Commit 2f8e16b2 authored by Eli Zaretskii's avatar Eli Zaretskii

Warning comments about subtleties of fetching characters from buffers/strings.

 src/buffer.h (FETCH_CHAR, FETCH_MULTIBYTE_CHAR):
 src/character.h (STRING_CHAR, STRING_CHAR_AND_LENGTH): Add comments
 about subtle differences between FETCH_CHAR* and STRING_CHAR*
 macros related to unification of CJK characters.  For the details,
 see the discussion following the message here:
 http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11073#14.
parent ea0ff314
2012-04-06 Eli Zaretskii <eliz@gnu.org>
* buffer.h (FETCH_CHAR, FETCH_MULTIBYTE_CHAR):
* character.h (STRING_CHAR, STRING_CHAR_AND_LENGTH): Add comments
about subtle differences between FETCH_CHAR* and STRING_CHAR*
macros related to unification of CJK characters. For the details,
see the discussion following the message here:
http://debbugs.gnu.org/cgi/bugreport.cgi?bug=11073#14.
2012-04-04 Chong Yidong <cyd@gnu.org>
* keyboard.c (Vdelayed_warnings_list): Doc fix.
......
......@@ -343,7 +343,8 @@ while (0)
- (ptr - (current_buffer)->text->beg <= GPT_BYTE - BEG_BYTE ? 0 : GAP_SIZE) \
+ BEG_BYTE)
/* Return character at byte position POS. */
/* Return character at byte position POS. See the caveat WARNING for
FETCH_MULTIBYTE_CHAR below. */
#define FETCH_CHAR(pos) \
(!NILP (BVAR (current_buffer, enable_multibyte_characters)) \
......@@ -359,7 +360,17 @@ extern unsigned char *_fetch_multibyte_char_p;
/* Return character code of multi-byte form at byte position POS. If POS
doesn't point the head of valid multi-byte form, only the byte at
POS is returned. No range checking. */
POS is returned. No range checking.
WARNING: The character returned by this macro could be "unified"
inside STRING_CHAR, if the original character in the buffer belongs
to one of the Private Use Areas (PUAs) of codepoints that Emacs
uses to support non-unified CJK characters. If that happens,
CHAR_BYTES will return a value that is different from the length of
the original multibyte sequence stored in the buffer. Therefore,
do _not_ use FETCH_MULTIBYTE_CHAR if you need to advance through
the buffer to the next character after fetching this one. Instead,
use either FETCH_CHAR_ADVANCE or STRING_CHAR_AND_LENGTH. */
#define FETCH_MULTIBYTE_CHAR(pos) \
(_fetch_multibyte_char_p = (((pos) >= GPT_BYTE ? GAP_SIZE : 0) \
......
......@@ -292,7 +292,9 @@ along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */
} while (0)
/* Return the character code of character whose multibyte form is at
P. */
P. Note that this macro unifies CJK characters whose codepoints
are in the Private Use Areas (PUAs), so it might return a different
codepoint from the one actually stored at P. */
#define STRING_CHAR(p) \
(!((p)[0] & 0x80) \
......@@ -309,7 +311,15 @@ along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */
/* Like STRING_CHAR, but set ACTUAL_LEN to the length of multibyte
form. */
form.
Note: This macro returns the actual length of the character's
multibyte sequence as it is stored in a buffer or string. The
character it returns might have a different codepoint that has a
different multibyte sequence of a different legth, due to possible
unification of CJK characters inside string_char. Therefore do NOT
assume that the length returned by this macro is identical to the
length of the multibyte sequence of the character it returns. */
#define STRING_CHAR_AND_LENGTH(p, actual_len) \
(!((p)[0] & 0x80) \
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment