================
@@ -149,6 +149,63 @@ class TextTokenRetokenizer {
addToken();
}
+ /// Check if this line starts with @par or \par
+ bool startsWithParCommand() {
+ unsigned Offset = 1;
+
+ /// Skip all whitespace characters at the beginning.
+ /// This needs to backtrack because Pos has already advanced past the
+ /// actual \par or @par command by the time this function is called.
+ while (isWhitespace(*(Pos.BufferPtr - Offset)))
+ Offset++;
----------------
AaronBallman wrote:
Okay, I think what's got me confused is that you calculate `Offset` but never
actually use it (or modify `Pos.BufferPtr`). I was thinking you would have been
able to use `ltrim().starts_with(...)` but I see now that the issue is we're at
the start of the comment, beyond seeing `par`, and we want to backtrack.
But the logic below seems to be that we already know we're parsing a par
command, so why do we need to do this check in the first place? (e.g., we get
into this code path via:
```
else if (Info->IsParCommand)
S.actOnBlockCommandArgs(BC,
parseParCommandArgs(Retokenizer, Info->NumArgs));
```
and eventually land here, so don't we already know this is a par command?)
https://github.com/llvm/llvm-project/pull/91100
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
