llvm.org/D14944
__
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
One minor nit, but looks good.
Comment at: source/Target/Process.cpp:4622
@@ +4621,3 @@
+
+if (interrupt_requested) {
+if (StateIsStoppedState
clayborg resigned from this revision.
clayborg removed a reviewer: clayborg.
clayborg added a comment.
Jason Molenda should be sufficient to review this.
http://reviews.llvm.org/D15046
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http:/
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
I am not sure that I like this new "platform process connect" command as it
isn't really clear what this command does. "platform connect" connects to the
remote platform, "proces
clayborg resigned from this revision.
clayborg removed a reviewer: clayborg.
clayborg added a comment.
I can't intelligently comment on low level ptrace stuff, so I will defer to
others that can.
http://reviews.llvm.org/D15042
___
lldb-commits mail
clayborg added a comment.
I am fine with adding a new platform GDB remote packet that queries for a list
of processes that we need to connect to after we attach. It might be nice for
clients to be able to vend N number of processes that should be attached to
when the platform is connected to. T
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks fine.
http://reviews.llvm.org/D11465
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/li
Author: gclayton
Date: Tue Dec 1 18:43:32 2015
New Revision: 254476
URL: http://llvm.org/viewvc/llvm-project?rev=254476&view=rev
Log:
Added support for -gmodule debugging when debug info is left in the .o files on
Darwin.
This is done by finding the types that are forward declarations that come
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
"process load" should probably be changed to have an options that allows us to
specify where the shared library needs to be installed:
(lldb) process load --install-path=/usr/l
clayborg added a comment.
The --install-path option is optional and doesn't need to be specified, but if
it is specified, we need all platforms, even the host platform, to install the
shared library to this location prior to loading it.
http://reviews.llvm.org/D15152
___
ng not
> conflicting with a type name in the std namespace then it works fine but it
> should work even in case of conflicting types as they are leave in different
> namespaces and it also worked before your change.
>
> Currently I don't fully understand your change but please l
t; not conflicting with a type name in the std namespace then it works fine
> > but it should work even in case of conflicting types as they are leave in
> > different namespaces and it also worked before your change.
> >
> > Currently I don't fully understand your change
.
Greg
> On Dec 2, 2015, at 11:26 AM, Greg Clayton via lldb-commits
> wrote:
>
> So I was able to repro this with the following code:
>
> #include
> #include
>
> namespace a
> {
>struct S
>{
>int s;
>int ss;
>};
clayborg added a comment.
Maybe Platform::LoadImage() should keep its current arguments and the path that
is specified should always be correct for the platform. Other code should do
the copy over if it needs to and call Platform::LoadImage() with a path that
makes sense for the platform. The w
clayborg added a comment.
Everything looks good, just wondering why we need the template code? See
inlined comments.
Comment at: include/lldb/Utility/JSON.h:102-104
@@ +101,5 @@
+
+template {} &&
+ std::is_unsigned{}>::type* =
n
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Ok, thanks for the explanation. You might add a comment saying something like
you just said above the template stuff so people can see why it is needed.
http://reviews.llvm.org/D15187
clayborg added a comment.
I like that solution.
http://reviews.llvm.org/D15152
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
clayborg added a comment.
Be sure to add nice HeaderDoc to the LoadImage() declaration in Platform.h so
everyone knows what is expected. We still might want to split this up so that
only Platform has LoadImage and we make everyone else just implement:
virtual uint32_t DoLoadImage(lldb_private
clayborg added a comment.
PlatformPOSIX::LoadImage() and PlatformAndroid::LoadImage() should now become
PlatformPOSIX::DoLoadImage() and PlatformAndroid::DoLoadImage() and the common
code between them removed, add Platform::LoadImage() and move the common code
that looks at both FileSpec argume
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
So my main issue with this is the new virtual
"Platform::GetPendingGdbServerList(...)" function. Can we remove this function
and just put the functionality into Platform::Contect
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
Add header doc for the new LoadImage and make Platform::LoadImage virtual in
case platforms want to override this and this is good to go.
Comment at: include/ll
Author: gclayton
Date: Fri Dec 4 12:37:48 2015
New Revision: 254743
URL: http://llvm.org/viewvc/llvm-project?rev=254743&view=rev
Log:
Fill in the generic register kind if in AugmentRegisterInfoViaABI if it is
available.
Modified:
lldb/trunk/source/Plugins/Process/gdb-remote/ProcessGDBRemote
clayborg accepted this revision.
clayborg added a comment.
This revision is now accepted and ready to land.
Looks good, just check me inline comment and make changes if you think you need
to.
Comment at: source/Commands/CommandObjectPlatform.cpp:414-419
@@ -413,1 +413,8 @@
+
+
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
So one of two things needs to happen here:
- ClangASTContext::DeclContextCountDeclLevels() becomes a function that is on
ClangASTContext only and the opaque arguments get changed
Author: gclayton
Date: Mon Dec 7 19:02:08 2015
New Revision: 254980
URL: http://llvm.org/viewvc/llvm-project?rev=254980&view=rev
Log:
Trying to submit 254476 one more time. This implement -gmodule debugging
support.
It was previously reverted due to issues that showed up only on linux. I was
a
clayborg added a comment.
So one thing I don't want to propagate here is the "go parse everything inside
a decl context" if we can avoid it. I believe the first CompilerDeclContext
patch did this, but I don't remember the exact details so I could be wrong. But
going forward I would like to see
clayborg added a comment.
That does make sense. Lets ignore the partial parsing until we find a
performance problem that needs to be fixed.
By I really do want to see the API on CompilerDeclContext and CompilerDecl get
a lot more member functions that can do useful things now that we have this
clayborg added a comment.
In http://reviews.llvm.org/D15312#304652, @dawn wrote:
> Thanks Greg! To address your main point:
>
> > So either make it generic, or clang specific.
>
>
> DeclContextCountDeclLevels's interface follows DeclContextFindDeclByName
> (authored by Paul Herman). If DeclCon
Author: gclayton
Date: Tue Dec 8 12:39:50 2015
New Revision: 255038
URL: http://llvm.org/viewvc/llvm-project?rev=255038&view=rev
Log:
Change DeclContextFindDeclByName to return a vector of CompilerDecl objects.
Opaque pointers should only be used for the decl context object. Also made a
default
clayborg added a comment.
BTW: I fixed DeclContextFindDeclByName to return a vector of CompilerDecl
objects:
% svn commit
Sendinginclude/lldb/Symbol/ClangASTContext.h
Sendinginclude/lldb/Symbol/GoASTContext.h
Sendinginclude/lldb/Symbol/TypeSystem.h
Sendingsource/
clayborg added a comment.
In http://reviews.llvm.org/D15312#305392, @dawn wrote:
> > Greg: But going forward I would like to see more of "find a struct named 'X'
>
>
> in CompilerDeclContext 'Y'" queries, instead of ...
>
> I think lldb should have both, a "search for anything named foo in my sco
clayborg accepted this revision.
clayborg added a comment.
Just fix the comment as mentioned in the inline comments and this is ready. Jim
should OK this as well.
Comment at: include/lldb/Target/Language.h:176
@@ -175,1 +175,3 @@
+// return the primary language, so if
clayborg added a comment.
In http://reviews.llvm.org/D15312#306201, @dawn wrote:
> Hi Greg, I'm working on a new revision to change the patch as you suggest
> (thanks for your review - you had some great suggestions!).
>
> Sorry, if I'm missing something obviously here, but there's still some th
clayborg added a comment.
Above I meant to say I don't understand what the "else if (find_name)" does. I
understand the rest of it. So in the "else if (find_name)" we are looking for
any decl, regardless of wether it matches "opaque_find_decl_ctx" as long as the
name and type are correct? Again
clayborg added a subscriber: clayborg.
clayborg added a comment.
I like the clang specific patch here better. Mainly because if you add a
function like this to TypeSystem.h:
virtual uint32_t
DeclContextCountDeclLevels (void *opaque_decl_ctx,
void *opaque_find_de
301 - 400 of 2985 matches
Mail list logo