r188863 - Issue fixits replacing invalid character literals with the equivalent \xNN

[Stephan Bergmann]

On 08/21/2013 06:18 PM, Richard Smith wrote:
> On 20 Aug 2013 21:15, "Nick Lewycky" <nicholas at mxc.ca
> <mailto:nicholas at mxc.ca>> wrote:
>  >
>  > Author: nicholas
>  > Date: Tue Aug 20 23:10:58 2013
>  > New Revision: 188863
>  >
>  > URL: http://llvm.org/viewvc/llvm-project?rev=188863&view=rev
>  > Log:
>  > Issue fixits replacing invalid character literals with the equivalent
> \xNN
>  > escape code.
>  >
>  > Added:
>  >     cfe/trunk/test/Lexer/char-literal-encoding-fixit.c
>  > Modified:
>  >     cfe/trunk/lib/Lex/LiteralSupport.cpp
>  >
>  > Modified: cfe/trunk/lib/Lex/LiteralSupport.cpp
>  > URL:
> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Lex/LiteralSupport.cpp?rev=188863&r1=188862&r2=188863&view=diff
>  >
> ==============================================================================
>  > --- cfe/trunk/test/Lexer/char-literal-encoding-fixit.c (added)
>  > +++ cfe/trunk/test/Lexer/char-literal-encoding-fixit.c Tue Aug 20
> 23:10:58 2013
>  > @@ -0,0 +1,11 @@
>  > +// RUN: cp %s %t
>  > +// RUN: %clang_cc1 -fixit -x c %t
>  > +// RUN: FileCheck -input-file=%t %t
>  > +
>  > +// Note that this file is not valid UTF-8.
>  > +
>  > +int test1 = 'ˆ';
>  > +// CHECK: int test1 = '\x88';
>  > +
>  > +int test2 = 'abˆc';
>  > +// CHECK: int test2 = 'ab\x88c';

But the latter will be parsed as "a", "b", "\x88c" rather than "a", "b", 
"\x88", "c".  I think the only safe way to do such a replacement esp. 
for character literals is to use octal \ooo escapes.  (For string 
literals, there would also be the possibility of splitting into adjacent 
string literals.)

Stephan