Am 2015-08-13 um 13:01 schrieb Mattias Gaertner:
On Thu, 13 Aug 2015 12:38:00 +0200 Jürgen Hestermann <juergen.hesterm...@gmx.de> wrote:Am 2015-08-13 um 11:55 schrieb Mattias Gaertner: > A string always ends with a #0, so checking byte by byte makes sure you > stay within range. Not quite true: ------------ if ((ord(p^) and %11110000) = %11100000) then begin // could be 3 byte character if ((ord(p[1]) and %11000000) = %10000000) and ((ord(p[2]) and %11000000) = %10000000) then ... ... ------------ In the above (current) code 3 bytes are accessed which may step behind the zero byte.The "and" operator stops evaluating if left side is already false.
Yes, I see now that I somehow missinterpreted the code. You are right that for the case that a zero byte exists it would not access further bytes within UTF8CharacterLength. Still I think it would be better to give back 3 in case the byte actually means 3 because 1 byte does not form a valid UTF-8 character. If I rely on this result I would try to use this 1 byte as a valid UTF-8 character which would be wrong so I have to apply further checks to cope with this situation anyway. Then I can also check whether the 3 or 4 bytes of the correct result exist. I would not loose anything for invalid UTF-8 strings but I would gain performance if I can guarantee valid UTF-8 string. And if no zero byte exists (for whatever reason) it currently fails anyway. -- _______________________________________________ Lazarus mailing list Lazarus@lists.lazarus.freepascal.org http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
