Commit db327c7e authored by Kenichi Handa's avatar Kenichi Handa
Browse files

Include "character.h" instead of "charset.h".

(Finsert_file_contents): Big change for the new code-conversion
API.
(choose_write_coding_system): Likewise.
(Fwrite_region): Likewise.
(build_annotations_2): Deleted.
(e_write): Big change for the new code-conversion API.
parent e37d7195
......@@ -87,7 +87,7 @@ extern int errno;
#include "lisp.h"
#include "intervals.h"
#include "buffer.h"
#include "charset.h"
#include "character.h"
#include "coding.h"
#include "window.h"
......@@ -3576,7 +3576,7 @@ actually used. */)
unsigned char buffer[1 << 14];
int replace_handled = 0;
int set_coding_system = 0;
int coding_system_decided = 0;
Lisp_Object coding_system;
int read_quit = 0;
if (current_buffer->base_buffer && ! NILP (visit))
......@@ -3713,20 +3713,20 @@ actually used. */)
}
}
/* The value Qnil means that the coding system is not yet
decided. */
coding_system = Qnil;
if (BEG < Z)
{
/* Decide the coding system to use for reading the file now
because we can't use an optimized method for handling
`coding:' tag if the current buffer is not empty. */
Lisp_Object val;
val = Qnil;
if (!NILP (Vcoding_system_for_read))
val = Vcoding_system_for_read;
coding_system = Vcoding_system_for_read;
else if (! NILP (replace))
/* In REPLACE mode, we can use the same coding system
that was used to visit the file. */
val = current_buffer->buffer_file_coding_system;
coding_system = current_buffer->buffer_file_coding_system;
else
{
/* Don't try looking inside a file for a coding system
......@@ -3773,8 +3773,8 @@ actually used. */)
current_buffer->enable_multibyte_characters = Qnil;
insert_1_both (read_buf, nread, nread, 0, 0, 0);
TEMP_SET_PT_BOTH (BEG, BEG_BYTE);
val = call2 (Vset_auto_coding_function,
filename, make_number (nread));
coding_system = call2 (Vset_auto_coding_function,
filename, make_number (nread));
set_buffer_internal (prev);
/* Remove the binding for standard-output. */
......@@ -3791,7 +3791,7 @@ actually used. */)
}
}
if (NILP (val))
if (NILP (coding_system))
{
/* If we have not yet decided a coding system, check
file-coding-system-alist. */
......@@ -3805,20 +3805,19 @@ actually used. */)
}
}
setup_coding_system (Fcheck_coding_system (val), &coding);
/* Ensure we set Vlast_coding_system_used. */
set_coding_system = 1;
if (NILP (coding_system))
coding_system = Qundecided;
else
CHECK_CODING_SYSTEM (coding_system);
if (NILP (current_buffer->enable_multibyte_characters)
&& ! NILP (val))
if (NILP (current_buffer->enable_multibyte_characters))
/* We must suppress all character code conversion except for
end-of-line conversion. */
setup_raw_text_coding_system (&coding);
coding_system = raw_text_coding_system (coding_system);
coding.src_multibyte = 0;
coding.dst_multibyte
= !NILP (current_buffer->enable_multibyte_characters);
coding_system_decided = 1;
setup_coding_system (coding_system, &coding);
/* Ensure we set Vlast_coding_system_used. */
set_coding_system = 1;
}
/* If requested, replace the accessible part of the buffer
......@@ -3837,7 +3836,8 @@ actually used. */)
and let the following if-statement handle the replace job. */
if (!NILP (replace)
&& BEGV < ZV
&& !(coding.common_flags & CODING_REQUIRE_DECODING_MASK))
&& (NILP (coding_system)
|| ! CODING_REQUIRE_DECODING (&coding)))
{
/* same_at_start and same_at_end count bytes,
because file access counts bytes
......@@ -3872,21 +3872,14 @@ actually used. */)
else if (nread == 0)
break;
if (coding.type == coding_type_undecided)
detect_coding (&coding, buffer, nread);
if (coding.common_flags & CODING_REQUIRE_DECODING_MASK)
/* We found that the file should be decoded somehow.
Let's give up here. */
if (CODING_REQUIRE_DETECTION (&coding))
{
giveup_match_end = 1;
break;
coding_system = detect_coding_system (buffer, nread, 1, 0,
coding_system);
setup_coding_system (coding_system, &coding);
}
if (coding.eol_type == CODING_EOL_UNDECIDED)
detect_eol (&coding, buffer, nread);
if (coding.eol_type != CODING_EOL_UNDECIDED
&& coding.eol_type != CODING_EOL_LF)
/* We found that the format of eol should be decoded.
if (CODING_REQUIRE_DECODING (&coding))
/* We found that the file should be decoded somehow.
Let's give up here. */
{
giveup_match_end = 1;
......@@ -4033,17 +4026,20 @@ actually used. */)
int same_at_end = ZV_BYTE;
int overlap;
int bufpos;
/* Make sure that the gap is large enough. */
int bufsize = 2 * st.st_size;
unsigned char *conversion_buffer = (unsigned char *) xmalloc (bufsize);
unsigned char *decoded;
int temp;
int this_count = BINDING_STACK_SIZE ();
Lisp_Object conversion_buffer
= make_conversion_work_buffer (! NILP (current_buffer
->enable_multibyte_characters));
record_unwind_protect (code_conversion_restore, save_excursion_save ());
/* First read the whole file, performing code conversion into
CONVERSION_BUFFER. */
if (lseek (fd, XINT (beg), 0) < 0)
{
xfree (conversion_buffer);
report_file_error ("Setting file position",
Fcons (orig_filename, Qnil));
}
......@@ -4055,71 +4051,41 @@ actually used. */)
while (how_much < total)
{
/* We read one bunch by one (READ_BUF_SIZE bytes) to allow
quitting while reading a huge while. */
/* try is reserved in some compilers (Microsoft C) */
int trytry = min (total - how_much, READ_BUF_SIZE - unprocessed);
unsigned char *destination = read_buf + unprocessed;
int this;
/* Allow quitting out of the actual I/O. */
immediate_quit = 1;
QUIT;
this = emacs_read (fd, destination, trytry);
this = emacs_read (fd, read_buf + unprocessed, trytry);
immediate_quit = 0;
if (this < 0 || this + unprocessed == 0)
if (this <= 0)
{
how_much = this;
if (this < 0)
how_much = this;
break;
}
how_much += this;
if (CODING_MAY_REQUIRE_DECODING (&coding))
{
int require, result;
this += unprocessed;
/* If we are using more space than estimated,
make CONVERSION_BUFFER bigger. */
require = decoding_buffer_size (&coding, this);
if (inserted + require + 2 * (total - how_much) > bufsize)
{
bufsize = inserted + require + 2 * (total - how_much);
conversion_buffer = (unsigned char *) xrealloc (conversion_buffer, bufsize);
}
/* Convert this batch with results in CONVERSION_BUFFER. */
if (how_much >= total) /* This is the last block. */
coding.mode |= CODING_MODE_LAST_BLOCK;
if (coding.composing != COMPOSITION_DISABLED)
coding_allocate_composition_data (&coding, BEGV);
result = decode_coding (&coding, read_buf,
conversion_buffer + inserted,
this, bufsize - inserted);
/* Save for next iteration whatever we didn't convert. */
unprocessed = this - coding.consumed;
bcopy (read_buf + coding.consumed, read_buf, unprocessed);
if (!NILP (current_buffer->enable_multibyte_characters))
this = coding.produced;
else
this = str_as_unibyte (conversion_buffer + inserted,
coding.produced);
}
inserted += this;
}
decode_coding_c_string (&coding, read_buf, unprocessed + this,
conversion_buffer);
unprocessed = coding.carryover_bytes;
if (coding.carryover_bytes > 0)
bcopy (coding.carryover, read_buf, unprocessed);
}
emacs_close (fd);
/* At this point, INSERTED is how many characters (i.e. bytes)
are present in CONVERSION_BUFFER.
HOW_MUCH should equal TOTAL,
or should be <= 0 if we couldn't read the file. */
/* At this point, HOW_MUCH should equal TOTAL, or should be <= 0
if we couldn't read the file. */
if (how_much < 0)
{
xfree (conversion_buffer);
if (how_much == -1)
error ("IO error reading %s: %s",
XSTRING (orig_filename)->data, emacs_strerror (errno));
......@@ -4127,21 +4093,30 @@ actually used. */)
error ("maximum buffer size exceeded");
}
/* Compare the beginning of the converted file
with the buffer text. */
if (unprocessed > 0)
{
coding.mode |= CODING_MODE_LAST_BLOCK;
decode_coding_c_string (&coding, read_buf, unprocessed,
conversion_buffer);
coding.mode &= ~CODING_MODE_LAST_BLOCK;
}
decoded = BUF_BEG_ADDR (XBUFFER (conversion_buffer));
inserted = BUF_Z_BYTE (XBUFFER (conversion_buffer));
/* Compare the beginning of the converted string with the buffer
text. */
bufpos = 0;
while (bufpos < inserted && same_at_start < same_at_end
&& FETCH_BYTE (same_at_start) == conversion_buffer[bufpos])
&& FETCH_BYTE (same_at_start) == decoded[bufpos])
same_at_start++, bufpos++;
/* If the file matches the buffer completely,
/* If the file matches the head of buffer completely,
there's no need to replace anything. */
if (bufpos == inserted)
{
xfree (conversion_buffer);
emacs_close (fd);
specpdl_ptr--;
/* Truncate the buffer to the size of the file. */
del_range_byte (same_at_start, same_at_end, 0);
......@@ -4149,8 +4124,8 @@ actually used. */)
goto handled;
}
/* Extend the start of non-matching text area to multibyte
character boundary. */
/* Extend the start of non-matching text area to the previous
multibyte character boundary. */
if (! NILP (current_buffer->enable_multibyte_characters))
while (same_at_start > BEGV_BYTE
&& ! CHAR_HEAD_P (FETCH_BYTE (same_at_start)))
......@@ -4163,11 +4138,11 @@ actually used. */)
/* Compare with same_at_start to avoid counting some buffer text
as matching both at the file's beginning and at the end. */
while (bufpos > 0 && same_at_end > same_at_start
&& FETCH_BYTE (same_at_end - 1) == conversion_buffer[bufpos - 1])
&& FETCH_BYTE (same_at_end - 1) == decoded[bufpos - 1])
same_at_end--, bufpos--;
/* Extend the end of non-matching text area to multibyte
character boundary. */
/* Extend the end of non-matching text area to the next
multibyte character boundary. */
if (! NILP (current_buffer->enable_multibyte_characters))
while (same_at_end < ZV_BYTE
&& ! CHAR_HEAD_P (FETCH_BYTE (same_at_end)))
......@@ -4185,11 +4160,11 @@ actually used. */)
/* Replace the chars that we need to replace,
and update INSERTED to equal the number of bytes
we are taking from the file. */
we are taking from the decoded string. */
inserted -= (Z_BYTE - same_at_end) + (same_at_start - BEG_BYTE);
if (same_at_end != same_at_start)
{
{
del_range_byte (same_at_start, same_at_end, 0);
temp = GPT;
same_at_start = GPT_BYTE;
......@@ -4200,18 +4175,16 @@ actually used. */)
}
/* Insert from the file at the proper position. */
SET_PT_BOTH (temp, same_at_start);
insert_1 (conversion_buffer + same_at_start - BEG_BYTE, inserted,
0, 0, 0);
if (coding.cmp_data && coding.cmp_data->used)
coding_restore_composition (&coding, Fcurrent_buffer ());
coding_free_composition_data (&coding);
insert_from_buffer (XBUFFER (conversion_buffer),
buf_bytepos_to_charpos (XBUFFER (conversion_buffer),
same_at_start),
buf_bytepos_to_charpos (XBUFFER (conversion_buffer),
same_at_start + inserted),
0);
/* Set `inserted' to the number of inserted characters. */
inserted = PT - temp;
xfree (conversion_buffer);
emacs_close (fd);
specpdl_ptr--;
unbind_to (this_count, Qnil);
goto handled;
}
......@@ -4255,7 +4228,7 @@ actually used. */)
inserted = 0;
/* Here, we don't do code conversion in the loop. It is done by
code_convert_region after all data are read into the buffer. */
decode_coding_gap after all data are read into the buffer. */
{
int gap_size = GAP_SIZE;
......@@ -4346,18 +4319,16 @@ actually used. */)
notfound:
if (! coding_system_decided)
if (NILP (coding_system))
{
/* The coding system is not yet decided. Decide it by an
optimized method for handling `coding:' tag.
Note that we can get here only if the buffer was empty
before the insertion. */
Lisp_Object val;
val = Qnil;
if (!NILP (Vcoding_system_for_read))
val = Vcoding_system_for_read;
coding_system = Vcoding_system_for_read;
else
{
/* Since we are sure that the current buffer was empty
......@@ -4378,11 +4349,11 @@ actually used. */)
if (inserted > 0 && ! NILP (Vset_auto_coding_function))
{
val = call2 (Vset_auto_coding_function,
filename, make_number (inserted));
coding_system = call2 (Vset_auto_coding_function,
filename, make_number (inserted));
}
if (NILP (val))
if (NILP (coding_system))
{
/* If the coding system is not yet decided, check
file-coding-system-alist. */
......@@ -4392,66 +4363,70 @@ actually used. */)
args[2] = visit, args[3] = beg, args[4] = end, args[5] = Qnil;
coding_systems = Ffind_operation_coding_system (6, args);
if (CONSP (coding_systems))
val = XCAR (coding_systems);
coding_system = XCAR (coding_systems);
}
unbind_to (count, Qnil);
inserted = Z_BYTE - BEG_BYTE;
}
/* The following kludgy code is to avoid some compiler bug.
We can't simply do
setup_coding_system (val, &coding);
on some system. */
{
struct coding_system temp_coding;
setup_coding_system (val, &temp_coding);
bcopy (&temp_coding, &coding, sizeof coding);
}
/* Ensure we set Vlast_coding_system_used. */
set_coding_system = 1;
if (NILP (coding_system))
coding_system = Qundecided;
else
CHECK_CODING_SYSTEM (coding_system);
if (NILP (current_buffer->enable_multibyte_characters)
&& ! NILP (val))
if (NILP (current_buffer->enable_multibyte_characters))
/* We must suppress all character code conversion except for
end-of-line conversion. */
setup_raw_text_coding_system (&coding);
coding.src_multibyte = 0;
coding.dst_multibyte
= !NILP (current_buffer->enable_multibyte_characters);
}
coding_system = raw_text_coding_system (coding_system);
if (!NILP (visit)
/* Can't do this if part of the buffer might be preserved. */
&& NILP (replace)
&& (coding.type == coding_type_no_conversion
|| coding.type == coding_type_raw_text))
{
/* Visiting a file with these coding system makes the buffer
unibyte. */
current_buffer->enable_multibyte_characters = Qnil;
coding.dst_multibyte = 0;
setup_coding_system (coding_system, &coding);
/* Ensure we set Vlast_coding_system_used. */
set_coding_system = 1;
}
if (inserted > 0 || coding.type == coding_type_ccl)
if (!NILP (visit))
{
if (CODING_MAY_REQUIRE_DECODING (&coding))
/* When we visit a file by raw-text, we change the buffer to
unibyte. If we have not yet decided how to decode a text,
decide it at first by detecting the file's encoding. */
if (CODING_REQUIRE_DETECTION (&coding))
{
code_convert_region (PT, PT_BYTE, PT + inserted, PT_BYTE + inserted,
&coding, 0, 0);
inserted = coding.produced_char;
coding_system = detect_coding_system (PT_ADDR, inserted, 1, 0,
coding_system);
setup_coding_system (coding_system, &coding);
}
else
adjust_after_insert (PT, PT_BYTE, PT + inserted, PT_BYTE + inserted,
inserted);
if (CODING_FOR_UNIBYTE (&coding)
/* Can't do this if part of the buffer might be preserved. */
&& NILP (replace))
/* Visiting a file with these coding system makes the buffer
unibyte. */
current_buffer->enable_multibyte_characters = Qnil;
}
if (CODING_REQUIRE_DETECTION (&coding)
|| CODING_REQUIRE_DECODING (&coding))
{
move_gap_both (PT, PT_BYTE);
GAP_SIZE += inserted;
ZV_BYTE -= inserted;
Z_BYTE -= inserted;
ZV -= inserted;
Z -= inserted;
decode_coding_gap (&coding, inserted, inserted);
inserted = coding.produced_char;
}
else if (inserted > 0)
adjust_after_insert (PT, PT_BYTE, PT + inserted, PT_BYTE + inserted,
inserted);
#ifdef DOS_NT
/* Use the conversion type to determine buffer-file-type
(find-buffer-file-type is now used to help determine the
conversion). */
if ((coding.eol_type == CODING_EOL_UNDECIDED
|| coding.eol_type == CODING_EOL_LF)
if ((coding.eol_type == eol_type_undecided
|| coding.eol_type == eol_type_lf)
&& ! CODING_REQUIRE_DECODING (&coding))
current_buffer->buffer_file_type = Qt;
else
......@@ -4515,7 +4490,7 @@ actually used. */)
}
if (set_coding_system)
Vlast_coding_system_used = coding.symbol;
Vlast_coding_system_used = coding_system;
/* Call after-change hooks for the inserted text, aside from the case
of normal visiting (not with REPLACE), which is done in a new buffer
......@@ -4560,8 +4535,6 @@ actually used. */)
}
static Lisp_Object build_annotations P_ ((Lisp_Object, Lisp_Object));
static Lisp_Object build_annotations_2 P_ ((Lisp_Object, Lisp_Object,
Lisp_Object, Lisp_Object));
/* If build_annotations switched buffers, switch back to BUF.
Kill the temporary buffer that was selected in the meantime.
......@@ -4643,40 +4616,36 @@ choose_write_coding_system (start, end, filename,
using_default_coding = 1;
}
if (! NILP (val) && ! force_raw_text)
{
Lisp_Object spec, attrs;
CHECK_CODING_SYSTEM_GET_SPEC (val, spec);
attrs = AREF (spec, 0);
if (EQ (CODING_ATTR_TYPE (attrs), Qraw_text))
force_raw_text = 1;
}
if (!force_raw_text
&& !NILP (Ffboundp (Vselect_safe_coding_system_function)))
/* Confirm that VAL can surely encode the current region. */
val = call3 (Vselect_safe_coding_system_function, start, end, val);
setup_coding_system (Fcheck_coding_system (val), coding);
if (coding->eol_type == CODING_EOL_UNDECIDED
&& !using_default_coding)
{
if (! EQ (default_buffer_file_coding.symbol,
buffer_defaults.buffer_file_coding_system))
setup_coding_system (buffer_defaults.buffer_file_coding_system,
&default_buffer_file_coding);
if (default_buffer_file_coding.eol_type != CODING_EOL_UNDECIDED)
{
Lisp_Object subsidiaries;
coding->eol_type = default_buffer_file_coding.eol_type;
subsidiaries = Fget (coding->symbol, Qeol_type);
if (VECTORP (subsidiaries)
&& XVECTOR (subsidiaries)->size == 3)
coding->symbol
= XVECTOR (subsidiaries)->contents[coding->eol_type];
}
}
/* If the decided coding-system doesn't specify end-of-line
format, we use that of
`default-buffer-file-coding-system'. */
if (! using_default_coding)
val = (coding_inherit_eol_type
(val, buffer_defaults.buffer_file_coding_system));
/* If we decide not to encode text, use `raw-text' or one of its
subsidiaries. */
if (force_raw_text)
setup_raw_text_coding_system (coding);
goto done_setup_coding;
val = raw_text_coding_system (val);
}
setup_coding_system (Fcheck_coding_system (val), coding);
done_setup_coding:
if (!STRINGP (start) && !NILP (current_buffer->selective_display))
coding->mode |= CODING_MODE_SELECTIVE_DISPLAY;
}
......@@ -4816,11 +4785,9 @@ This does code conversion according to the value of
unsavable chars (as was the case with X-Symbol). */
choose_write_coding_system (start, end, filename,
append, visit, lockname, &coding);
Vlast_coding_system_used = coding.symbol;
Vlast_coding_system_used = CODING_ID_NAME (coding.id);
given_buffer = current_buffer;
annotations = build_annotations_2 (start, end,
coding.pre_write_conversion, annotations);
if (current_buffer != given_buffer)
{
XSETFASTINT (start, BEGV);
......@@ -4963,6 +4930,9 @@ This does code conversion according to the value of
if (GPT > BEG && GPT_ADDR[-1] != '\n')
move_gap (find_next_newline (GPT, 1));
#else
#if 0
/* The new encoding routine doesn't require the following. */
/* Whether VMS or not, we must move the gap to the next of newline
when we must put designation sequences at beginning of line. */
if (INTEGERP (start)
......@@ -4975,6 +4945,7 @@ This does code conversion according to the value of
move_gap_both (PT, PT_BYTE);
SET_PT_BOTH (opoint, opoint_byte);
}
#endif
#endif
failure = 0;
......@@ -4988,23 +4959,10 @@ This does code conversion according to the value of
}
else if (XINT (start) != XINT (end))
{
tem = CHAR_TO_BYTE (XINT (start));
if (XINT (start) < GPT)
{
failure = 0 > a_write (desc, Qnil, XINT (start),
min (GPT, XINT (end)) - XINT (start),
&annotations, &coding);
save_errno = errno;
}
if (XINT (end) > GPT && !failure)
{
tem = max (XINT (start), GPT);
failure = 0 > a_write (desc, Qnil, tem , XINT (end) - tem,
&annotations, &coding);
save_errno = errno;
}
failure = 0 > a_write (desc, Qnil,
XINT (start), XINT (end) - XINT (start),
&annotations, &coding);
save_errno = errno;
}
else
{
......@@ -5020,7 +4978,7 @@ This does code conversion according to the value of
{
/* We have to flush out a data. */
coding.mode |= CODING_MODE_LAST_BLOCK;
failure = 0 > e_write (desc, Qnil, 0, 0, &coding);
failure = 0 > e_write (desc, Qnil, 1, 1, &coding);
save_errno = errno;
}
......@@ -5196,30 +5154,6 @@ build_annotations (start, end)
return annotations;
}
static Lisp_Object
build_annotations_2 (start, end, pre_write_conversion, annotations)
Lisp_Object start, end, pre_write_conversion, annotations;
{
struct gcpro gcpro1;
Lisp_Object res;
GCPRO1 (annotations);
/* At last, do the same for the function PRE_WRITE_CONVERSION
implied by the current coding-system. */
if (!NILP (pre_write_conversion))
{
struct buffer *given_buffer = current_buffer;
Vwrite_region_annotations_so_far = annotations;
res = call2 (pre_write_conversion, start, end);
Flength (res);
annotations = (current_buffer != given_buffer
? res
: merge (annotations, res, Qcar_less_than_car));
}
UNGCPRO;
return annotations;
}
/* Write to descriptor DESC the NCHARS chars starting at POS of STRING.
If STRING is nil, POS is the character position in the current buffer.
......@@ -5291,76 +5225,45 @@ e_write (desc, string, start, end, coding)
int start, end;
struct coding_system *coding;
{
register char *addr;
register int nbytes;
char buf[WRITE_BUF_SIZE];
int return_val = 0;
if (start >= end)