cor3ntin wrote:

I'd be oppose to try to invent what would effectively be yet another string 
semantics for this.
However, there maybe a spec bug here
Presumably the intent of the feature is to allow header-names and so maybe the 
syntax should be that of and header-name?

The two options are
 - status quo, these things are strings and invalid escape sequences should be 
diagnose - because `\` would be interpreted as escape sequences
 - these things are header names and `\` is always just a `\`. 

I mean, consider `#line 1 "C:\test"` - that does not do what you think it does 
https://godbolt.org/z/EhKnere7v

@hubert-reinterpretcast 

https://github.com/llvm/llvm-project/pull/201413
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to