[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Emacs-diffs] Changes to emacs/src/coding.c [emacs-unicode-2]
|
From: |
Kenichi Handa |
|
Subject: |
[Emacs-diffs] Changes to emacs/src/coding.c [emacs-unicode-2] |
|
Date: |
Sun, 12 Sep 2004 20:46:33 -0400 |
Index: emacs/src/coding.c
diff -c emacs/src/coding.c:1.289.2.36 emacs/src/coding.c:1.289.2.37
*** emacs/src/coding.c:1.289.2.36 Wed Apr 28 12:52:51 2004
--- emacs/src/coding.c Mon Sep 13 00:41:15 2004
***************
*** 1418,1443 ****
}
else if (c1 >= 0 && c2 >= 0)
{
- unsigned char b1[256], b2[256];
- int b1_variants = 1, b2_variants = 1;
- int n;
-
- bzero (b1, 256), bzero (b2, 256);
- b1[c1]++, b2[c2]++;
- for (n = 0; n < 256 && src < src_end; n++)
- {
- src_base = src;
- ONE_MORE_BYTE (c1);
- ONE_MORE_BYTE (c2);
- if (c1 < 0 || c2 < 0)
- break;
- if (! b1[c1++]) b1_variants++;
- if (! b2[c2++]) b2_variants++;
- }
- if (b1_variants < b2_variants)
- detect_info->found |= CATEGORY_MASK_UTF_16_BE_NOSIG;
- else
- detect_info->found |= CATEGORY_MASK_UTF_16_LE_NOSIG;
detect_info->rejected
|= (CATEGORY_MASK_UTF_16_BE | CATEGORY_MASK_UTF_16_LE);
}
--- 1418,1423 ----
***************
*** 5421,5473 ****
if (EQ (CODING_ATTR_TYPE (CODING_ID_ATTRS (coding->id)), Qundecided))
{
int c, i;
for (i = 0, src = coding->source; src < src_end; i++, src++)
{
c = *src;
! if (c & 0x80 || (c < 0x20 && (c == ISO_CODE_ESC
! || c == ISO_CODE_SI
! || c == ISO_CODE_SO)))
break;
}
coding->head_ascii = src - (coding->source + coding->consumed);
! if (coding->head_ascii < coding->src_bytes)
{
- struct coding_detection_info detect_info;
enum coding_category category;
struct coding_system *this;
! detect_info.checked = detect_info.found = detect_info.rejected = 0;
! for (i = 0; i < coding_category_raw_text; i++)
! {
! category = coding_priorities[i];
! this = coding_categories + category;
! if (this->id < 0)
! {
! /* No coding system of this category is defined. */
! detect_info.rejected |= (1 << category);
! }
! else if (category >= coding_category_raw_text)
! continue;
! else if (detect_info.checked & (1 << category))
! {
! if (detect_info.found & (1 << category))
! break;
! }
! else if ((*(this->detector)) (coding, &detect_info)
! && detect_info.found & (1 << category))
! {
! if (category == coding_category_utf_16_auto)
! {
! if (detect_info.found & CATEGORY_MASK_UTF_16_LE)
! category = coding_category_utf_16_le;
! else
! category = coding_category_utf_16_be;
! }
break;
! }
! }
if (i < coding_category_raw_text)
setup_coding_system (CODING_ID_NAME (this->id), coding);
else if (detect_info.rejected == CATEGORY_MASK_ANY)
--- 5401,5478 ----
if (EQ (CODING_ATTR_TYPE (CODING_ID_ATTRS (coding->id)), Qundecided))
{
int c, i;
+ struct coding_detection_info detect_info;
+ detect_info.checked = detect_info.found = detect_info.rejected = 0;
for (i = 0, src = coding->source; src < src_end; i++, src++)
{
c = *src;
! if (c & 0x80)
break;
+ if (c < 0x20
+ && (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
+ && ! inhibit_iso_escape_detection
+ && ! detect_info.checked)
+ {
+ coding->head_ascii = src - (coding->source + coding->consumed);
+ if (detect_coding_iso_2022 (coding, &detect_info))
+ {
+ /* We have scanned the whole data. */
+ if (! (detect_info.rejected & CATEGORY_MASK_ISO_7_ELSE))
+ /* We didn't find an 8-bit code. */
+ src = src_end;
+ break;
+ }
+ }
}
coding->head_ascii = src - (coding->source + coding->consumed);
! if (coding->head_ascii == coding->src_bytes
! || detect_info.found)
{
enum coding_category category;
struct coding_system *this;
! if (coding->head_ascii == coding->src_bytes)
! /* As all bytes are 7-bit, we can ignore non-ISO-2022 codings. */
! for (i = 0; i < coding_category_raw_text; i++)
! {
! category = coding_priorities[i];
! this = coding_categories + category;
! if (detect_info.found & (1 << category))
break;
! }
! else
! for (i = 0; i < coding_category_raw_text; i++)
! {
! category = coding_priorities[i];
! this = coding_categories + category;
! if (this->id < 0)
! {
! /* No coding system of this category is defined. */
! detect_info.rejected |= (1 << category);
! }
! else if (category >= coding_category_raw_text)
! continue;
! else if (detect_info.checked & (1 << category))
! {
! if (detect_info.found & (1 << category))
! break;
! }
! else if ((*(this->detector)) (coding, &detect_info)
! && detect_info.found & (1 << category))
! {
! if (category == coding_category_utf_16_auto)
! {
! if (detect_info.found & CATEGORY_MASK_UTF_16_LE)
! category = coding_category_utf_16_le;
! else
! category = coding_category_utf_16_be;
! }
! break;
! }
! }
!
if (i < coding_category_raw_text)
setup_coding_system (CODING_ID_NAME (this->id), coding);
else if (detect_info.rejected == CATEGORY_MASK_ANY)
***************
*** 7120,7168 ****
for (i = 0; src < src_end; i++, src++)
{
c = *src;
! if (c & 0x80 || (c < 0x20 && (c == ISO_CODE_ESC
! || c == ISO_CODE_SI
! || c == ISO_CODE_SO)))
break;
}
coding.head_ascii = src - coding.source;
! if (src < src_end)
! for (i = 0; i < coding_category_raw_text; i++)
! {
! category = coding_priorities[i];
! this = coding_categories + category;
!
! if (this->id < 0)
! {
! /* No coding system of this category is defined. */
! detect_info.rejected |= (1 << category);
! }
! else if (category >= coding_category_raw_text)
! continue;
! else if (detect_info.checked & (1 << category))
{
! if (highest
! && (detect_info.found & (1 << category)))
break;
}
! else
{
! if ((*(this->detector)) (&coding, &detect_info)
! && highest
! && (detect_info.found & (1 << category)))
{
! if (category == coding_category_utf_16_auto)
{
! if (detect_info.found & CATEGORY_MASK_UTF_16_LE)
! category = coding_category_utf_16_le;
! else
! category = coding_category_utf_16_be;
}
- break;
}
}
! }
if (detect_info.rejected == CATEGORY_MASK_ANY)
{
--- 7125,7197 ----
for (i = 0; src < src_end; i++, src++)
{
c = *src;
! if (c & 0x80)
break;
+ if (c < 0x20
+ && (c == ISO_CODE_ESC || c == ISO_CODE_SI || c == ISO_CODE_SO)
+ && inhibit_iso_escape_detection)
+ {
+ coding.head_ascii = src - coding.source;
+ if (detect_coding_iso_2022 (&coding, &detect_info))
+ {
+ /* We have scanned the whole data. */
+ if (! (detect_info.rejected & CATEGORY_MASK_ISO_7_ELSE))
+ /* We didn't find an 8-bit code. */
+ src = src_end;
+ break;
+ }
+ }
}
coding.head_ascii = src - coding.source;
! if (src < src_end
! || detect_info.found)
! {
! if (src == src_end)
! /* As all bytes are 7-bit, we can ignore non-ISO-2022 codings. */
! for (i = 0; i < coding_category_raw_text; i++)
{
! category = coding_priorities[i];
! if (detect_info.found & (1 << category))
break;
}
! else
! for (i = 0; i < coding_category_raw_text; i++)
{
! category = coding_priorities[i];
! this = coding_categories + category;
!
! if (this->id < 0)
{
! /* No coding system of this category is defined. */
! detect_info.rejected |= (1 << category);
! }
! else if (category >= coding_category_raw_text)
! continue;
! else if (detect_info.checked & (1 << category))
! {
! if (highest
! && (detect_info.found & (1 << category)))
! break;
! }
! else
! {
! if ((*(this->detector)) (&coding, &detect_info)
! && highest
! && (detect_info.found & (1 << category)))
{
! if (category == coding_category_utf_16_auto)
! {
! if (detect_info.found & CATEGORY_MASK_UTF_16_LE)
! category = coding_category_utf_16_le;
! else
! category = coding_category_utf_16_be;
! }
! break;
}
}
}
! }
if (detect_info.rejected == CATEGORY_MASK_ANY)
{
- [Emacs-diffs] Changes to emacs/src/coding.c [emacs-unicode-2],
Kenichi Handa <=