Re: lexer.ll: Warn about non-UTF-8 characters (issue 5505090)

[dak]

On 2012/01/01 22:06:52, Keith wrote:
> On 2012/01/01 10:12:27, dak wrote:

> Sorry, I wasn't making much sense.  As a reader I want to *recognize*
what the
> but switch/case is doing rather than trying to figure it out.  Maybe :
>   // Test if these bytes are a UTF-8 encoding of a Unicode character,
>   // and warn if not.   Trap overly-long UTF-8 encodings, but we don't
>   // need to worry about finer details like some filters do.

What are the finer details?  UTF-16 surrogates?  I was just too lazy to
figure out the patterns for them.

> because your test is almost but not quite equivalent to the regex in
the back of
> the Flex manual.

I can't find a regex there.  Got a link or section name?  Is this a
released version?

http://codereview.appspot.com/5505090/