Author: paulherman
Date: Thu Sep 17 14:32:02 2015
New Revision: 247923
URL: http://llvm.org/viewvc/llvm-project?rev=247923&view=rev
Log:
Fix caching for clang::Decl in DWARFASTParserClang
Reviewers: sivachandra, chaoren, clayborg, tberghammer
Subscribers: lldb-commits
Differential Revision: htt
paulherman updated this revision to Diff 35026.
paulherman added a comment.
Fix caching for clang::Decl in DWARFASTParserClang
Changed from std::set to switch. Rebased the patch.
http://reviews.llvm.org/D12942
Files:
source/Plugins/SymbolFile/DWARF/DWARFASTParserClang.cpp
Index: source/Plug
paulherman updated this revision to Diff 35019.
paulherman added a comment.
Fix caching for clang::Decl in DWARFASTParserClang
Fixed small mistakes from previous commit.
http://reviews.llvm.org/D12942
Files:
source/Plugins/SymbolFile/DWARF/DWARFASTParserClang.cpp
Index: source/Plugins/Symbo
paulherman updated this revision to Diff 35017.
paulherman added a comment.
Fix caching for clang::Decl in DWARFASTParserClang
http://reviews.llvm.org/D12942
Files:
source/Plugins/SymbolFile/DWARF/DWARFASTParserClang.cpp
Index: source/Plugins/SymbolFile/DWARF/DWARFASTParserClang.cpp
paulherman created this revision.
paulherman added reviewers: sivachandra, chaoren, tberghammer, clayborg.
paulherman added a subscriber: lldb-commits.
http://reviews.llvm.org/D12942
Files:
source/Plugins/SymbolFile/DWARF/DWARFASTParserClang.cpp
Index: source/Plugins/SymbolFile/DWARF/DWARFASTP
Author: paulherman
Date: Wed Sep 16 13:48:30 2015
New Revision: 247836
URL: http://llvm.org/viewvc/llvm-project?rev=247836&view=rev
Log:
Add using directives to the clang::DeclContext and fix decls for variables
inside namespaces
Summary: Supports the parsing of the "using namespace XXX" and "us
paulherman updated this revision to Diff 34911.
paulherman added a comment.
Add using directives to the clang::DeclContext and fix decls for variables
inside namespaces
Rebased the patch.
http://reviews.llvm.org/D12897
Files:
include/lldb/Symbol/SymbolFile.h
source/Expression/ClangExpress
paulherman added a comment.
The need for ParseDeclsInContext is to get a list of namespaces that are
contained within a using-directive. This does not actually parse anything but
DW_TAG_imported_namespace and DW_TAG_imported_decl. The actual contents of the
namespace are parsed only when the se
paulherman added a comment.
Please also look at http://reviews.llvm.org/D12897
It fixes the imported decls and some bugs that I have missed (evaluating
Fun::fun_var reports an error in TestCppNsImport). Unfortunately I don't think
there is a way to report ambiguity without for "imported" withou
paulherman created this revision.
paulherman added reviewers: sivachandra, chaoren, clayborg.
paulherman added a subscriber: lldb-commits.
Supports the parsing of the "using namespace XXX" and "using XXX::XXX"
directives. Added ambiguity errors when it two decls with the same name are
encountere
paulherman added a comment.
The test was supposed to be marked as an XFAIL. I'm currently writing a fix for
this that reports ambiguity in a context and deals with imported decls.
http://reviews.llvm.org/D12658
___
lldb-commits mailing list
lldb-co
Author: paulherman
Date: Tue Sep 15 19:29:10 2015
New Revision: 247760
URL: http://llvm.org/viewvc/llvm-project?rev=247760&view=rev
Log:
Fix double include in CompilerDeclContext.h
Modified:
lldb/trunk/include/lldb/Symbol/CompilerDeclContext.h
Modified: lldb/trunk/include/lldb/Symbol/Compile
Author: paulherman
Date: Tue Sep 15 19:26:12 2015
New Revision: 247759
URL: http://llvm.org/viewvc/llvm-project?rev=247759&view=rev
Log:
Fix missing include in CompilerDeclContext.h
Modified:
lldb/trunk/include/lldb/Symbol/CompilerDeclContext.h
Modified: lldb/trunk/include/lldb/Symbol/Compil
Author: paulherman
Date: Tue Sep 15 18:44:17 2015
New Revision: 247746
URL: http://llvm.org/viewvc/llvm-project?rev=247746&view=rev
Log:
Search variables based on clang::DeclContext and clang::Decl tree
Summary: SymbolFileDWARF now creates VarDecl and BlockDecl and adds them to the
Decl tree. Th
paulherman updated this revision to Diff 34850.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
Rebased the patch.
http://reviews.llvm.org/D12658
Files:
include/lldb/Symbol/ClangASTContext.h
include/lldb/Symbol/CompilerDecl.h
include/lldb/Sym
paulherman updated this revision to Diff 34823.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
This adds handling of imported declarations, but there is no call to actually
process them as I am not sure where to do this. I believe that the right pla
paulherman updated this revision to Diff 34808.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
http://reviews.llvm.org/D12658
Files:
include/lldb/Symbol/ClangASTContext.h
include/lldb/Symbol/CompilerDecl.h
include/lldb/Symbol/CompilerDeclCont
paulherman updated this revision to Diff 34768.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
Moved the DeclLinkToObject inside Variable::GetDecl.
http://reviews.llvm.org/D12658
Files:
include/lldb/Symbol/ClangASTContext.h
include/lldb/Symbol
paulherman added a comment.
I just ran a grep through the source. It seems that everywhere it is created as
a shared_ptr. In a previous attempt I tried storing it in the TypeSystem map
from decl to object as Variable* instead of VariableSP and it seemed like at
some point between launching a qu
paulherman added a comment.
In http://reviews.llvm.org/D12658#245851, @clayborg wrote:
> So I don't like it in ParseVariableDIE() because it means we must create the
> decl right away for the variable when parsing it. ParseVariableDIE is used to
> parse all variables everywhere and we don't nee
paulherman added a comment.
I thought about the problem a bit more and I believe that having the call to
DeclLinkDeclToObject inside ParseVariableDIE makes sense since this gets called
only when the variable will actually be used in a search together with its
context. Also, since ParseVariableD
paulherman added a comment.
If I move the call to DeclLinkDeclToObject inside any of the GetDeclXXX
methods, then there is no way to access the VariableSP needed for the call and
it is probably harder to reconstruct the parameters needed for ParseVariableDIE
starting from just the DIE itself.
paulherman added a comment.
Are you suggesting that the DeclLinkToObject is moved inside the GetDecl call
and I only remove that? This is the only place where GetDecl is called, so it
would be equivalent from a computational point of view.
http://reviews.llvm.org/D12658
paulherman added a comment.
I think the only concern left might be the call to DeclLinkToObject in
ParseVariableDIE. It doesn't have anything to do with clang but I feel like
there should be a way to do this lazily.
Otherwise, I believe that the comments have been addressed.
Thank you!
http:
paulherman added a comment.
In http://reviews.llvm.org/D12658#245466, @clayborg wrote:
> In http://reviews.llvm.org/D12658#244710, @paulherman wrote:
>
> > [WIP] Search variables based on clang::DeclContext and clang::Decl tree
> >
> > This revision fixes some of the comments. There are some thin
paulherman updated this revision to Diff 34711.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
I think this fixes the comments. It also adds a cache for clang::decl so that
there is no memory wasted.
http://reviews.llvm.org/D12658
Files:
includ
paulherman updated this revision to Diff 34616.
paulherman added a comment.
Search variables based on clang::DeclContext and clang::Decl tree
I believe I fixed most of the comments. I removed the handling of using
declarations for now and made the test an expected failure because of this. I
thi
paulherman updated this revision to Diff 34609.
paulherman added a comment.
[WIP] Search variables based on clang::DeclContext and clang::Decl tree
This revision fixes some of the comments. There are some things I'm not sure
about. The problem is that at some point there will be the need to link
paulherman added a comment.
I believe that the approach of CompilerDeclContext::FindDecls could be better.
But then this kind of forces CompilerDeclContext to inherit CompilerDecl (a
function is both a DeclContext and a Decl) and I believe that creating a class
for each possible entity is an ov
paulherman added a comment.
I agree with most of your comments and will fix them.
One thing I noticed is that Block::GetDeclContext() does not do the right thing
(see inline comment).
About creating a CompilerDecl class, I agree with you. I used
CompilerDeclContext since it was just a void* an
paulherman added a comment.
Ping.
http://reviews.llvm.org/D12658
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
paulherman updated this revision to Diff 33345.
paulherman added a comment.
Add scope tree for variable searching
http://reviews.llvm.org/D12304
Files:
include/lldb/Symbol/Block.h
include/lldb/Symbol/CompileUnit.h
include/lldb/Symbol/Function.h
include/lldb/Symbol/Symbol.h
include/lld
Author: paulherman
Date: Wed Aug 19 17:44:48 2015
New Revision: 245515
URL: http://llvm.org/viewvc/llvm-project?rev=245515&view=rev
Log:
Fix tests to work on remote targets.
Modified:
lldb/trunk/test/lang/cpp/chained-calls/TestCppChainedCalls.py
lldb/trunk/test/lang/cpp/global_operators/T
Author: paulherman
Date: Wed Aug 19 16:44:56 2015
New Revision: 245508
URL: http://llvm.org/viewvc/llvm-project?rev=245508&view=rev
Log:
Fix evaluation of global operators in C++
Added:
lldb/trunk/test/lang/cpp/global_operators/
lldb/trunk/test/lang/cpp/global_operators/Makefile
lldb/
Author: paulherman
Date: Wed Aug 19 16:23:01 2015
New Revision: 245505
URL: http://llvm.org/viewvc/llvm-project?rev=245505&view=rev
Log:
Improve tests regarding imported namespaces and chained calls in C++
Modified:
lldb/trunk/test/lang/cpp/chained-calls/TestCppChainedCalls.py
lldb/trunk/
paulherman updated this revision to Diff 32593.
paulherman added a comment.
Improve tests regarding imported namespaces and chained calls in C++
http://reviews.llvm.org/D12165
Files:
test/lang/cpp/chained-calls/TestCppChainedCalls.py
test/lang/cpp/chained-calls/main.cpp
test/lang/cpp/nsim
paulherman created this revision.
paulherman added a reviewer: sivachandra.
paulherman added a subscriber: lldb-commits.
http://reviews.llvm.org/D12165
Files:
test/lang/cpp/chained-calls/TestCppChainedCalls.py
test/lang/cpp/chained-calls/main.cpp
test/lang/cpp/nsimport/TestCppNsImport.py
paulherman added a comment.
Ping
http://reviews.llvm.org/D11543
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
Author: paulherman
Date: Tue Aug 18 17:46:57 2015
New Revision: 245381
URL: http://llvm.org/viewvc/llvm-project?rev=245381&view=rev
Log:
Fix resolution conflict between global and class static variables in C++
Added:
lldb/trunk/test/lang/cpp/scope/
lldb/trunk/test/lang/cpp/scope/Makefile
Author: paulherman
Date: Tue Aug 18 17:43:37 2015
New Revision: 245380
URL: http://llvm.org/viewvc/llvm-project?rev=245380&view=rev
Log:
Test chaned function calls and imported namespaces in C++
Added:
lldb/trunk/test/lang/cpp/chained-calls/
lldb/trunk/test/lang/cpp/chained-calls/Makefile
paulherman updated this revision to Diff 32345.
paulherman added a comment.
Fix resolution conflict between global and class static variables in C++
http://reviews.llvm.org/D12044
Files:
include/lldb/Symbol/Variable.h
include/lldb/Symbol/VariableList.h
source/Plugins/SymbolFile/DWARF/Symb
paulherman created this revision.
paulherman added reviewers: clayborg, sivachandra, spyffe.
paulherman added a subscriber: lldb-commits.
Consider having a global variable 'a' and a static variable with the same name
inside a class. This resolves the arbitrary choice when resolving the name 'a'.
paulherman added a comment.
This patch does exactly that. It detects the language of the frame and upgrades
it according to the rules you said (I think I might've missed ObjC -> ObjC++,
but that can be added).
Regarding the global setting of ObjC, I believe it is not helpful. What I mean
is, w
paulherman added a comment.
The idea is that, as I user, I do not expect the identifiers "Class" and "id"
to not be available - I don't think I've seen a warning or notice about that
when evaluating expressions.
I believe that setting the language based on the current frame is a good guess.
I
44 matches
Mail list logo
