/*
This file is part of LilyPond, the GNU music typesetter.
- Copyright (C) 1997--2011 Jan Nieuwenhuizen <janneke@gnu.org>
+ Copyright (C) 1997--2012 Jan Nieuwenhuizen <janneke@gnu.org>
Han-Wen Nienhuys <hanwen@xs4all.nl>
LilyPond is free software: you can redistribute it and/or modify
get_counts (context_str0, &l, &ch, &col, &offset);
return name_string () + ":" + to_string (l)
- + ":" + to_string (col);
+ + ":" + to_string (col + 1);
}
}
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;
}
}
{
int current_line = get_line (pos_str0);
line_offset_ += line - current_line;
-
+
assert (line == get_line (pos_str0));
}
else