X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=lily%2Fsource-file.cc;h=d847605425b0d9be7a9ec79bf66f0d2b462a8004;hb=f42142c17b68d84721597f8abf400fa18364b118;hp=e6f7a4fb1951a0f66c401368faea76c3da517c39;hpb=0c61221b46addec50e2406e04af44a7d460443d4;p=lilypond.git diff --git a/lily/source-file.cc b/lily/source-file.cc index e6f7a4fb19..d847605425 100644 --- a/lily/source-file.cc +++ b/lily/source-file.cc @@ -1,7 +1,7 @@ /* This file is part of LilyPond, the GNU music typesetter. - Copyright (C) 1997--2011 Jan Nieuwenhuizen + Copyright (C) 1997--2012 Jan Nieuwenhuizen Han-Wen Nienhuys LilyPond is free software: you can redistribute it and/or modify @@ -277,30 +277,27 @@ Source_file::get_counts (char const *pos_str0, char const *line_start = (char const *)data + line[LEFT]; ssize left = (char const *) pos_str0 - line_start; + *byte_offset = left; + string line_begin (line_start, left); char const *line_chars = line_begin.c_str (); - while (left > 0) + for (; left > 0; --left, ++line_chars) { - size_t thislen = utf8_char_len (*line_chars); - - if (thislen == 1 && line_chars[0] == '\t') + // Skip UTF-8 continuation bytes. This is simplistic but + // robust, and we warn against non-UTF-8 input in the lexer + // already. In the case of non-UTF-8 or of this function being + // called in mid-character, the results are somewhat arbitrary, + // but there is no really sane definition anyway. + if ((*line_chars & 0xc0) == 0x80) + continue; + + if (*line_chars == '\t') (*column) = (*column / 8 + 1) * 8; else (*column)++; (*line_char)++; - - /* - To have decent output in UTF-8 aware terminals, - we must keep track of the number of bytes from - the left edge of the terminal. - */ - *byte_offset += thislen; - - /* Advance past this character. */ - line_chars += thislen; - left -= thislen; } }