#endif
#include <stdio.h>
-#include <ctype.h>
#include <string.h>
#include <unistd.h>
#include <unicase.h>
#include "libguile/hashtab.h"
#include "libguile/hash.h"
#include "libguile/ports.h"
+#include "libguile/ports-internal.h"
#include "libguile/fports.h"
#include "libguile/root.h"
#include "libguile/strings.h"
{
int chr;
- chr = scm_get_byte_or_eof (port);
+ chr = scm_get_byte_or_eof_unlocked (port);
if (chr == EOF)
return 0;
else if (CHAR_IS_DELIMITER (chr))
{
- scm_unget_byte (chr, port);
+ scm_unget_byte_unlocked (chr, port);
return 0;
}
else
{
scm_t_wchar c;
while (1)
- switch (c = scm_getc (port))
+ switch (c = scm_getc_unlocked (port))
{
case EOF:
goteof:
case ';':
lp:
- switch (c = scm_getc (port))
+ switch (c = scm_getc_unlocked (port))
{
case EOF:
goto goteof;
break;
case '#':
- switch (c = scm_getc (port))
+ switch (c = scm_getc_unlocked (port))
{
case EOF:
eoferr = "read_sharp";
}
/* fall through */
default:
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
return '#';
}
break;
if (terminating_char == c)
return SCM_EOL;
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
tmp = scm_read_expression (port, opts);
/* Note that it is possible for scm_read_expression to return
"in pair: mismatched close paren: ~A",
scm_list_1 (SCM_MAKE_CHAR (c)));
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
tmp = scm_read_expression (port, opts);
/* See above note about scm_sym_dot. */
c = 0; \
while (i < ndigits) \
{ \
- a = scm_getc (port); \
+ a = scm_getc_unlocked (port); \
if (a == EOF) \
goto str_eof; \
if (terminator \
do
{
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
if (c == EOF)
return;
}
while (c == '\t' || uc_is_general_category (c, UC_SPACE_SEPARATOR));
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
}
static SCM
long line = SCM_LINUM (port);
int column = SCM_COL (port) - 1;
- while ('"' != (c = scm_getc (port)))
+ while ('"' != (c = scm_getc_unlocked (port)))
{
if (c == EOF)
{
if (c == '\\')
{
- switch (c = scm_getc (port))
+ switch (c = scm_getc_unlocked (port))
{
case EOF:
goto str_eof;
SCM result, str = SCM_EOL;
char local_buffer[READER_BUFFER_SIZE], *buffer;
size_t bytes_read;
- scm_t_port *pt = SCM_PTAB_ENTRY (port);
/* Need to capture line and column numbers here. */
long line = SCM_LINUM (port);
int column = SCM_COL (port) - 1;
- scm_ungetc (chr, port);
+ scm_ungetc_unlocked (chr, port);
buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
&bytes_read);
- str = scm_from_stringn (buffer, bytes_read, pt->encoding, pt->ilseq_handler);
+ str = scm_from_port_stringn (buffer, bytes_read, port);
result = scm_string_to_number (str, SCM_UNDEFINED);
if (scm_is_false (result))
size_t bytes_read;
int postfix = (opts->keyword_style == KEYWORD_STYLE_POSTFIX);
char local_buffer[READER_BUFFER_SIZE], *buffer;
- scm_t_port *pt = SCM_PTAB_ENTRY (port);
SCM str;
- scm_ungetc (chr, port);
+ scm_ungetc_unlocked (chr, port);
buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
&bytes_read);
if (bytes_read > 0)
if (postfix && ends_with_colon && (bytes_read > 1))
{
- str = scm_from_stringn (buffer, bytes_read - 1,
- pt->encoding, pt->ilseq_handler);
+ str = scm_from_port_stringn (buffer, bytes_read - 1, port);
if (opts->case_insensitive_p)
str = scm_string_downcase_x (str);
}
else
{
- str = scm_from_stringn (buffer, bytes_read,
- pt->encoding, pt->ilseq_handler);
+ str = scm_from_port_stringn (buffer, bytes_read, port);
if (opts->case_insensitive_p)
str = scm_string_downcase_x (str);
char local_buffer[READER_BUFFER_SIZE], *buffer;
unsigned int radix;
SCM str;
- scm_t_port *pt;
switch (chr)
{
break;
default:
- scm_ungetc (chr, port);
- scm_ungetc ('#', port);
+ scm_ungetc_unlocked (chr, port);
+ scm_ungetc_unlocked ('#', port);
radix = 10;
}
buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
&read);
- pt = SCM_PTAB_ENTRY (port);
- str = scm_from_stringn (buffer, read, pt->encoding, pt->ilseq_handler);
+ str = scm_from_port_stringn (buffer, read, port);
result = scm_string_to_number (str, scm_from_uint (radix));
{
scm_t_wchar c;
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
if ('@' == c)
p = scm_sym_uq_splicing;
else
{
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
p = scm_sym_unquote;
}
break;
{
int c;
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
if ('@' == c)
p = sym_unsyntax_splicing;
else
{
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
p = sym_unsyntax;
}
break;
/* We use the get_byte here because there is no need to get the
locale correct with comment input. This presumes that newline
always represents itself no matter what the encoding is. */
- for (c = scm_get_byte_or_eof (port);
+ for (c = scm_get_byte_or_eof_unlocked (port);
(c != EOF) && (c != '\n');
- c = scm_get_byte_or_eof (port));
+ c = scm_get_byte_or_eof_unlocked (port));
return SCM_UNSPECIFIED;
}
size_t charname_len, bytes_read;
scm_t_wchar cp;
int overflow;
- scm_t_port *pt;
+ scm_t_port_internal *pti;
overflow = read_token (port, opts, buffer, READER_CHAR_NAME_MAX_SIZE,
&bytes_read);
if (bytes_read == 0)
{
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
if (chr == EOF)
scm_i_input_error (FUNC_NAME, port, "unexpected end of file "
"while reading character", SCM_EOL);
return (SCM_MAKE_CHAR (chr));
}
- pt = SCM_PTAB_ENTRY (port);
+ pti = SCM_PORT_GET_INTERNAL (port);
/* Simple ASCII characters can be processed immediately. Also, simple
ISO-8859-1 characters can be processed immediately if the encoding for this
port is ISO-8859-1. */
- if (bytes_read == 1 && ((unsigned char) buffer[0] <= 127 || pt->encoding == NULL))
+ if (bytes_read == 1 &&
+ ((unsigned char) buffer[0] <= 127
+ || pti->encoding_mode == SCM_PORT_ENCODING_MODE_LATIN1))
{
SCM_COL (port) += 1;
return SCM_MAKE_CHAR (buffer[0]);
/* Otherwise, convert the buffer into a proper scheme string for
processing. */
- charname = scm_from_stringn (buffer, bytes_read, pt->encoding,
- pt->ilseq_handler);
+ charname = scm_from_port_stringn (buffer, bytes_read, port);
charname_len = scm_i_string_length (charname);
SCM_COL (port) += charname_len;
cp = scm_i_string_ref (charname, 0);
if (c == '-')
{
sign = -1;
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
}
while ('0' <= c && c <= '9')
"number too large", SCM_EOL);
res = 10*res + c-'0';
got_it = 1;
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
}
if (got_it)
/* Disambiguate between '#f' and uniform floating point vectors. */
if (c == 'f')
{
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
if (c != '3' && c != '6')
{
if (c != EOF)
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
return SCM_BOOL_F;
}
rank = 1;
&& tag_len < sizeof tag_buf / sizeof tag_buf[0])
{
tag_buf[tag_len++] = c;
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
}
if (tag_len == 0)
tag = SCM_BOOL_T;
if (c == '@')
{
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
c = read_decimal_integer (port, c, &lbnd);
}
if (c == ':')
{
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
c = read_decimal_integer (port, c, &len);
if (len < 0)
scm_i_input_error (NULL, port,
scm_read_bytevector (scm_t_wchar chr, SCM port, scm_t_read_opts *opts,
long line, int column)
{
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
if (chr != 'u')
goto syntax;
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
if (chr != '8')
goto syntax;
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
if (chr != '(')
goto syntax;
terribly inefficient but who cares? */
SCM s_bits = SCM_EOL;
- for (chr = scm_getc (port);
+ for (chr = scm_getc_unlocked (port);
(chr != EOF) && ((chr == '0') || (chr == '1'));
- chr = scm_getc (port))
+ chr = scm_getc_unlocked (port))
{
s_bits = scm_cons ((chr == '0') ? SCM_BOOL_F : SCM_BOOL_T, s_bits);
}
if (chr != EOF)
- scm_ungetc (chr, port);
+ scm_ungetc_unlocked (chr, port);
return maybe_annotate_source
(scm_bitvector (scm_reverse_x (s_bits, SCM_EOL)),
for (;;)
{
- int c = scm_getc (port);
+ int c = scm_getc_unlocked (port);
if (c == EOF)
scm_i_input_error ("skip_block_comment", port,
while (i <= READER_DIRECTIVE_NAME_MAX_SIZE)
{
- c = scm_getc (port);
+ c = scm_getc_unlocked (port);
if (c == EOF)
scm_i_input_error ("skip_block_comment", port,
"unterminated `#! ... !#' comment", SCM_EOL);
name[i++] = c;
else if (CHAR_IS_DELIMITER (c))
{
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
name[i] = '\0';
if (0 == strcmp ("r6rs", name))
; /* Silently ignore */
}
else
{
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
break;
}
}
while (i > 0)
- scm_ungetc (name[--i], port);
+ scm_ungetc_unlocked (name[--i], port);
return scm_read_scsh_block_comment (chr, port);
}
nested. So care must be taken. */
int nesting_level = 1;
- int a = scm_getc (port);
+ int a = scm_getc_unlocked (port);
if (a == EOF)
scm_i_input_error ("scm_read_r6rs_block_comment", port,
while (nesting_level > 0)
{
- int b = scm_getc (port);
+ int b = scm_getc_unlocked (port);
if (b == EOF)
scm_i_input_error ("scm_read_r6rs_block_comment", port,
if (EOF == c)
scm_i_input_error ("read_commented_expression", port,
"no expression after #; comment", SCM_EOL);
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
scm_read_expression (port, opts);
return SCM_UNSPECIFIED;
}
buf = scm_i_string_start_writing (buf);
- while ((chr = scm_getc (port)) != EOF)
+ while ((chr = scm_getc_unlocked (port)) != EOF)
{
if (saw_brace)
{
that the extended read syntax would never put a `\' before
an `x'. For now, we just ignore other instances of
backslash in the string. */
- switch ((chr = scm_getc (port)))
+ switch ((chr = scm_getc_unlocked (port)))
{
case EOF:
goto done;
{
SCM result;
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
result = scm_read_sharp_extension (chr, port, opts);
if (!scm_is_eq (result, SCM_UNSPECIFIED))
case '0': case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
case '@':
-#if SCM_ENABLE_DEPRECATED
- /* See below for 'i' and 'e'. */
- case 'a':
- case 'y':
- case 'h':
- case 'l':
-#endif
return (scm_read_array (chr, port, opts, line, column));
case 'i':
case 'e':
-#if SCM_ENABLE_DEPRECATED
- {
- /* When next char is '(', it really is an old-style
- uniform array. */
- scm_t_wchar next_c = scm_getc (port);
- if (next_c != EOF)
- scm_ungetc (next_c, port);
- if (next_c == '(')
- return scm_read_array (chr, port, opts, line, column);
- /* Fall through. */
- }
-#endif
case 'b':
case 'B':
case 'o':
{
scm_t_wchar chr;
- chr = scm_getc (port);
+ chr = scm_getc_unlocked (port);
switch (chr)
{
int c = flush_ws (port, opts, (char *) NULL);
if (c == EOF)
return SCM_EOF_VAL;
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
line = SCM_LINUM (port);
column = SCM_COL (port);
}
new expression. For example, f{n - 1}(x) => ((f (- n 1)) x). */
for (;;)
{
- int chr = scm_getc (port);
+ int chr = scm_getc_unlocked (port);
if (chr == '(')
/* e(...) => (e ...) */
else
{
if (chr != EOF)
- scm_ungetc (chr, port);
+ scm_ungetc_unlocked (chr, port);
break;
}
maybe_annotate_source (expr, port, opts, line, column);
c = flush_ws (port, &opts, (char *) NULL);
if (EOF == c)
return SCM_EOF_VAL;
- scm_ungetc (c, port);
+ scm_ungetc_unlocked (c, port);
return (scm_read_expression (port, &opts));
}
#define SCM_ENCODING_SEARCH_SIZE (500)
+static int
+is_encoding_char (char c)
+{
+ if (c >= 'a' && c <= 'z') return 1;
+ if (c >= 'A' && c <= 'Z') return 1;
+ if (c >= '0' && c <= '9') return 1;
+ return strchr ("_-.:/,+=()", c) != NULL;
+}
+
/* Search the first few hundred characters of a file for an Emacs-like coding
declaration. Returns either NULL or a string whose storage has been
allocated with `scm_gc_malloc ()'. */
pt = SCM_PTAB_ENTRY (port);
if (pt->rw_active == SCM_PORT_WRITE)
- scm_flush (port);
+ scm_flush_unlocked (port);
if (pt->rw_random)
pt->rw_active = SCM_PORT_READ;
if (pt->read_pos == pt->read_end)
{
/* We can use the read buffer, and thus avoid a seek. */
- if (scm_fill_input (port) == EOF)
+ if (scm_fill_input_unlocked (port) == EOF)
return NULL;
bytes_read = pt->read_end - pt->read_pos;
if (SCM_FPORTP (port) && !SCM_FDES_RANDOM_P (SCM_FPORT_FDES (port)))
return NULL;
- bytes_read = scm_c_read (port, header, SCM_ENCODING_SEARCH_SIZE);
+ bytes_read = scm_c_read_unlocked (port, header, SCM_ENCODING_SEARCH_SIZE);
header[bytes_read] = '\0';
scm_seek (port, scm_from_int (0), scm_from_int (SEEK_SET));
}
i = 0;
while (encoding_start + i - header <= SCM_ENCODING_SEARCH_SIZE
&& encoding_start + i - header < bytes_read
- && (isalnum ((int) encoding_start[i])
- || strchr ("_-.:/,+=()", encoding_start[i]) != NULL))
+ && is_encoding_char (encoding_start[i]))
i++;
encoding_length = i;
return NULL;
encoding = scm_gc_strndup (encoding_start, encoding_length, "encoding");
- for (i = 0; i < encoding_length; i++)
- encoding[i] = toupper ((int) encoding[i]);
/* push backwards to make sure we were in a comment */
in_comment = 0;
return SCM_BOOL_F;
else
{
- s_enc = scm_from_locale_string (enc);
+ s_enc = scm_string_upcase (scm_from_locale_string (enc));
return s_enc;
}
unsigned int read_options;
new_value &= READ_OPTION_MASK;
+
+ scm_dynwind_begin (0);
+ scm_dynwind_lock_port (port);
+
scm_read_options = scm_i_port_property (port, sym_port_read_options);
if (scm_is_unsigned_integer (scm_read_options, 0, READ_OPTIONS_MAX_VALUE))
read_options = scm_to_uint (scm_read_options);
read_options |= new_value << option;
scm_read_options = scm_from_uint (read_options);
scm_i_set_port_property_x (port, sym_port_read_options, scm_read_options);
+
+ scm_dynwind_end ();
}
/* Set OPTS and PORT's case-insensitivity according to VALUE. */
HCoop / bpt / guile.git / blobdiff