void
html_ul(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
unsigned char *al;
/* dump_html_stack(html_context); */
par_format.list_level++;
par_format.list_number = 0;
par_format.flags = P_DISC;
al = get_attr_val(a, (unsigned char *)"type", html_context->doc_cp);
if (al) {
if (!c_strcasecmp((const char *)al, "disc"))
par_format.flags = P_DISC;
else if (!c_strcasecmp((const char *)al, "circle"))
par_format.flags = P_O;
else if (!c_strcasecmp((const char *)al, "square"))
par_format.flags = P_SQUARE;
mem_free(al);
}
par_format.leftmargin += 2 + (par_format.list_level > 1);
if (!html_context->table_level)
int_upper_bound(&par_format.leftmargin, par_format.width / 2);
par_format.align = ALIGN_LEFT;
html_top->type = ELEMENT_DONT_KILL;
}
/* Returns the encoding specified by ENCODING, which must be in one of the
forms described at the top of encoding-guesser.h. The returned string might
be ENCODING itself or a suffix of it, or it might be a statically allocated
string. */
const char *
encoding_guess_parse_encoding (const char *encoding)
{
if (encoding == NULL
|| !c_strcasecmp (encoding, "auto")
|| !c_strcasecmp (encoding, "auto,locale")
|| !c_strcasecmp (encoding, "locale"))
return locale_charset ();
else if (!c_strncasecmp (encoding, "auto,", 5))
return encoding + 5;
else
return encoding;
}
void
html_button(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
unsigned char *al;
struct form_control *fc;
enum form_type type = FC_SUBMIT;
int cp = html_context->doc_cp;
html_focusable(html_context, a);
al = get_attr_val(a, (unsigned char *)"type", cp);
if (!al) goto no_type_attr;
if (!c_strcasecmp((const char *)al, "button")) {
type = FC_BUTTON;
} else if (!c_strcasecmp((const char *)al, "reset")) {
type = FC_RESET;
} else if (c_strcasecmp((const char *)al, "submit")) {
/* unknown type */
mem_free(al);
return;
}
mem_free(al);
no_type_attr:
fc = init_form_control(type, a, html_context);
if (!fc) return;
fc->id = get_attr_val(a, (unsigned char *)"id", cp);
fc->name = get_attr_val(a, (unsigned char *)"name", cp);
fc->default_value = get_attr_val(a, (unsigned char *)"value", cp);
if (!fc->default_value) {
if (fc->type == FC_SUBMIT)
fc->default_value = stracpy((const unsigned char *)"Submit");
else if (fc->type == FC_RESET)
fc->default_value = stracpy((const unsigned char *)"Reset");
else if (fc->type == FC_BUTTON)
fc->default_value = stracpy((const unsigned char *)"Button");
}
if (!fc->default_value)
fc->default_value = stracpy((const unsigned char *)"");
html_context->special_f(html_context, SP_CONTROL, fc);
format.form = fc;
format.style.attr |= AT_BOLD;
}
/* Canonicalize an encoding name. */
const char *
po_charset_canonicalize (const char *charset)
{
/* The list of charsets supported by glibc's iconv() and by the portable
iconv() across platforms. Taken from intl/config.charset. */
static const char *standard_charsets[] =
{
ascii, "ANSI_X3.4-1968", "US-ASCII", /* i = 0..2 */
"ISO-8859-1", "ISO_8859-1", /* i = 3, 4 */
"ISO-8859-2", "ISO_8859-2",
"ISO-8859-3", "ISO_8859-3",
"ISO-8859-4", "ISO_8859-4",
"ISO-8859-5", "ISO_8859-5",
"ISO-8859-6", "ISO_8859-6",
"ISO-8859-7", "ISO_8859-7",
"ISO-8859-8", "ISO_8859-8",
"ISO-8859-9", "ISO_8859-9",
"ISO-8859-13", "ISO_8859-13",
"ISO-8859-14", "ISO_8859-14",
"ISO-8859-15", "ISO_8859-15", /* i = 25, 26 */
"KOI8-R",
"KOI8-U",
"KOI8-T",
"CP850",
"CP866",
"CP874",
"CP932",
"CP949",
"CP950",
"CP1250",
"CP1251",
"CP1252",
"CP1253",
"CP1254",
"CP1255",
"CP1256",
"CP1257",
"GB2312",
"EUC-JP",
"EUC-KR",
"EUC-TW",
"BIG5",
"BIG5-HKSCS",
"GBK",
"GB18030",
"SHIFT_JIS",
"JOHAB",
"TIS-620",
"VISCII",
"GEORGIAN-PS",
utf8
};
size_t i;
for (i = 0; i < SIZEOF (standard_charsets); i++)
if (c_strcasecmp (charset, standard_charsets[i]) == 0)
return standard_charsets[i < 3 ? 0 : i < 27 ? ((i - 3) & ~1) + 3 : i];
return NULL;
}
void
html_linebrk(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
unsigned char *al = get_attr_val(a, (unsigned char *)"align", html_context->doc_cp);
if (al) {
if (!c_strcasecmp((const char *)al, "left")) par_format.align = ALIGN_LEFT;
else if (!c_strcasecmp((const char *)al, "right")) par_format.align = ALIGN_RIGHT;
else if (!c_strcasecmp((const char *)al, "center")) {
par_format.align = ALIGN_CENTER;
if (!html_context->table_level)
par_format.leftmargin = par_format.rightmargin = 0;
} else if (!c_strcasecmp((const char *)al, "justify")) par_format.align = ALIGN_JUSTIFY;
mem_free(al);
}
}
static widget_creation_function
find_in_table (const char *type, const widget_creation_entry *table)
{
const widget_creation_entry* cur;
for (cur = table; cur->type; cur++)
if (!c_strcasecmp (type, cur->type))
return cur->function;
return NULL;
}
/* Parse a link and return results in @link.
* It tries to identify known types. */
static int
html_link_parse(struct html_context *html_context, unsigned char *a,
struct hlink *link)
{
int i;
assert(a && link);
memset(link, 0, sizeof(*link));
link->href = get_url_val(a, (unsigned char *)"href", html_context->doc_cp);
if (!link->href) return 0;
link->lang = get_attr_val(a, (unsigned char *)"lang", html_context->doc_cp);
link->hreflang = get_attr_val(a, (unsigned char *)"hreflang", html_context->doc_cp);
link->title = get_attr_val(a, (unsigned char *)"title", html_context->doc_cp);
link->content_type = get_attr_val(a, (unsigned char *)"type", html_context->doc_cp);
link->media = get_attr_val(a, (unsigned char *)"media", html_context->doc_cp);
link->name = get_attr_val(a, (unsigned char *)"rel", html_context->doc_cp);
if (link->name) {
link->direction = LD_REL;
} else {
link->name = get_attr_val(a, (unsigned char *)"rev", html_context->doc_cp);
if (link->name) link->direction = LD_REV;
}
if (!link->name) return 1;
/* TODO: fastfind */
for (i = 0; lt_names[i].str; i++)
if (!c_strcasecmp((const char *)link->name, (const char *)lt_names[i].str)) {
link->type = lt_names[i].type;
return 1;
}
if (c_strcasestr((const char *)link->name, "icon") ||
(link->content_type && c_strcasestr((const char *)link->content_type, "icon"))) {
link->type = LT_ICON;
} else if (c_strcasestr((const char *)link->name, "alternate")) {
link->type = LT_ALTERNATE;
if (link->lang)
link->type = LT_ALTERNATE_LANG;
else if (c_strcasestr((const char *)link->name, "stylesheet") ||
(link->content_type && c_strcasestr((const char *)link->content_type, "css")))
link->type = LT_ALTERNATE_STYLESHEET;
else if (link->media)
link->type = LT_ALTERNATE_MEDIA;
} else if (link->content_type && c_strcasestr((const char *)link->content_type, "css")) {
link->type = LT_STYLESHEET;
}
return 1;
}
int
name_to_language(const unsigned char *name)
{
int i;
for (i = 0; languages[i].name; i++) {
if (c_strcasecmp(languages[i].name, name))
continue;
return i;
}
return 1;
}
int
compare_opt(struct document_options *o1, struct document_options *o2)
{
return memcmp(o1, o2, offsetof(struct document_options, framename))
|| c_strcasecmp(o1->framename, o2->framename)
|| (o1->box.x != o2->box.x)
|| (o1->box.y != o2->box.y)
|| ((o1->needs_height || o2->needs_height)
&& o1->box.height != o2->box.height)
|| ((o1->needs_width || o2->needs_width)
&& o1->box.width != o2->box.width);
}
char *
str_iconveha (const char *src,
const char *from_codeset, const char *to_codeset,
bool transliterate,
enum iconv_ilseq_handler handler)
{
if (*src == '\0' || c_strcasecmp (from_codeset, to_codeset) == 0)
{
char *result = strdup (src);
if (result == NULL)
errno = ENOMEM;
return result;
}
/* When using GNU libc >= 2.2 or GNU libiconv >= 1.5,
we want to use transliteration. */
#if (((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 2) || __GLIBC__ > 2) \
&& !defined __UCLIBC__) \
|| _LIBICONV_VERSION >= 0x0105
if (transliterate)
{
char *result;
size_t len = strlen (to_codeset);
char *to_codeset_suffixed = (char *) malloca (len + 10 + 1);
memcpy (to_codeset_suffixed, to_codeset, len);
memcpy (to_codeset_suffixed + len, "//TRANSLIT", 10 + 1);
result = str_iconveha_notranslit (src, from_codeset, to_codeset_suffixed,
handler);
freea (to_codeset_suffixed);
return result;
}
else
#endif
return str_iconveha_notranslit (src, from_codeset, to_codeset, handler);
}
/* ToUnicode(). May realloc() utf8in. Will free utf8in unconditionally. */
static int
idna_to_unicode_internal (char *utf8in,
uint32_t * out, size_t * outlen, int flags)
{
int rc;
char tmpout[64];
size_t utf8len = strlen (utf8in) + 1;
size_t addlen = 0;
/*
* ToUnicode consists of the following steps:
*
* 1. If the sequence contains any code points outside the ASCII range
* (0..7F) then proceed to step 2, otherwise skip to step 3.
*/
{
size_t i;
int inasciirange;
inasciirange = 1;
for (i = 0; utf8in[i]; i++)
if (utf8in[i] & ~0x7F)
inasciirange = 0;
if (inasciirange)
goto step3;
}
/*
* 2. Perform the steps specified in [NAMEPREP] and fail if there is an
* error. (If step 3 of ToASCII is also performed here, it will not
* affect the overall behavior of ToUnicode, but it is not
* necessary.) The AllowUnassigned flag is used in [NAMEPREP].
*/
do
{
char *newp = realloc (utf8in, utf8len + addlen);
if (newp == NULL)
{
free (utf8in);
return IDNA_MALLOC_ERROR;
}
utf8in = newp;
if (flags & IDNA_ALLOW_UNASSIGNED)
rc = stringprep_nameprep (utf8in, utf8len + addlen);
else
rc = stringprep_nameprep_no_unassigned (utf8in, utf8len + addlen);
addlen += 1;
}
while (rc == STRINGPREP_TOO_SMALL_BUFFER);
if (rc != STRINGPREP_OK)
{
free (utf8in);
return IDNA_STRINGPREP_ERROR;
}
/* 3. Verify that the sequence begins with the ACE prefix, and save a
* copy of the sequence.
* ... The ToASCII and ToUnicode operations MUST recognize the ACE
prefix in a case-insensitive manner.
*/
step3:
if (c_strncasecmp (utf8in, IDNA_ACE_PREFIX, strlen (IDNA_ACE_PREFIX)) != 0)
{
free (utf8in);
return IDNA_NO_ACE_PREFIX;
}
/* 4. Remove the ACE prefix.
*/
memmove (utf8in, &utf8in[strlen (IDNA_ACE_PREFIX)],
strlen (utf8in) - strlen (IDNA_ACE_PREFIX) + 1);
/* 5. Decode the sequence using the decoding algorithm in [PUNYCODE]
* and fail if there is an error. Save a copy of the result of
* this step.
*/
(*outlen)--; /* reserve one for the zero */
rc = punycode_decode (strlen (utf8in), utf8in, outlen, out, NULL);
if (rc != PUNYCODE_SUCCESS)
{
free (utf8in);
return IDNA_PUNYCODE_ERROR;
}
out[*outlen] = 0; /* add zero */
/* 6. Apply ToASCII.
*/
rc = idna_to_ascii_4i (out, *outlen, tmpout, flags);
if (rc != IDNA_SUCCESS)
{
free (utf8in);
return rc;
}
/* 7. Verify that the result of step 6 matches the saved copy from
* step 3, using a case-insensitive ASCII comparison.
*/
if (c_strcasecmp (utf8in, tmpout + strlen (IDNA_ACE_PREFIX)) != 0)
{
free (utf8in);
return IDNA_ROUNDTRIP_VERIFY_ERROR;
}
/* 8. Return the saved copy from step 5.
*/
free (utf8in);
return IDNA_SUCCESS;
}
char *
str_iconv (const char *src, const char *from_codeset, const char *to_codeset)
{
if (*src == '\0' || c_strcasecmp (from_codeset, to_codeset) == 0)
{
char *result = strdup (src);
if (result == NULL)
errno = ENOMEM;
return result;
}
else
{
#if HAVE_ICONV
iconv_t cd;
char *result;
/* Avoid glibc-2.1 bug with EUC-KR. */
# if (__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) && !defined _LIBICONV_VERSION
if (c_strcasecmp (from_codeset, "EUC-KR") == 0
|| c_strcasecmp (to_codeset, "EUC-KR") == 0)
{
errno = EINVAL;
return NULL;
}
# endif
cd = iconv_open (to_codeset, from_codeset);
if (cd == (iconv_t) -1)
return NULL;
result = str_cd_iconv (src, cd);
if (result == NULL)
{
/* Close cd, but preserve the errno from str_cd_iconv. */
int saved_errno = errno;
iconv_close (cd);
errno = saved_errno;
}
else
{
if (iconv_close (cd) < 0)
{
/* Return NULL, but free the allocated memory, and while doing
that, preserve the errno from iconv_close. */
int saved_errno = errno;
free (result);
errno = saved_errno;
return NULL;
}
}
return result;
#else
/* This is a different error code than if iconv_open existed but didn't
support from_codeset and to_codeset, so that the caller can emit
an error message such as
"iconv() is not supported. Installing GNU libiconv and
then reinstalling this package would fix this." */
errno = ENOSYS;
return NULL;
#endif
}
}
static int
cmp_names(const void *item, void *data)
{
return c_strcasecmp(item, data);
}
const char *
proper_name_utf8 (const char *name_ascii, const char *name_utf8)
{
/* See whether there is a translation. */
const char *translation = gettext (name_ascii);
/* Try to convert NAME_UTF8 to the locale encoding. */
const char *locale_code = locale_charset ();
char *alloc_name_converted = NULL;
char *alloc_name_converted_translit = NULL;
const char *name_converted = NULL;
const char *name_converted_translit = NULL;
const char *name;
if (c_strcasecmp (locale_code, "UTF-8") != 0)
{
#if HAVE_ICONV
name_converted = alloc_name_converted =
xstr_iconv (name_utf8, "UTF-8", locale_code);
# if (__GLIBC__ == 2 && __GLIBC_MINOR__ >= 2) || __GLIBC__ > 2 \
|| _LIBICONV_VERSION >= 0x0105
{
size_t len = strlen (locale_code);
char *locale_code_translit = (char *) xmalloc (len + 10 + 1);
memcpy (locale_code_translit, locale_code, len);
memcpy (locale_code_translit + len, "//TRANSLIT", 10 + 1);
name_converted_translit = alloc_name_converted_translit =
xstr_iconv (name_utf8, "UTF-8", locale_code_translit);
free (locale_code_translit);
}
# endif
#endif
}
else
{
name_converted = name_utf8;
name_converted_translit = name_utf8;
}
/* The name in locale encoding. */
name = (name_converted != NULL ? name_converted :
name_converted_translit != NULL ? name_converted_translit :
name_ascii);
if (translation != name_ascii)
{
/* See whether the translation contains the original name.
A multibyte-aware strstr() is not absolutely necessary here. */
if (c_strstr (translation, name_ascii) != NULL
|| (name_converted != NULL
&& strstr (translation, name_converted) != NULL)
|| (name_converted_translit != NULL
&& strstr (translation, name_converted_translit) != NULL))
{
if (alloc_name_converted != NULL)
free (alloc_name_converted);
if (alloc_name_converted_translit != NULL)
free (alloc_name_converted_translit);
return translation;
}
else
{
/* Return "TRANSLATION (NAME)". */
char *result =
(char *) xmalloc (strlen (translation) + 2 + strlen (name) + 1 + 1);
sprintf (result, "%s (%s)", translation, name);
if (alloc_name_converted != NULL)
free (alloc_name_converted);
if (alloc_name_converted_translit != NULL)
free (alloc_name_converted_translit);
return result;
}
}
else
{
if (alloc_name_converted != NULL && alloc_name_converted != name)
free (alloc_name_converted);
if (alloc_name_converted_translit != NULL
&& alloc_name_converted_translit != name)
free (alloc_name_converted_translit);
return name;
}
}
static bool
is_encoding_utf32 (const char *encoding)
{
return (!c_strcasecmp (encoding, "utf-32")
|| !c_strcasecmp (encoding, "utf32"));
}
int
iconveh_open (const char *to_codeset, const char *from_codeset, iconveh_t *cdp)
{
iconv_t cd;
iconv_t cd1;
iconv_t cd2;
/* Avoid glibc-2.1 bug with EUC-KR. */
# if ((__GLIBC__ == 2 && __GLIBC_MINOR__ <= 1) && !defined __UCLIBC__) \
&& !defined _LIBICONV_VERSION
if (c_strcasecmp (from_codeset, "EUC-KR") == 0
|| c_strcasecmp (to_codeset, "EUC-KR") == 0)
{
errno = EINVAL;
return -1;
}
# endif
cd = iconv_open (to_codeset, from_codeset);
if (STRCASEEQ (from_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0))
cd1 = (iconv_t)(-1);
else
{
cd1 = iconv_open ("UTF-8", from_codeset);
if (cd1 == (iconv_t)(-1))
{
int saved_errno = errno;
if (cd != (iconv_t)(-1))
iconv_close (cdp->cd);
errno = saved_errno;
return -1;
}
}
if (STRCASEEQ (to_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0)
# if (((__GLIBC__ == 2 && __GLIBC_MINOR__ >= 2) || __GLIBC__ > 2) \
&& !defined __UCLIBC__) \
|| _LIBICONV_VERSION >= 0x0105
|| c_strcasecmp (to_codeset, "UTF-8//TRANSLIT") == 0
# endif
)
cd2 = (iconv_t)(-1);
else
{
cd2 = iconv_open (to_codeset, "UTF-8");
if (cd2 == (iconv_t)(-1))
{
int saved_errno = errno;
if (cd1 != (iconv_t)(-1))
iconv_close (cd1);
if (cd != (iconv_t)(-1))
iconv_close (cd);
errno = saved_errno;
return -1;
}
}
cdp->cd = cd;
cdp->cd1 = cd1;
cdp->cd2 = cd2;
return 0;
}
/* Parse all the aggregate functions. */
static bool
parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict,
struct agr_proc *agr)
{
struct agr_var *tail; /* Tail of linked list starting at agr->vars. */
/* Parse everything. */
tail = NULL;
for (;;)
{
char **dest;
char **dest_label;
size_t n_dest;
struct string function_name;
enum mv_class exclude;
const struct agr_func *function;
int func_index;
union agr_argument arg[2];
const struct variable **src;
size_t n_src;
size_t i;
dest = NULL;
dest_label = NULL;
n_dest = 0;
src = NULL;
function = NULL;
n_src = 0;
arg[0].c = NULL;
arg[1].c = NULL;
ds_init_empty (&function_name);
/* Parse the list of target variables. */
while (!lex_match (lexer, T_EQUALS))
{
size_t n_dest_prev = n_dest;
if (!parse_DATA_LIST_vars (lexer, dict, &dest, &n_dest,
(PV_APPEND | PV_SINGLE | PV_NO_SCRATCH
| PV_NO_DUPLICATE)))
goto error;
/* Assign empty labels. */
{
int j;
dest_label = xnrealloc (dest_label, n_dest, sizeof *dest_label);
for (j = n_dest_prev; j < n_dest; j++)
dest_label[j] = NULL;
}
if (lex_is_string (lexer))
{
dest_label[n_dest - 1] = xstrdup (lex_tokcstr (lexer));
lex_get (lexer);
}
}
/* Get the name of the aggregation function. */
if (lex_token (lexer) != T_ID)
{
lex_error (lexer, _("expecting aggregation function"));
goto error;
}
ds_assign_substring (&function_name, lex_tokss (lexer));
exclude = ds_chomp_byte (&function_name, '.') ? MV_SYSTEM : MV_ANY;
for (function = agr_func_tab; function->name; function++)
if (!c_strcasecmp (function->name, ds_cstr (&function_name)))
break;
if (NULL == function->name)
{
msg (SE, _("Unknown aggregation function %s."),
ds_cstr (&function_name));
goto error;
}
ds_destroy (&function_name);
func_index = function - agr_func_tab;
lex_get (lexer);
/* Check for leading lparen. */
if (!lex_match (lexer, T_LPAREN))
{
if (function->src_vars == AGR_SV_YES)
{
lex_force_match (lexer, T_LPAREN);
goto error;
}
}
else
{
/* Parse list of source variables. */
{
int pv_opts = PV_NO_SCRATCH;
if (func_index == SUM || func_index == MEAN || func_index == SD)
pv_opts |= PV_NUMERIC;
else if (function->n_args)
pv_opts |= PV_SAME_TYPE;
if (!parse_variables_const (lexer, dict, &src, &n_src, pv_opts))
goto error;
}
/* Parse function arguments, for those functions that
require arguments. */
if (function->n_args != 0)
for (i = 0; i < function->n_args; i++)
{
int type;
lex_match (lexer, T_COMMA);
if (lex_is_string (lexer))
{
arg[i].c = recode_string (dict_get_encoding (agr->dict),
"UTF-8", lex_tokcstr (lexer),
-1);
type = VAL_STRING;
}
else if (lex_is_number (lexer))
{
arg[i].f = lex_tokval (lexer);
type = VAL_NUMERIC;
}
else
{
msg (SE, _("Missing argument %zu to %s."),
i + 1, function->name);
goto error;
}
lex_get (lexer);
if (type != var_get_type (src[0]))
{
msg (SE, _("Arguments to %s must be of same type as "
"source variables."),
function->name);
goto error;
}
}
/* Trailing rparen. */
if (!lex_force_match (lexer, T_RPAREN))
goto error;
/* Now check that the number of source variables match
the number of target variables. If we check earlier
than this, the user can get very misleading error
message, i.e. `AGGREGATE x=SUM(y t).' will get this
error message when a proper message would be more
like `unknown variable t'. */
if (n_src != n_dest)
{
msg (SE, _("Number of source variables (%zu) does not match "
"number of target variables (%zu)."),
n_src, n_dest);
goto error;
}
if ((func_index == PIN || func_index == POUT
|| func_index == FIN || func_index == FOUT)
&& (var_is_numeric (src[0])
? arg[0].f > arg[1].f
: str_compare_rpad (arg[0].c, arg[1].c) > 0))
{
union agr_argument t = arg[0];
arg[0] = arg[1];
arg[1] = t;
msg (SW, _("The value arguments passed to the %s function "
"are out-of-order. They will be treated as if "
"they had been specified in the correct order."),
function->name);
}
}
/* Finally add these to the linked list of aggregation
variables. */
for (i = 0; i < n_dest; i++)
{
struct agr_var *v = xzalloc (sizeof *v);
/* Add variable to chain. */
if (agr->agr_vars != NULL)
tail->next = v;
else
agr->agr_vars = v;
tail = v;
tail->next = NULL;
v->moments = NULL;
/* Create the target variable in the aggregate
dictionary. */
{
struct variable *destvar;
v->function = func_index;
if (src)
{
v->src = src[i];
if (var_is_alpha (src[i]))
{
v->function |= FSTRING;
v->string = xmalloc (var_get_width (src[i]));
}
if (function->alpha_type == VAL_STRING)
destvar = dict_clone_var_as (agr->dict, v->src, dest[i]);
else
{
assert (var_is_numeric (v->src)
|| function->alpha_type == VAL_NUMERIC);
destvar = dict_create_var (agr->dict, dest[i], 0);
if (destvar != NULL)
{
struct fmt_spec f;
if ((func_index == N || func_index == NMISS)
&& dict_get_weight (dict) != NULL)
f = fmt_for_output (FMT_F, 8, 2);
else
f = function->format;
var_set_both_formats (destvar, &f);
}
}
} else {
struct fmt_spec f;
v->src = NULL;
destvar = dict_create_var (agr->dict, dest[i], 0);
if (destvar != NULL)
{
if ((func_index == N || func_index == NMISS)
&& dict_get_weight (dict) != NULL)
f = fmt_for_output (FMT_F, 8, 2);
else
f = function->format;
var_set_both_formats (destvar, &f);
}
}
if (!destvar)
{
msg (SE, _("Variable name %s is not unique within the "
"aggregate file dictionary, which contains "
"the aggregate variables and the break "
"variables."),
dest[i]);
goto error;
}
free (dest[i]);
if (dest_label[i])
var_set_label (destvar, dest_label[i]);
v->dest = destvar;
}
v->exclude = exclude;
if (v->src != NULL)
{
int j;
if (var_is_numeric (v->src))
for (j = 0; j < function->n_args; j++)
v->arg[j].f = arg[j].f;
else
for (j = 0; j < function->n_args; j++)
v->arg[j].c = xstrdup (arg[j].c);
}
}
if (src != NULL && var_is_alpha (src[0]))
for (i = 0; i < function->n_args; i++)
{
free (arg[i].c);
arg[i].c = NULL;
}
free (src);
free (dest);
free (dest_label);
if (!lex_match (lexer, T_SLASH))
{
if (lex_token (lexer) == T_ENDCMD)
return true;
lex_error (lexer, "expecting end of command");
return false;
}
continue;
error:
ds_destroy (&function_name);
for (i = 0; i < n_dest; i++)
{
free (dest[i]);
free (dest_label[i]);
}
free (dest);
free (dest_label);
free (arg[0].c);
free (arg[1].c);
if (src && n_src && var_is_alpha (src[0]))
for (i = 0; i < function->n_args; i++)
{
free (arg[i].c);
arg[i].c = NULL;
}
free (src);
return false;
}
}
void
html_input(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
unsigned char *al;
struct form_control *fc;
int cp = html_context->doc_cp;
fc = init_form_control(FC_TEXT, a, html_context);
if (!fc) return;
al = get_attr_val(a, (unsigned char *)"type", cp);
if (al) {
if (!c_strcasecmp((const char *)al, "text")) fc->type = FC_TEXT;
else if (!c_strcasecmp((const char *)al, "hidden")) fc->type = FC_HIDDEN;
else if (!c_strcasecmp((const char *)al, "button")) fc->type = FC_BUTTON;
else if (!c_strcasecmp((const char *)al, "checkbox")) fc->type = FC_CHECKBOX;
else if (!c_strcasecmp((const char *)al, "radio")) fc->type = FC_RADIO;
else if (!c_strcasecmp((const char *)al, "password")) fc->type = FC_PASSWORD;
else if (!c_strcasecmp((const char *)al, "submit")) fc->type = FC_SUBMIT;
else if (!c_strcasecmp((const char *)al, "reset")) fc->type = FC_RESET;
else if (!c_strcasecmp((const char *)al, "file")) fc->type = FC_FILE;
else if (!c_strcasecmp((const char *)al, "image")) fc->type = FC_IMAGE;
/* else unknown type, let it default to FC_TEXT. */
mem_free(al);
}
if (fc->type == FC_HIDDEN)
fc->default_value = get_lit_attr_val(a, (unsigned char *)"value", cp);
else if (fc->type != FC_FILE)
fc->default_value = get_attr_val(a, (unsigned char *)"value", cp);
if (!fc->default_value) {
if (fc->type == FC_CHECKBOX)
fc->default_value = stracpy((const unsigned char *)"on");
else if (fc->type == FC_SUBMIT)
fc->default_value = stracpy((const unsigned char *)"Submit");
else if (fc->type == FC_RESET)
fc->default_value = stracpy((const unsigned char *)"Reset");
else if (fc->type == FC_BUTTON)
fc->default_value = stracpy((const unsigned char *)"Button");
}
if (!fc->default_value)
fc->default_value = stracpy((const unsigned char *)"");
fc->id = get_attr_val(a, (unsigned char *)"id", cp);
fc->name = get_attr_val(a, (unsigned char *)"name", cp);
fc->size = get_num(a, (unsigned char *)"size", cp);
if (fc->size == -1)
fc->size = html_context->options->default_form_input_size;
fc->size++;
if (fc->size > html_context->options->box.width)
fc->size = html_context->options->box.width;
fc->maxlength = get_num(a, (unsigned char *)"maxlength", cp);
if (fc->maxlength == -1) fc->maxlength = INT_MAX;
if (fc->type == FC_CHECKBOX || fc->type == FC_RADIO)
fc->default_state = has_attr(a, (unsigned char *)"checked", cp);
if (fc->type == FC_IMAGE)
fc->alt = get_attr_val(a, (unsigned char *)"alt", cp);
if (fc->type != FC_HIDDEN) {
html_input_format(html_context, a, fc);
}
html_context->special_f(html_context, SP_CONTROL, fc);
}
static int
alias_compare(const struct alias_map *map1, const struct alias_map *map2)
{
return c_strcasecmp(map1->alias, map2->alias);
}
void
html_textarea(struct html_context *html_context, unsigned char *attr,
unsigned char *html, unsigned char *eof, unsigned char **end)
{
struct form_control *fc;
unsigned char *p, *t_name, *wrap_attr;
int t_namelen;
int cols, rows;
int i;
html_focusable(html_context, attr);
while (html < eof && (*html == '\n' || *html == '\r')) html++;
p = html;
while (p < eof && *p != '<') {
pp:
p++;
}
if (p >= eof) {
*end = eof;
return;
}
if (parse_element(p, eof, &t_name, &t_namelen, NULL, end)) goto pp;
if (c_strlcasecmp(t_name, t_namelen, (const unsigned char *)"/TEXTAREA", 9)) goto pp;
fc = init_form_control(FC_TEXTAREA, attr, html_context);
if (!fc) return;
fc->id = get_attr_val(attr, (unsigned char *)"id", html_context->doc_cp);
fc->name = get_attr_val(attr, (unsigned char *)"name", html_context->doc_cp);
fc->default_value = convert_string(NULL, html, p - html,
html_context->doc_cp,
CSM_DEFAULT, NULL, NULL, NULL);
for (p = fc->default_value; p && p[0]; p++) {
/* FIXME: We don't cope well with entities here. Bugzilla uses
* inside of textarea and we fail miserably upon that
* one. --pasky */
if (p[0] == '\r') {
if (p[1] == '\n'
|| (p > fc->default_value && p[-1] == '\n')) {
memmove(p, p + 1, strlen((const char *)p));
p--;
} else {
p[0] = '\n';
}
}
}
cols = get_num(attr, (unsigned char *)"cols", html_context->doc_cp);
if (cols <= 0)
cols = html_context->options->default_form_input_size;
cols++; /* Add 1 column, other browsers may have different
behavior here (mozilla adds 2) --Zas */
if (cols > html_context->options->box.width)
cols = html_context->options->box.width;
fc->cols = cols;
rows = get_num(attr, (unsigned char *)"rows", html_context->doc_cp);
if (rows <= 0) rows = 1;
if (rows > html_context->options->box.height)
rows = html_context->options->box.height;
fc->rows = rows;
html_context->options->needs_height = 1;
wrap_attr = get_attr_val(attr, (unsigned char *)"wrap", html_context->doc_cp);
if (wrap_attr) {
if (!c_strcasecmp((const char *)wrap_attr, "hard")
|| !c_strcasecmp((const char *)wrap_attr, "physical")) {
fc->wrap = FORM_WRAP_HARD;
} else if (!c_strcasecmp((const char *)wrap_attr, "soft")
|| !c_strcasecmp((const char *)wrap_attr, "virtual")) {
fc->wrap = FORM_WRAP_SOFT;
} else if (!c_strcasecmp((const char *)wrap_attr, "none")
|| !c_strcasecmp((const char *)wrap_attr, "off")) {
fc->wrap = FORM_WRAP_NONE;
}
mem_free(wrap_attr);
} else if (has_attr(attr, (unsigned char *)"nowrap", html_context->doc_cp)) {
fc->wrap = FORM_WRAP_NONE;
} else {
fc->wrap = FORM_WRAP_SOFT;
}
fc->maxlength = get_num(attr, (unsigned char *)"maxlength", html_context->doc_cp);
if (fc->maxlength == -1) fc->maxlength = INT_MAX;
if (rows > 1) ln_break(html_context, 1);
else put_chrs(html_context, (unsigned char *)" ", 1);
html_stack_dup(html_context, ELEMENT_KILLABLE);
format.form = fc;
format.style.attr |= AT_BOLD;
for (i = 0; i < rows; i++) {
int j;
for (j = 0; j < cols; j++)
put_chrs(html_context, (unsigned char *)"_", 1);
if (i < rows - 1)
ln_break(html_context, 1);
}
pop_html_element(html_context);
if (rows > 1)
ln_break(html_context, 1);
else
put_chrs(html_context, (unsigned char *)" ", 1);
html_context->special_f(html_context, SP_CONTROL, fc);
}
void
html_form(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
unsigned char *al;
struct form *form;
html_context->was_br = 1;
form = init_form();
if (!form) return;
form->method = FORM_METHOD_GET;
form->form_num = a - html_context->startf;
al = get_attr_val(a, (unsigned char *)"method", html_context->doc_cp);
if (al) {
if (!c_strcasecmp((const char *)al, "post")) {
unsigned char *enctype;
enctype = get_attr_val(a, (unsigned char *)"enctype",
html_context->doc_cp);
form->method = FORM_METHOD_POST;
if (enctype) {
if (!c_strcasecmp((const char *)enctype, "multipart/form-data"))
form->method = FORM_METHOD_POST_MP;
else if (!c_strcasecmp((const char *)enctype, "text/plain"))
form->method = FORM_METHOD_POST_TEXT_PLAIN;
mem_free(enctype);
}
}
mem_free(al);
}
form->onsubmit = get_attr_val(a, (unsigned char *)"onsubmit", html_context->doc_cp);
al = get_attr_val(a, (unsigned char *)"name", html_context->doc_cp);
if (al) form->name = al;
al = get_attr_val(a, (unsigned char *)"action", html_context->doc_cp);
/* The HTML specification at
* http://www.w3.org/TR/REC-html40/interact/forms.html#h-17.3 states
* that the behavior of an empty action attribute should be undefined.
* Mozilla handles action="" as action="<current-URI>" which seems
* reasonable. (bug 615) */
if (al && *al) {
form->action = join_urls(html_context->base_href, trim_chars(al, ' ', NULL));
mem_free(al);
} else {
enum uri_component components = URI_ORIGINAL;
mem_free_if(al);
/* We have to do following for GET method, because we would end
* up with two '?' otherwise. */
if (form->method == FORM_METHOD_GET)
components = URI_FORM_GET;
form->action = get_uri_string(html_context->base_href, components);
/* No action URI should contain post data */
assert(!form->action || !strchr((char *)form->action, POST_CHAR));
/* GET method URIs should not have '?'. */
assert(!form->action
|| form->method != FORM_METHOD_GET
|| !strchr((char *)form->action, '?'));
}
al = get_target(html_context->options, a);
form->target = al ? al : stracpy(html_context->base_target);
html_context->special_f(html_context, SP_FORM, form);
}
int
main (int argc, char *argv[])
{
if (argc > 1)
{
/* configure should already have checked that the locale is supported. */
if (setlocale (LC_ALL, "") == NULL)
return 1;
}
ASSERT (c_strcasecmp ("paragraph", "Paragraph") == 0);
ASSERT (c_strcasecmp ("paragrapH", "parAgRaph") == 0);
ASSERT (c_strcasecmp ("paragraph", "paraLyzed") < 0);
ASSERT (c_strcasecmp ("paraLyzed", "paragraph") > 0);
ASSERT (c_strcasecmp ("para", "paragraph") < 0);
ASSERT (c_strcasecmp ("paragraph", "para") > 0);
/* The following tests shows how c_strcasecmp() is different from
strcasecmp(). */
ASSERT (c_strcasecmp ("\311mile", "\351mile") < 0);
ASSERT (c_strcasecmp ("\351mile", "\311mile") > 0);
/* The following tests shows how c_strcasecmp() is different from
mbscasecmp(). */
ASSERT (c_strcasecmp ("\303\266zg\303\274r", "\303\226ZG\303\234R") > 0); /* özgür */
ASSERT (c_strcasecmp ("\303\226ZG\303\234R", "\303\266zg\303\274r") < 0); /* özgür */
/* This test shows how strings of different size cannot compare equal. */
ASSERT (c_strcasecmp ("turkish", "TURK\304\260SH") < 0);
ASSERT (c_strcasecmp ("TURK\304\260SH", "turkish") > 0);
return 0;
}
void
html_link(struct html_context *html_context, unsigned char *a,
unsigned char *xxx3, unsigned char *xxx4, unsigned char **xxx5)
{
int link_display = html_context->options->meta_link_display;
unsigned char *name;
struct hlink link;
struct string text;
int name_neq_title = 0;
int first = 1;
#ifndef CONFIG_CSS
if (!link_display) return;
#endif
if (!html_link_parse(html_context, a, &link)) return;
if (!link.href) goto free_and_return;
#ifdef CONFIG_CSS
if (link.type == LT_STYLESHEET
&& supports_html_media_attr(link.media)) {
int len = strlen((const char *)link.href);
import_css_stylesheet(&html_context->css_styles,
html_context->base_href, link.href, len);
}
if (!link_display) goto free_and_return;
#endif
/* Ignore few annoying links.. */
if (link_display < 5 &&
(link.type == LT_ICON ||
link.type == LT_AUTHOR ||
link.type == LT_STYLESHEET ||
link.type == LT_ALTERNATE_STYLESHEET)) goto free_and_return;
if (!link.name || link.type != LT_UNKNOWN)
/* Give preference to our default names for known types. */
name = get_lt_default_name(&link);
else
name = link.name;
if (!name) goto free_and_return;
if (!init_string(&text)) goto free_and_return;
html_focusable(html_context, a);
if (link.title) {
add_to_string(&text, link.title);
name_neq_title = strcmp((const char *)link.title, (const char *)name);
} else
add_to_string(&text, name);
if (link_display == 1) goto put_link_line; /* Only title */
#define APPEND(what) do { \
add_to_string(&text, first ? (const unsigned char *)" (" : (const unsigned char *)", "); \
add_to_string(&text, (what)); \
first = 0; \
} while (0)
if (name_neq_title) {
APPEND(name);
}
if (link_display >= 3 && link.hreflang) {
APPEND(link.hreflang);
}
if (link_display >= 4 && link.content_type) {
APPEND(link.content_type);
}
if (link.lang && link.type == LT_ALTERNATE_LANG &&
(link_display < 3 || (link.hreflang &&
c_strcasecmp((const char *)link.hreflang, (const char *)link.lang)))) {
APPEND(link.lang);
}
if (link.media) {
APPEND(link.media);
}
#undef APPEND
if (!first) add_char_to_string(&text, ')');
put_link_line:
{
unsigned char *prefix = (link.direction == LD_REL)
? (unsigned char *)"Link: " : (unsigned char *)"Reverse link: ";
unsigned char *link_name = (text.length)
? text.source : name;
put_link_line(prefix, link_name, link.href,
html_context->base_target, html_context);
if (text.source) done_string(&text);
}
free_and_return:
html_link_clear(&link);
}
iconv_t
rpl_iconv_open (const char *tocode, const char *fromcode)
#undef iconv_open
{
char fromcode_upper[32];
char tocode_upper[32];
char *fromcode_upper_end;
char *tocode_upper_end;
#if REPLACE_ICONV_UTF
/* Special handling of conversion between UTF-8 and UTF-{16,32}{BE,LE}.
Do this here, before calling the real iconv_open(), because OSF/1 5.1
iconv() to these encoding inserts a BOM, which is wrong.
We do not need to handle conversion between arbitrary encodings and
UTF-{16,32}{BE,LE}, because the 'striconveh' module implements two-step
conversion through UTF-8.
The _ICONV_* constants are chosen to be disjoint from any iconv_t
returned by the system's iconv_open() functions. Recall that iconv_t
is a scalar type. */
if (c_toupper (fromcode[0]) == 'U'
&& c_toupper (fromcode[1]) == 'T'
&& c_toupper (fromcode[2]) == 'F'
&& fromcode[3] == '-')
{
if (c_toupper (tocode[0]) == 'U'
&& c_toupper (tocode[1]) == 'T'
&& c_toupper (tocode[2]) == 'F'
&& tocode[3] == '-')
{
if (strcmp (fromcode + 4, "8") == 0)
{
if (c_strcasecmp (tocode + 4, "16BE") == 0)
return _ICONV_UTF8_UTF16BE;
if (c_strcasecmp (tocode + 4, "16LE") == 0)
return _ICONV_UTF8_UTF16LE;
if (c_strcasecmp (tocode + 4, "32BE") == 0)
return _ICONV_UTF8_UTF32BE;
if (c_strcasecmp (tocode + 4, "32LE") == 0)
return _ICONV_UTF8_UTF32LE;
}
else if (strcmp (tocode + 4, "8") == 0)
{
if (c_strcasecmp (fromcode + 4, "16BE") == 0)
return _ICONV_UTF16BE_UTF8;
if (c_strcasecmp (fromcode + 4, "16LE") == 0)
return _ICONV_UTF16LE_UTF8;
if (c_strcasecmp (fromcode + 4, "32BE") == 0)
return _ICONV_UTF32BE_UTF8;
if (c_strcasecmp (fromcode + 4, "32LE") == 0)
return _ICONV_UTF32LE_UTF8;
}
}
}
#endif
/* Do *not* add special support for 8-bit encodings like ASCII or ISO-8859-1
here. This would lead to programs that work in some locales (such as the
"C" or "en_US" locales) but do not work in East Asian locales. It is
better if programmers make their programs depend on GNU libiconv (except
on glibc systems), e.g. by using the AM_ICONV macro and documenting the
dependency in an INSTALL or DEPENDENCIES file. */
/* Try with the original names first.
This covers the case when fromcode or tocode is a lowercase encoding name
that is understood by the system's iconv_open but not listed in our
mappings table. */
{
iconv_t cd = iconv_open (tocode, fromcode);
if (cd != (iconv_t)(-1))
return cd;
}
/* Convert the encodings to upper case, because
1. in the arguments of iconv_open() on AIX, HP-UX, and OSF/1 the case
matters,
2. it makes searching in the table faster. */
{
const char *p = fromcode;
char *q = fromcode_upper;
while ((*q = c_toupper (*p)) != '\0')
{
p++;
q++;
if (q == &fromcode_upper[SIZEOF (fromcode_upper)])
{
errno = EINVAL;
return (iconv_t)(-1);
}
}
fromcode_upper_end = q;
}
{
const char *p = tocode;
char *q = tocode_upper;
while ((*q = c_toupper (*p)) != '\0')
{
p++;
q++;
if (q == &tocode_upper[SIZEOF (tocode_upper)])
{
errno = EINVAL;
return (iconv_t)(-1);
}
}
tocode_upper_end = q;
}
#ifdef ICONV_FLAVOR
/* Apply the mappings. */
{
const struct mapping *m =
mapping_lookup (fromcode_upper, fromcode_upper_end - fromcode_upper);
fromcode = (m != NULL ? m->vendor_name : fromcode_upper);
}
{
const struct mapping *m =
mapping_lookup (tocode_upper, tocode_upper_end - tocode_upper);
tocode = (m != NULL ? m->vendor_name : tocode_upper);
}
#else
fromcode = fromcode_upper;
tocode = tocode_upper;
#endif
return iconv_open (tocode, fromcode);
}
void
html_script(struct html_context *html_context, unsigned char *a,
unsigned char *html, unsigned char *eof, unsigned char **end)
{
#ifdef CONFIG_ECMASCRIPT
/* TODO: <noscript> processing. Well, same considerations apply as to
* CSS property display: none processing. */
/* TODO: Charsets for external scripts. */
unsigned char *type, *language, *src;
int in_comment = 0;
#endif
html_skip(html_context, a);
#ifdef CONFIG_ECMASCRIPT
/* We try to process nested <script> if we didn't process the parent
* one. That's why's all the fuzz. */
/* Ref:
* http://www.ietf.org/internet-drafts/draft-hoehrmann-script-types-03.txt
*/
type = get_attr_val(a, (unsigned char *)"type", html_context->doc_cp);
if (type) {
unsigned char *pos = type;
if (!c_strncasecmp((const char *)type, "text/", 5)) {
pos += 5;
} else if (!c_strncasecmp((const char *)type, "application/", 12)) {
pos += 12;
} else {
mem_free(type);
not_processed:
/* Permit nested scripts and retreat. */
html_top->invisible++;
return;
}
if (!c_strncasecmp((const char *)pos, "javascript", 10)) {
int len = strlen((const char *)pos);
if (len > 10 && !isdigit(pos[10])) {
mem_free(type);
goto not_processed;
}
} else if (c_strcasecmp((const char *)pos, "ecmascript")
&& c_strcasecmp((const char *)pos, "jscript")
&& c_strcasecmp((const char *)pos, "livescript")
&& c_strcasecmp((const char *)pos, "x-javascript")
&& c_strcasecmp((const char *)pos, "x-ecmascript")) {
mem_free(type);
goto not_processed;
}
mem_free(type);
}
/* Check that the script content is ecmascript. The value of the
* language attribute can be JavaScript with optional version digits
* postfixed (like: ``JavaScript1.1'').
* That attribute is deprecated in favor of type by HTML 4.01 */
language = get_attr_val(a, (unsigned char *)"language", html_context->doc_cp);
if (language) {
int languagelen = strlen((const char *)language);
if (languagelen < 10
|| (languagelen > 10 && !isdigit(language[10]))
|| c_strncasecmp((const char *)language, "javascript", 10)) {
mem_free(language);
html_top->invisible++;
return;
//goto not_processed;
}
mem_free(language);
}
if (html_context->part->document
&& (src = get_attr_val(a, (unsigned char *)"src", html_context->doc_cp))) {
/* External reference. */
unsigned char *import_url;
struct uri *uri;
if (!get_opt_bool((const unsigned char *)"ecmascript.enable", NULL)) {
mem_free(src);
html_top->invisible++;
return;
//goto not_processed;
}
/* HTML <head> urls should already be fine but we can.t detect them. */
import_url = join_urls(html_context->base_href, src);
mem_free(src);
if (!import_url) goto imported;
uri = get_uri(import_url, URI_BASE);
if (!uri) goto imported;
/* Request the imported script as part of the document ... */
html_context->special_f(html_context, SP_SCRIPT, uri);
done_uri(uri);
/* Create URL reference onload snippet. */
insert_in_string(&import_url, 0, (const unsigned char *)"^", 1);
add_to_string_list(&html_context->part->document->onload_snippets,
import_url, -1);
imported:
/* Retreat. Do not permit nested scripts, tho'. */
if (import_url) mem_free(import_url);
return;
}
/* Positive, grab the rest and interpret it. */
/* First position to the real script start. */
while (html < eof && *html <= ' ') html++;
if (eof - html > 4 && !strncmp((const char *)html, "<!--", 4)) {
in_comment = 1;
/* We either skip to the end of line or to -->. */
for (; *html != '\n' && *html != '\r' && eof - html >= 3; html++) {
if (!strncmp((const char *)html, "-->", 3)) {
/* This means the document is probably broken.
* We will now try to process the rest of
* <script> contents, which is however likely
* to be empty. Should we try to process the
* comment too? Currently it seems safer but
* less tolerant to broken pages, if there are
* any like this. */
html += 3;
in_comment = 0;
break;
}
}
}
*end = html;
/* Now look ahead for the script end. The <script> contents is raw
* CDATA, so we just look for the ending tag and need not care for
* any quote marks counting etc - YET, we are more tolerant and permit
* </script> stuff inside of the script if the whole <script> element
* contents is wrapped in a comment. See i.e. Mozilla bug 26857 for fun
* reading regarding this. */
for (; *end < eof; (*end)++) {
unsigned char *name;
int namelen;
if (in_comment) {
/* TODO: If we ever get some standards-quirk mode
* distinction, this should be disabled in the
* standards mode (and we should just look for CDATA
* end, which is "</"). --pasky */
if (eof - *end >= 3 && !strncmp((const char *)*end, "-->", 3)) {
/* Next iteration will jump passed the ending '>' */
(*end) += 2;
in_comment = 0;
}
continue;
/* XXX: Scan for another comment? That's admittelly
* already stretching things a little bit to an
* extreme ;-). */
}
if (**end != '<')
continue;
/* We want to land before the closing element, that's why we
* don't pass @end also as the appropriate parse_element()
* argument. */
if (parse_element(*end, eof, &name, &namelen, NULL, NULL))
continue;
if (c_strlcasecmp(name, namelen, (const unsigned char *)"/script", 7))
continue;
/* We have won! */
break;
}
if (*end >= eof) {
/* Either the document is not completely loaded yet or it's
* broken. At any rate, run away screaming. */
*end = eof; /* Just for sanity. */
return;
}
if (html_context->part->document && *html != '^') {
add_to_string_list(&html_context->part->document->onload_snippets,
html, *end - html);
}
#endif
}