On Thursday 31 December 2015 19:29:11 Thiago Macieira wrote: > > The offset: 494 > > > > On Ubuntu 15.10, with the commercial 5.5.1, the output is Valid. > > > > > For that matter, in the file that produces the error, is it using CRLF > > > line- endings? The one from your email does. > > > > It doesn't matter: I ran dos2unix on it, but it gives the same result. > > With the same offset?
With CRLF line endings, offset 494 is nowhere near an UTF-8 sequence (it's the first 'y' in "Styczny normalny pędzel"). With LF lineendings, offset 494 is one byte before that 'ę'. I've checked and QJsonDocument's parsing does NOT have an off-by-one error reporting of UTF-8 invalid sequences. $ echo -e "[\"\x80\"]" | ./jsonvalidator /dev/stdin Invalid "invalid UTF8 string" 2 There are no changes to either qjsonparser.cpp or qutfcodec_p.h since v5.5.1 that could account for this bug being fixed. The only two things that I can think of to explain this problem are that either CentOS packages applied a patch that broke the UTF-8 parsing or that their compiler is generating bad code. Sorry. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center _______________________________________________ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
