code review 6625062: cmd/gc: skip over reported BOMs

cmd/gc: skip over reported BOMs

This keeps the BOM runes from causing other errors.

[rsc, 2012-10-07 20:00:52]

Hello golang-dev@googlegroups.com,

I'd like you to review this change to
https://go.googlecode.com/hg/

[remyoudompheng, 2012-10-07 20:04:46]

I just tried that and it doesn't work: in bombad.go at the first BOM, the lexer
has gone in identifier mode for some reason.

It reads the first byte of the BOM, goes into identifier mode, skips the BOM,
ends up with '/', and ocmplains.

[remyoudompheng, 2012-10-07 20:06:18]

Maybe because the BOM is \xef\xbb\xbf and U+00EF is ï, which is indeed a letter.

[rsc, 2012-10-07 20:11:56]

Hello golang-dev@googlegroups.com, remyoudompheng@gmail.com (cc:
golang-dev@googlegroups.com),

Please take another look.

[rsc, 2012-10-07 20:12:40]

On Sun, Oct 7, 2012 at 4:06 PM,  <remyoudompheng@gmail.com> wrote:
> Maybe because the BOM is \xef\xbb\xbf and U+00EF is ï, which is indeed a
> letter.

Yes, effectively. Not because of U+00EF but because the assumption is
that any non-ASCII encountered must be part of an identifier, so we
end up calling getr in the identifier loop. To skip the BOMs we have
to check in getc instead.

[remyoudompheng, 2012-10-07 20:17:26]

LGTM

bombad.go passes again

[rsc, 2012-10-07 20:35:48]

*** Submitted as http://code.google.com/p/go/source/detail?r=3d58333e8e2a ***

cmd/gc: skip over reported BOMs

This keeps the BOM runes from causing other errors.

R=golang-dev, remyoudompheng
CC=golang-dev
http://codereview.appspot.com/6625062