commit hash 8abb5fb68f81b0e42d824bf080b1cef9a61559d6
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
clang/include/clang/Tooling/Syntax/Nodes.h | ||
---|---|---|
101–105 | All of this is ephemeral, we are gonna have just one Name Specifier with a PointerUnion. | |
204–258 | (bis) All of this is temporary, we are gonna have just one Name Specifier with a PointerUnion. | |
clang/lib/Tooling/Syntax/BuildTree.cpp | ||
763–764 | We extracted the logic into getUnqualifiedSourceRange | |
843–862 | This is the same logic as DeclRefExpr! Exacly the same code! | |
clang/lib/Tooling/Syntax/Nodes.cpp | ||
119–127 | (bis) All of this is temporary, we are gonna have just one Name Specifier with a PointerUnion. | |
clang/unittests/Tooling/Syntax/TreeTest.cpp | ||
964–970 | I noticed that we don't need that as everything is dependent on the template anyways |
What is this diff based on? On the left I see, for example, NamespaceNameSpecifier, which is not in the repository yet.
clang/lib/Tooling/Syntax/BuildTree.cpp | ||
---|---|---|
746–821 | s/treat/support/ Also, use llvm::report_fatal_error instead? assert is not supposed to ever trigger. |
- Improve getLocalSourceRange
- nested-name-specifier is now a ::-separated list of name-specifiers
clang/include/clang/Tooling/Syntax/Nodes.h | ||
---|---|---|
297 | You can remove this todo now. | |
clang/lib/Tooling/Syntax/BuildTree.cpp | ||
748 | Please use isa<DecltypeType>(T) and inline this expression into its only user. | |
753 | Ditto, please use isa and inline into the only user. | |
803 | ... except in the case of incorrect code. Feel free to just add a TODO. | |
819 | Please use isa<GlobalNameSpecifier>(NS). | |
821 | Could you add an overload for foldNode that takes a NestedNameSpecifierLoc but ignores it, just like we have an overload of foldNode that takes a TypeLoc but ignores it? | |
831 | Please don't use C++17, Clang uses C++14 now. | |
834 | WDYM by "upstream"? | |
clang/unittests/Tooling/Syntax/TreeTest.cpp | ||
879–881 | TS => ST (struct template?) | |
885 | ST |
- Answer code review
- Simpler logic for getUnqualifiedIdSourceRange and inline it
- Remove ambiguously named variable NNS
clang/lib/Tooling/Syntax/BuildTree.cpp | ||
---|---|---|
834 | I meant to put that logic under the DeclRefExpr node, instead of here. But I found a way of writing this logic in a simpler way :). So I just inlined it! |
clang/lib/Tooling/Syntax/BuildTree.cpp | ||
---|---|---|
784 | Newbie mistake. Corrected in latter commit |
rebase to add this commit from a further patch
- [SyntaxTree] Fix crash on name specifier.
All of this is ephemeral, we are gonna have just one Name Specifier with a PointerUnion.