I wrote: > On Mon, Jul 26, 2021 at 7:55 AM Vladimir Sitnikov < sitnikov.vladi...@gmail.com> wrote: > > > > >+ utf8_advance(s, state, len); > > >+ > > >+ /* > > >+ * If we saw an error during the loop, let the caller handle it. We treat > > >+ * all other states as success. > > >+ */ > > >+ if (state == ERR) > > >+ return 0; > > > > Did you mean state = utf8_advance(s, state, len); there? (reassign state variable) > > Yep, that's a bug, thanks for catching!
Fixed in v21, with a regression test added. Also, utf8_advance() now directly changes state by a passed pointer rather than returning a value. Some cosmetic changes: s/valid_bytes/non_error_bytes/ since the former is kind of misleading now. Some other var name and symbol changes. In my first DFA experiment, ASC conflicted with the parser or scanner somehow, but it doesn't here, so it's clearer to use this. Rewrote a lot of comments about the state machine and regression tests. -- John Naylor EDB: http://www.enterprisedb.com
v21-0001-Add-fast-paths-for-validating-UTF-8-text.patch
Description: Binary data
