owenpan added inline comments.
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:123-125
+ bool braceTokenMatchesScope(FormatToken &Token) {
+ if (!Token.is(tok::l_brace) || Scopes.empty())
+ return false;
----------------
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:129
+ case TT_LambdaLBrace:
+ return Scopes.back() == ST_Function;
+ case TT_ClassLBrace:
----------------
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:868
+ assert(!Scopes.empty());
+ assert(braceTokenMatchesScope(OpeningBrace));
+ Scopes.pop_back();
----------------
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:1169-1181
+ switch (Tok->getType()) {
+ case TT_FunctionLBrace:
+ case TT_LambdaLBrace:
+ Scopes.push_back(ST_Function);
+ break;
+ case TT_ClassLBrace:
+ case TT_StructLBrace:
----------------
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:2491
+ auto IsChainedOperatorAmpOrMember = [](const FormatToken *token) {
+ return token && token->isOneOf(tok::amp, tok::period, tok::arrow,
+ tok::arrowstar, tok::periodstar);
----------------
Sorry about that!
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:2498-2499
+ if (Tok.is(tok::amp) && PrevToken && PrevToken->Tok.isAnyIdentifier() &&
+ (!PrevToken->getPreviousNonComment() ||
+ IsChainedOperatorAmpOrMember(PrevToken->getPreviousNonComment())) &&
+ NextToken && NextToken->Tok.isAnyIdentifier()) {
----------------
Now we only need the lambda.
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:1195-1198
+ // Handle unbalanced braces.
+ if (!Scopes.empty())
+ Scopes.pop_back();
// Lines can start with '}'.
----------------
dkt01 wrote:
> owenpan wrote:
> > I don't think it's about unbalanced braces here.
> `if (!Scopes.empty())` handles unbalanced braces. `if(Tok->Previous)`
> handles the case where a line starts with an rbrace.
I can't think of an example. Do you have one?
================
Comment at: clang/lib/Format/TokenAnnotator.cpp:2481-2482
+ if (Tok.is(tok::amp) && (PrevToken && PrevToken->Tok.isAnyIdentifier()) &&
+ (!PrevToken->getPreviousNonComment() ||
+ IsChainedOperatorAmpOrMember(PrevToken->getPreviousNonComment())) &&
+ (NextToken && NextToken->Tok.isAnyIdentifier()) &&
----------------
dkt01 wrote:
> owenpan wrote:
> > The lambda would check that `PrevToken` is nonnull.
> `!PrevToken->getPreviousNonComment() ||` is a different check than the null
> check in the lambda. It's acceptable to have no token two tokens prior
> because that indicates the ampersand is the second token of the line.
I actually wanted fold this into the lambda. (See above.)
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D141959/new/
https://reviews.llvm.org/D141959
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
