nyje
/
medit
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
medit/moo/mooedit/gtksourceview/gtksourcelanguage-parser-2.c

1794 lines
43 KiB
C
Raw Blame History

/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8; coding: utf-8 -*-
* gtksourcelanguage-parser-2.c
* Language specification parser for 2.0 version .lang files
*
* Copyright (C) 2003 - Gustavo Giráldez <gustavo.giraldez@gmx.net>
* Copyright (C) 2005, 2006 - Emanuele Aina, Marco Barisione
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU Library General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#undef ENABLE_DEBUG
#ifdef ENABLE_DEBUG
#define DEBUG(x) x
#else
#define DEBUG(x)
#endif
#include <string.h>
#include <fcntl.h>
#include <libxml/xmlreader.h>
#include <glib/gstdio.h>
#include "gtksourceview-i18n.h"
#include "gtksourcebuffer.h"
#include "gtksourcelanguage.h"
#include "gtksourcelanguagesmanager.h"
#include "gtksourcelanguage-private.h"
#include "gtksourcecontextengine.h"
#include <eggregex.h>
#define PARSER_ERROR parser_error_quark ()
typedef enum {
PARSER_ERROR_CANNOT_OPEN = 0,
PARSER_ERROR_CANNOT_VALIDATE,
PARSER_ERROR_INVALID_DOC,
PARSER_ERROR_WRONG_VERSION,
PARSER_ERROR_WRONG_ID,
PARSER_ERROR_MALFORMED_REGEX,
PARSER_ERROR_MALFORMED_MAP_TO
} ParserError;
struct _ParserState
{
/* The args passed to _file_parse_version2() */
xmlTextReader *reader;
GtkSourceLanguage *language;
GtkSourceContextEngine *engine;
/* A stack of id that representing parent contexts */
GQueue *curr_parents;
/* The id of the current language (used to decorate ids) */
gchar *current_lang_id;
/* An hash table with the defined regex as strings used to
* resolve references (the key is the id) */
GHashTable *defined_regexes;
/* The mapping between style ids and their default styles */
GHashTable *styles_mapping;
/* The list of loaded languages (the item are XmlChar pointers) */
GSList **loaded_lang_ids;
/* A serial number incremented to get unique generated names */
guint id_cookie;
/* The default flags used by the regexes */
EggRegexCompileFlags regex_compile_flags;
gchar *opening_delimiter;
gchar *closing_delimiter;
GError **error;
};
typedef struct _ParserState ParserState;
static GQuark parser_error_quark (void);
static gboolean str_to_bool (gchar *string);
static gchar *generate_new_id (ParserState *parser_state);
static gboolean id_is_decorated (gchar *id,
gchar **lang_id);
static gchar *decorate_id (ParserState *parser_state,
gchar *id);
static ParserState *parser_state_new (GtkSourceLanguage *language,
GtkSourceContextEngine *engine,
GHashTable *defined_regexes,
GHashTable *styles_mapping,
xmlTextReader *reader,
GSList **loaded_lang_ids);
static void parser_state_destroy (ParserState *parser_state);
static gboolean file_parse (gchar *filename,
GtkSourceLanguage *language,
GtkSourceContextEngine *engine,
GHashTable *defined_regexes,
GHashTable *styles,
GSList **loaded_lang_ids,
GError **error);
static EggRegexCompileFlags
update_regex_flags (EggRegexCompileFlags flags,
const gchar *option_name,
gboolean value);
static gboolean create_definition (ParserState *parser_state,
gchar *id,
gchar *parent_id,
gchar *style,
GError **error);
static void handle_context_element (ParserState *parser_state,
GError **error);
static void handle_language_element (ParserState *parser_state,
GError **error);
static void handle_define_regex_element (ParserState *parser_state,
GError **error);
static void handle_default_regex_options_element
(ParserState *parser_state,
GError **error);
static void element_start (ParserState *parser_state,
GError **error);
static void element_end (ParserState *parser_state);
static gboolean replace_by_id (const EggRegex *egg_regex,
const gchar *regex,
GString *expanded_regex,
gpointer data);
static gchar *expand_regex (ParserState *parser_state,
gchar *regex,
EggRegexCompileFlags flags,
gboolean do_expand_vars,
gboolean insert_parentheses,
GError **error);
static GQuark
parser_error_quark (void)
{
static GQuark err_q = 0;
if (err_q == 0)
err_q = g_quark_from_static_string (
"parser-error-quark");
return err_q;
}
static gboolean
str_to_bool (gchar *string)
{
g_assert (string);
if (g_ascii_strcasecmp ("true", string) == 0)
return TRUE;
else
return FALSE;
}
static gchar *
generate_new_id (ParserState *parser_state)
{
gchar *id;
id = g_strdup_printf ("unnamed-%u", parser_state->id_cookie);
parser_state->id_cookie++;
DEBUG (g_message ("generated id %s", id));
return id;
}
static gboolean
id_is_decorated (gchar *id, gchar **lang_id)
{
/* This function is quite simple because the XML validator check for
* the correctness of the id with a regex */
gchar **tokens;
gboolean is_decorated;
tokens = g_strsplit (id, ":", 2);
g_return_val_if_fail (tokens != NULL, FALSE);
if (tokens [1] == NULL)
{
/* There is no ":" in the id */
is_decorated = FALSE;
}
else if (tokens [1] != NULL && strcmp ("*", tokens[1]) == 0)
{
/* This is an undecorated "import all", and not a decorated
* reference */
is_decorated = FALSE;
}
else
{
is_decorated = TRUE;
if (lang_id != NULL)
{
*lang_id = g_strdup (tokens[0]);
}
}
g_strfreev (tokens);
return is_decorated;
}
static gchar *
decorate_id (ParserState *parser_state,
gchar *id)
{
gchar *decorated_id;
decorated_id = g_strdup_printf ("%s:%s",
parser_state->current_lang_id, id);
DEBUG (g_message ("decorated '%s' to '%s'", id, decorated_id));
return decorated_id;
}
static gboolean
lang_id_is_already_loaded (ParserState *parser_state, gchar *lang_id)
{
GSList *l;
gchar *loaded_lang;
l = *(parser_state->loaded_lang_ids);
g_return_val_if_fail (lang_id != NULL, FALSE);
while (l != NULL)
{
loaded_lang = l->data;
if (strcmp (loaded_lang, lang_id) == 0)
{
return TRUE;
}
l = g_slist_next (l);
}
return FALSE;
}
static gboolean
create_definition (ParserState *parser_state,
gchar *id,
gchar *parent_id,
gchar *style,
GError **error)
{
gchar *match = NULL, *start = NULL, *end = NULL;
gchar *prefix = NULL, *suffix = NULL;
xmlChar *tmp;
gboolean extend_parent = TRUE;
gboolean end_at_line_end = FALSE;
GtkSourceContextMatchOptions options = 0;
xmlNode *context_node, *child;
GString *all_items;
EggRegexCompileFlags flags, match_flags = 0, start_flags = 0, end_flags = 0;
GError *tmp_error = NULL;
g_assert (parser_state->engine != NULL);
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
/* extend-parent */
tmp = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "extend-parent");
if (tmp)
extend_parent = str_to_bool ((gchar *)tmp);
xmlFree (tmp);
/* end-at-line-end */
tmp = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "end-at-line-end");
if (tmp)
end_at_line_end = str_to_bool ((gchar *)tmp);
xmlFree (tmp);
DEBUG (g_message ("creating context %s, child of %s", id, parent_id ? parent_id : "(null)"));
/* Fetch the content of the sublements using the tree API on
* the current node */
context_node = xmlTextReaderExpand (parser_state->reader);
/* The file should be validated so this should not happen */
g_assert (context_node != NULL);
child = context_node->children;
while (child != NULL)
{
xmlAttr *attribute;
if (child->type != XML_READER_TYPE_ELEMENT)
{
child = child->next;
continue;
}
/* FIXME: add PCRE_EXTRA support in EggRegex */
flags = parser_state->regex_compile_flags;
for (attribute = child->properties;
attribute != NULL;
attribute = attribute->next)
{
g_assert (attribute->children);
tmp = attribute->children->content;
if (xmlStrcmp (BAD_CAST "extended", attribute->name) == 0)
{
flags = update_regex_flags (flags,
"extended",
str_to_bool ((gchar *)tmp));
}
else if (xmlStrcmp (BAD_CAST "case-insensitive", attribute->name) == 0)
{
flags = update_regex_flags (flags,
"case-insensitive",
str_to_bool ((gchar *)tmp));
}
else if (xmlStrcmp (BAD_CAST "dot-match-all", attribute->name) == 0)
{
flags = update_regex_flags (flags,
"dot-match-all",
str_to_bool ((gchar *)tmp));
}
}
/*flags = update_regex_flags (parser_state->regex_compile_flags,
"extended", TRUE);*/
g_assert (child->name);
if (xmlStrcmp (BAD_CAST "match", child->name) == 0
&& child->children)
{
/* <match> */
match = g_strdup ((gchar *)child->children->content);
match_flags = flags;
}
else if (xmlStrcmp (BAD_CAST "start", child->name) == 0)
{
/* <start> */
if (child->children)
{
start = g_strdup ((gchar *)child->children->content);
}
else
{
/* If the <start> element is present but
* has no content use an empty string */
start = g_strdup ("");
}
start_flags = flags;
}
else if (xmlStrcmp (BAD_CAST "end", child->name) == 0)
{
/* <end> */
if (child->children)
end = g_strdup ((gchar *)child->children->content);
else
{
/* If the <end> element is present but
* has no content use an empty string */
end = g_strdup ("");
}
end_flags = flags;
}
else if (xmlStrcmp (BAD_CAST "prefix", child->name) == 0)
{
/* <prefix> */
if (child->children != NULL)
prefix = g_strdup ((gchar*) child->children->content);
else
prefix = g_strdup ("");
}
else if (xmlStrcmp (BAD_CAST "suffix", child->name) == 0)
{
/* <suffix> */
if (child->children != NULL)
suffix = g_strdup ((gchar*) child->children->content);
else
suffix = g_strdup ("");
}
else if (xmlStrcmp (BAD_CAST "keyword", child->name) == 0 &&
child->children != NULL)
{
/* <keyword> */
all_items = g_string_new (NULL);
if (prefix != NULL)
g_string_append (all_items, prefix);
else
g_string_append (all_items,
parser_state->opening_delimiter);
g_string_append (all_items, "(");
/* Read every keyword/symbol concatenating them
* in a single string */
while (TRUE)
{
/* TODO: this could be done destructively,
* modifing the string in place without the
* copy */
g_string_append (all_items, (gchar*) child->children->content);
child = child->next;
/* Skip text nodes */
if (child != NULL && !xmlStrcmp (BAD_CAST "text", child->name))
child = child->next;
/* These are the conditions which control
* the loop */
if (child == NULL)
break;
if (xmlStrcmp (BAD_CAST "keyword", child->name) != 0)
break;
if (child->children == NULL)
break;
/* If we get here the next node will be added
* to the keywords/symbols list, so we add the
* separator */
g_string_append (all_items, "|");
}
g_string_append (all_items, ")");
if (suffix != NULL)
g_string_append (all_items, suffix);
else
g_string_append (all_items,
parser_state->closing_delimiter);
match = g_string_free (all_items, FALSE);
match_flags = flags;
}
if (child != NULL)
child = child->next;
}
DEBUG (g_message ("start: '%s'", start ? start : "(null)"));
DEBUG (g_message ("end: '%s'", end ? end : "(null)"));
DEBUG (g_message ("match: '%s'", match ? match : "(null)"));
if (tmp_error == NULL && start != NULL)
{
gchar *tmp = start;
start = expand_regex (parser_state, start, start_flags,
TRUE, FALSE, &tmp_error);
g_free (tmp);
}
if (tmp_error == NULL && end != NULL)
{
gchar *tmp = end;
end = expand_regex (parser_state, end, end_flags,
TRUE, FALSE, &tmp_error);
g_free (tmp);
}
if (tmp_error == NULL && match != NULL)
{
gchar *tmp = match;
match = expand_regex (parser_state, match, match_flags,
TRUE, FALSE, &tmp_error);
g_free (tmp);
}
if (extend_parent)
options |= GTK_SOURCE_CONTEXT_EXTEND_PARENT;
if (end_at_line_end)
options |= GTK_SOURCE_CONTEXT_END_AT_LINE_END;
if (tmp_error == NULL)
_gtk_source_context_engine_define_context (parser_state->engine,
id,
parent_id,
match,
start,
end,
style,
options,
&tmp_error);
g_free (match);
g_free (start);
g_free (end);
g_free (prefix);
g_free (suffix);
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return FALSE;
}
return TRUE;
}
static gboolean
add_ref (ParserState *parser_state, gchar *ref, GError **error)
{
gchar *container_id;
gboolean all = FALSE;
gchar *ref_id;
gchar *lang_id;
GtkSourceLanguagesManager *lm;
GtkSourceLanguage *imported_language;
GError *tmp_error = NULL;
/* Return if an error is already set */
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
if (id_is_decorated (ref, &lang_id))
{
if (!lang_id_is_already_loaded (parser_state, lang_id))
{
lm = _gtk_source_language_get_languages_manager (parser_state->language);
imported_language = gtk_source_languages_manager_get_language_by_id (lm, lang_id);
if (imported_language == NULL)
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_WRONG_ID,
"unable to resolve language '%s' in ref '%s'",
lang_id, ref);
}
else
{
file_parse (imported_language->priv->lang_file_name,
parser_state->language,
parser_state->engine,
parser_state->defined_regexes,
parser_state->styles_mapping,
parser_state->loaded_lang_ids,
&tmp_error);
}
}
ref_id = g_strdup (ref);
}
else
{
ref_id = decorate_id (parser_state, ref);
}
if (tmp_error == NULL && parser_state->engine != NULL)
{
if (g_str_has_suffix (ref, ":*"))
{
all = TRUE;
ref_id [strlen (ref_id) - 2] = '\0';
}
container_id = g_queue_peek_head (parser_state->curr_parents);
/* If the document is validated container_id is never NULL */
g_assert (container_id);
_gtk_source_context_engine_add_ref (parser_state->engine,
container_id,
ref_id,
all,
&tmp_error);
DEBUG (g_message ("appended %s in %s", ref_id, container_id));
}
g_free (ref_id);
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return FALSE;
}
return TRUE;
}
static gboolean
create_sub_pattern (ParserState *parser_state,
gchar *id,
gchar *sub_pattern,
gchar *style,
GError **error)
{
gchar *container_id;
xmlChar *where;
GError *tmp_error = NULL;
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
container_id = g_queue_peek_head (parser_state->curr_parents);
/* If the document is validated container is never NULL */
g_assert (container_id);
where = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "where");
_gtk_source_context_engine_add_sub_pattern (parser_state->engine,
id,
container_id,
sub_pattern,
(gchar*) where,
style,
&tmp_error);
xmlFree (where);
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return FALSE;
}
return TRUE;
}
static void
handle_context_element (ParserState *parser_state,
GError **error)
{
gchar *id, *parent_id, *style_ref;
xmlChar *ref, *sub_pattern, *tmp;
int is_empty;
gboolean success;
GError *tmp_error = NULL;
/* Return if an error is already set */
g_return_if_fail (error == NULL || *error == NULL);
ref = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "ref");
sub_pattern = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "sub-pattern");
tmp = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "style-ref");
if (tmp == NULL || id_is_decorated ((gchar *)tmp, NULL))
style_ref = g_strdup ((gchar *)tmp);
else
style_ref = decorate_id (parser_state, (gchar *)tmp);
xmlFree (tmp);
if (style_ref && !g_hash_table_lookup (parser_state->styles_mapping, style_ref))
{
g_warning ("style '%s' not defined", style_ref);
}
if (ref != NULL)
{
add_ref (parser_state, (gchar *)ref, &tmp_error);
}
else
{
tmp = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "id");
if (tmp == NULL)
tmp = xmlStrdup (BAD_CAST generate_new_id (parser_state));
if (id_is_decorated ((gchar *)tmp, NULL))
id = g_strdup ((gchar *)tmp);
else
id = decorate_id (parser_state, (gchar *)tmp);
xmlFree (tmp);
if (parser_state->engine != NULL)
{
if (sub_pattern != NULL)
{
create_sub_pattern (parser_state, id,
(gchar *)sub_pattern, style_ref,
&tmp_error);
}
else
{
parent_id = g_queue_peek_head (
parser_state->curr_parents);
is_empty = xmlTextReaderIsEmptyElement (
parser_state->reader);
success = create_definition (parser_state, id, parent_id,
style_ref, &tmp_error);
if (success && !is_empty)
{
/* Push the new context in the curr_parents
* stack only if other contexts can be
* defined inside it */
g_queue_push_head (parser_state->curr_parents,
g_strdup (id));
}
}
}
g_free (id);
}
g_free (style_ref);
xmlFree (sub_pattern);
xmlFree (ref);
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
handle_language_element (ParserState *parser_state,
GError **error)
{
/* FIXME: check that the language name, version, etc. are the
* right ones - Paolo */
xmlChar *lang_id, *lang_version;
xmlChar *expected_version = BAD_CAST "2.0";
/* Return if an error is already set */
g_return_if_fail (error == NULL || *error == NULL);
lang_version = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "version");
if (lang_version == NULL ||
(xmlStrcmp (expected_version, lang_version) != 0))
{
g_set_error (error,
PARSER_ERROR,
PARSER_ERROR_WRONG_VERSION,
"wrong language version '%s', expected '%s'",
lang_version, (gchar *)expected_version);
}
lang_id = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "id");
parser_state->current_lang_id = g_strdup ((gchar *)lang_id);
*(parser_state->loaded_lang_ids) = g_slist_prepend (
*(parser_state->loaded_lang_ids), lang_id);
xmlFree (lang_version);
}
static gboolean
replace_by_id (const EggRegex *egg_regex,
const gchar *regex,
GString *expanded_regex,
gpointer data)
{
gchar *id, *subst, *escapes;
gchar *tmp;
GError *tmp_error = NULL;
ParserState *parser_state = data;
escapes = egg_regex_fetch (egg_regex, 1, regex);
tmp = egg_regex_fetch (egg_regex, 2, regex);
g_strstrip (tmp);
if (id_is_decorated (tmp, NULL))
id = g_strdup (tmp);
else
id = decorate_id (parser_state, tmp);
g_free (tmp);
subst = g_hash_table_lookup (parser_state->defined_regexes, id);
if (subst == NULL)
{
g_set_error (&tmp_error,
PARSER_ERROR, PARSER_ERROR_WRONG_ID,
_ ("wrong id '%s' in regex '%s'"), id, regex);
}
if (tmp_error == NULL)
{
g_string_append (expanded_regex, escapes);
g_string_append (expanded_regex, subst);
}
g_free (escapes);
g_free (id);
if (tmp_error != NULL)
{
g_propagate_error (parser_state->error, tmp_error);
return TRUE;
}
return FALSE;
}
static EggRegexCompileFlags
update_regex_flags (EggRegexCompileFlags flags,
const gchar *option_name,
gboolean value)
{
EggRegexCompileFlags single_flag;
DEBUG (g_message ("setting the '%s' regex flag to %d", option_name, value));
if (strcmp ("case-insensitive", option_name) == 0)
{
single_flag = EGG_REGEX_CASELESS;
}
else if (strcmp ("extended", option_name) == 0)
{
single_flag = EGG_REGEX_EXTENDED;
}
else if (strcmp ("dot-match-all", option_name) == 0)
{
single_flag = EGG_REGEX_DOTALL;
}
else
return flags;
if (value)
flags |= single_flag;
else
flags &= ~single_flag;
return flags;
}
static gboolean
calc_regex_flags (const gchar *flags_string, EggRegexCompileFlags *flags)
{
gboolean add = TRUE;
while (*flags_string != '\0')
{
EggRegexCompileFlags single_flag = 0;
switch (g_utf8_get_char (flags_string))
{
case '-':
add = FALSE;
break;
case '+':
add = TRUE;
break;
case 'i':
single_flag = EGG_REGEX_CASELESS;
break;
case 'x':
single_flag = EGG_REGEX_EXTENDED;
break;
case 's':
/* FIXME: do we need this? */
single_flag = EGG_REGEX_DOTALL;
break;
default:
return FALSE;
}
if (single_flag != 0)
{
if (add)
*flags |= single_flag;
else
*flags &= ~single_flag;
}
flags_string = g_utf8_next_char (flags_string);
}
return TRUE;
}
static gchar *
expand_regex_vars (ParserState *parser_state, gchar *regex, gint len, GError **error)
{
/* This is the commented regex without the doubled escape needed
* in a C string:
*
* (?<!\\)(\\\\)*\\%\{([^@]*?)\}
* |------------||---||------||-|
* | | | |
* | the string | the ending
* | "\%{" | bracket "}"
* | |
* a even sequence |
* of escapes or the id of the
* a char that included regex
* is not an (not greedy)
* escape (this
* is not matched)
*
* The first block is needed to ensure that the sequence is
* not escaped.
* Matches with an id containing a "@" sign are ignored and
* passed to the engine because they are references to subpatterns
* in a different regex (i.e. in the start regex while we are in the
* end regex.)
* The sub pattern containing the id is the second.
*/
const gchar *re = "(?<!\\\\)(\\\\\\\\)*\\\\%\\{([^@]*?)\\}";
gchar *expanded_regex;
EggRegex *egg_re;
if (regex == NULL)
return NULL;
egg_re = egg_regex_new (re, 0, 0, NULL);
/* Use parser_state to pass the GError because we cannot pass
* it directly to the callback */
parser_state->error = error;
expanded_regex = egg_regex_replace_eval (egg_re, regex, len, 0, 0,
replace_by_id, parser_state);
parser_state->error = NULL;
if (*error == NULL)
{
DEBUG (g_message ("expanded regex vars '%s' to '%s'",
regex, expanded_regex));
}
egg_regex_unref (egg_re);
if (*error != NULL)
{
g_free (expanded_regex);
return NULL;
}
return expanded_regex;
}
static gboolean
replace_delimiter (const EggRegex *egg_regex,
const gchar *regex,
GString *expanded_regex,
gpointer data)
{
gchar *delim, *escapes;
ParserState *parser_state = data;
escapes = egg_regex_fetch (egg_regex, 1, regex);
g_string_append (expanded_regex, escapes);
delim = egg_regex_fetch (egg_regex, 2, regex);
DEBUG (g_message ("replacing '\\%%%s'", delim));
switch (delim[0])
{
case '[':
g_string_append (expanded_regex,
parser_state->opening_delimiter);
break;
case ']':
g_string_append (expanded_regex,
parser_state->closing_delimiter);
break;
}
g_free (delim);
g_free (escapes);
return FALSE;
}
static gchar *
expand_regex_delimiters (ParserState *parser_state,
gchar *regex,
gint len)
{
/* This is the commented regex without the doubled escape needed
* in a C string:
*
* (?<!\\)(\\\\)*\\%([\[|\])
* |------------||---------|
* | |
* | the strings
* | "\%[" or "\%]"
* |
* a even sequence of escapes or
* a char that is not an escape
* (this is not matched)
*
* The first block is needed to ensure that the sequence is
* not escaped.
*/
const gchar *re = "(?<!\\\\)(\\\\\\\\)*\\\\%(\\[|\\])";
gchar *expanded_regex;
EggRegex *egg_re;
if (regex == NULL)
return NULL;
egg_re = egg_regex_new (re, 0, 0, NULL);
expanded_regex = egg_regex_replace_eval (egg_re, regex, len, 0, 0,
replace_delimiter, parser_state);
DEBUG (g_message ("expanded regex delims '%s' to '%s'",
regex, expanded_regex));
egg_regex_unref (egg_re);
return expanded_regex;
}
static gchar *
expand_regex (ParserState *parser_state,
gchar *regex,
EggRegexCompileFlags flags,
gboolean do_expand_vars,
gboolean insert_parentheses,
GError **error)
{
gchar *tmp_regex;
GString *expanded_regex;
/* Length in *bytes* of the regex passed to expand_regex_vars. */
gint len;
g_assert (parser_state != NULL);
g_return_val_if_fail (error == NULL || *error == NULL, NULL);
if (regex == NULL)
return NULL;
if (g_utf8_get_char (regex) == '/')
{
gchar *regex_end = g_utf8_strrchr (regex, -1, '/');
if (regex_end == regex)
{
g_set_error (error,
PARSER_ERROR, PARSER_ERROR_MALFORMED_REGEX,
_ ("malformed regex '%s'"), regex);
return NULL;
}
regex = g_utf8_next_char (regex);
len = regex_end - regex;
}
else
{
len = -1;
}
if (do_expand_vars)
{
tmp_regex = expand_regex_vars (parser_state, regex, len, error);
}
else
{
if (len == -1)
len = strlen (regex);
tmp_regex = g_strndup (regex, len);
}
regex = tmp_regex;
tmp_regex = expand_regex_delimiters (parser_state, regex, len);
g_free (regex);
/* Set the options and add not capturing parentheses if
* insert_parentheses is TRUE (this is needed for included
* regular expressions.) */
expanded_regex = g_string_new ("");
if (insert_parentheses)
g_string_append (expanded_regex, "(?:");
g_string_append (expanded_regex, "(?");
if (flags != 0)
{
if (flags & EGG_REGEX_CASELESS)
g_string_append (expanded_regex, "i");
if (flags & EGG_REGEX_EXTENDED)
g_string_append (expanded_regex, "x");
if (flags & EGG_REGEX_DOTALL)
g_string_append (expanded_regex, "s");
}
if (flags != (EGG_REGEX_CASELESS | EGG_REGEX_EXTENDED |
EGG_REGEX_DOTALL))
{
g_string_append (expanded_regex, "-");
if (!(flags & EGG_REGEX_CASELESS))
g_string_append (expanded_regex, "i");
if (!(flags & EGG_REGEX_EXTENDED))
g_string_append (expanded_regex, "x");
if (!(flags & EGG_REGEX_DOTALL))
g_string_append (expanded_regex, "s");
}
g_string_append (expanded_regex, ")");
g_string_append (expanded_regex, tmp_regex);
if (insert_parentheses)
{
/* The '\n' is needed otherwise, if the regex is "extended"
* and it ends with a comment, the ')' is appended inside the
* comment itself */
if (flags & EGG_REGEX_EXTENDED)
{
g_string_append (expanded_regex, "\n");
}
/* FIXME: if the regex is not extended this doesn't works */
g_string_append (expanded_regex, ")");
}
g_free (tmp_regex);
return g_string_free (expanded_regex, FALSE);
}
static void
handle_define_regex_element (ParserState *parser_state,
GError **error)
{
gchar *id;
xmlChar *regex;
xmlChar *tmp;
gchar *expanded_regex;
int i;
const gchar *regex_options[] = {"extended", "case-insensitive", "dot-match-all", NULL};
EggRegexCompileFlags flags;
GError *tmp_error = NULL;
int type;
/* Return if an error is already set */
g_return_if_fail (error == NULL || *error == NULL);
if (parser_state->engine == NULL)
return;
tmp = xmlTextReaderGetAttribute (parser_state->reader, BAD_CAST "id");
/* If the file is validated <define-regex> must have an id
* attribute */
g_assert (tmp != NULL);
if (id_is_decorated ((gchar *)tmp, NULL))
id = g_strdup ((gchar *)tmp);
else
id = decorate_id (parser_state, (gchar *)tmp);
xmlFree (tmp);
flags = parser_state->regex_compile_flags;
for (i=0; regex_options[i] != NULL; i++)
{
tmp = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST regex_options[i]);
if (tmp != NULL)
flags = update_regex_flags (flags, regex_options[i],
str_to_bool ((gchar *)tmp));
xmlFree (tmp);
}
xmlTextReaderRead (parser_state->reader);
type = xmlTextReaderNodeType (parser_state->reader);
if (type == XML_READER_TYPE_TEXT || type == XML_READER_TYPE_CDATA)
regex = xmlTextReaderValue (parser_state->reader);
else
regex = xmlStrdup (BAD_CAST "");
expanded_regex = expand_regex (parser_state, (gchar*) regex, flags,
FALSE, TRUE, &tmp_error);
if (tmp_error == NULL)
{
DEBUG (g_message ("defined regex %s: \"%s\"", id, (gchar *)regex));
g_hash_table_insert (parser_state->defined_regexes, id, expanded_regex);
}
xmlFree (regex);
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
handle_default_regex_options_element (ParserState *parser_state,
GError **error)
{
xmlChar *options;
int ret, type;
/* Return if an error is already set */
g_return_if_fail (error == NULL || *error == NULL);
if (!parser_state->engine)
return;
do {
ret = xmlTextReaderRead (parser_state->reader);
g_assert (ret == 1);
type = xmlTextReaderNodeType (parser_state->reader);
} while (type != XML_READER_TYPE_TEXT &&
type != XML_READER_TYPE_CDATA);
options = xmlTextReaderValue (parser_state->reader);
g_strstrip ((gchar*) options);
if (!calc_regex_flags ((gchar*) options, &(parser_state->regex_compile_flags)))
{
g_set_error (error,
PARSER_ERROR,
PARSER_ERROR_MALFORMED_REGEX,
_ ("malformed regex options '%s'"),
(gchar*) options);
}
xmlFree (options);
}
static void
map_style (ParserState *parser_state,
gchar *style_id,
gchar *map_to,
GError **error)
{
const gchar *mapped_style;
g_return_if_fail (error == NULL || *error == NULL);
if (map_to != NULL)
mapped_style = g_hash_table_lookup (parser_state->styles_mapping,
map_to);
else
mapped_style = style_id;
DEBUG (g_message ("mapping the style of '%s' to '%s' -> '%s'",
style_id,
map_to ? map_to : "(null)",
mapped_style));
if (mapped_style)
g_hash_table_insert (parser_state->styles_mapping,
g_strdup (style_id),
g_strdup (mapped_style));
else
g_set_error (error,
PARSER_ERROR,
PARSER_ERROR_WRONG_ID,
"unable to map style '%s' to '%s'",
style_id, map_to);
}
// static GtkTextTag *
// create_tag (ParserState *parser_state,
// gchar *id,
// gchar *name,
// gchar *default_style,
// GError **error)
// {
// GtkTextTag *tag;
// GtkSourceTagStyle *ts = NULL;
//
// g_return_val_if_fail (parser_state != NULL, NULL);
// g_return_val_if_fail (id != NULL, NULL);
// g_return_val_if_fail (name != NULL, NULL);
//
// DEBUG (g_message ("new tag id '%s', name '%s'", id, name));
//
// tag = gtk_source_tag_new (name, id);
//
// DEBUG (g_message ("mapping the style of '%s' to '%s'",
// id, default_style));
//
// if (default_style != NULL)
// {
// ts = gtk_source_language_get_tag_style (parser_state->language,
// default_style);
// }
//
// if (ts != NULL)
// {
// gtk_source_tag_set_style (GTK_SOURCE_TAG (tag), ts);
// gtk_source_tag_style_free (ts);
// }
//
// return tag;
// }
static void
parse_language_with_id (ParserState *parser_state,
gchar *lang_id,
GError **error)
{
GtkSourceLanguagesManager *lm;
GtkSourceLanguage *imported_language;
GError *tmp_error = NULL;
g_return_if_fail (error == NULL || *error == NULL);
lm = _gtk_source_language_get_languages_manager (parser_state->language);
imported_language = gtk_source_languages_manager_get_language_by_id (lm, lang_id);
if (imported_language == NULL)
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_WRONG_ID,
"unable to resolve language '%s'",
lang_id);
}
else
{
file_parse (imported_language->priv->lang_file_name,
parser_state->language,
parser_state->engine,
parser_state->defined_regexes,
parser_state->styles_mapping,
parser_state->loaded_lang_ids,
&tmp_error);
}
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
parse_style (ParserState *parser_state,
GError **error)
{
gchar *id;
xmlChar *name, *map_to;
xmlChar *tmp;
gchar *lang_id = NULL;
GError *tmp_error = NULL;
g_return_if_fail (error == NULL || *error == NULL);
tmp = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "id");
if (id_is_decorated ((gchar*) tmp, NULL))
id = g_strdup ((gchar*) tmp);
else
id = decorate_id (parser_state, (gchar*) tmp);
xmlFree (tmp);
name = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "_name");
if (name)
{
tmp = xmlStrdup (BAD_CAST dgettext (parser_state->language->priv->translation_domain,
(gchar*) name));
xmlFree (name);
name = tmp;
}
else
{
name = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "name");
}
map_to = xmlTextReaderGetAttribute (parser_state->reader,
BAD_CAST "map-to");
if (map_to && !id_is_decorated ((gchar*) map_to, &lang_id))
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_MALFORMED_MAP_TO,
"the map-to attribute '%s' for the style '%s' lacks the prefix",
map_to, id);
}
if (!tmp_error && lang_id && !lang_id[0])
{
g_free (lang_id);
lang_id = NULL;
}
if (!tmp_error && lang_id && !lang_id_is_already_loaded (parser_state, lang_id))
parse_language_with_id (parser_state, lang_id, &tmp_error);
DEBUG (g_message ("style %s (%s) to be mapped to '%s'",
name, id, map_to ? (char*) map_to : "(null)"));
if (!tmp_error)
map_style (parser_state, id, (gchar*) map_to, &tmp_error);
g_free (lang_id);
g_free (id);
xmlFree (name);
xmlFree (map_to);
if (tmp_error)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
handle_keyword_char_class_element (ParserState *parser_state,
GError **error)
{
xmlChar *char_class;
int ret, type;
/* Return if an error is already set */
g_return_if_fail (!error || !*error);
if (!parser_state->engine)
return;
do {
ret = xmlTextReaderRead (parser_state->reader);
g_assert (ret == 1);
type = xmlTextReaderNodeType (parser_state->reader);
}
while (type != XML_READER_TYPE_TEXT && type != XML_READER_TYPE_CDATA);
char_class = xmlTextReaderValue (parser_state->reader);
g_free (parser_state->opening_delimiter);
g_free (parser_state->closing_delimiter);
parser_state->opening_delimiter = g_strdup_printf ("(?!<%s)(?=%s)",
char_class, char_class);
parser_state->closing_delimiter = g_strdup_printf ("(?<=%s)(?!%s)",
char_class, char_class);
xmlFree (char_class);
}
static void
handle_styles_element (ParserState *parser_state,
GError **error)
{
int ret, type;
const xmlChar *tag_name;
GError *tmp_error = NULL;
g_return_if_fail (error == NULL || *error == NULL);
while (TRUE)
{
ret = xmlTextReaderRead (parser_state->reader);
if (!xmlTextReaderIsValid (parser_state->reader))
{
/* TODO: get the error message from the
* xml parser */
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_INVALID_DOC,
"invalid language file");
break;
}
tag_name = xmlTextReaderConstName (parser_state->reader);
type = xmlTextReaderNodeType (parser_state->reader);
/* End at the closing </styles> tag */
if (tag_name && type == XML_READER_TYPE_END_ELEMENT &&
!xmlStrcmp (BAD_CAST "styles", tag_name))
break;
/* Skip nodes that aren't <style> elements */
if (!tag_name || xmlStrcmp (BAD_CAST "style", tag_name))
continue;
/* Handle <style> elements */
parse_style (parser_state, &tmp_error);
if (tmp_error != NULL)
break;
}
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
element_start (ParserState *parser_state,
GError **error)
{
const xmlChar *name;
GError *tmp_error = NULL;
/* TODO: check the namespace and ignore everithing is not in our namespace */
name = xmlTextReaderConstName (parser_state->reader);
if (xmlStrcmp (BAD_CAST "context", name) == 0)
{
handle_context_element (parser_state, &tmp_error);
}
else if (xmlStrcmp (BAD_CAST "define-regex", name) == 0)
{
handle_define_regex_element (parser_state, &tmp_error);
}
else if (xmlStrcmp (BAD_CAST "language", name) == 0)
{
handle_language_element (parser_state, &tmp_error);
}
else if (xmlStrcmp (BAD_CAST "styles", name) == 0)
{
handle_styles_element (parser_state, &tmp_error);
}
else if (xmlStrcmp (BAD_CAST "keyword-char-class", name) == 0)
{
handle_keyword_char_class_element (parser_state, &tmp_error);
}
else if (xmlStrcmp (BAD_CAST "default-regex-options", name) == 0)
{
handle_default_regex_options_element (parser_state, &tmp_error);
}
if (tmp_error != NULL)
{
g_propagate_error (error, tmp_error);
return;
}
}
static void
element_end (ParserState *parser_state)
{
const xmlChar *name;
gchar *popped_id;
name = xmlTextReaderConstName (parser_state->reader);
if (!xmlStrcmp (name, BAD_CAST "context"))
{
/* pop the first element in the curr_parents list */
popped_id = g_queue_pop_head (parser_state->curr_parents);
g_free (popped_id);
}
}
static gboolean
file_parse (gchar *filename,
GtkSourceLanguage *language,
GtkSourceContextEngine *engine,
GHashTable *defined_regexes,
GHashTable *styles,
GSList **loaded_lang_ids,
GError **error)
{
ParserState *parser_state;
xmlTextReader *reader = NULL;
int ret;
int fd;
GError *tmp_error = NULL;
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
DEBUG (g_message ("loading file '%s'", filename));
/*
* Use fd instead of filename so that it's utf8 safe on w32.
*/
fd = g_open (filename, O_RDONLY, 0);
if (fd != -1)
reader = xmlReaderForFd (fd, filename, NULL, 0);
if (reader == NULL)
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_CANNOT_OPEN,
"unable to open the file");
}
if (tmp_error == NULL)
{
GtkSourceLanguagesManager *lm;
const gchar *rng_lang_schema;
lm = _gtk_source_language_get_languages_manager (language);
rng_lang_schema = _gtk_source_languages_manager_get_rng_file (lm);
if (!rng_lang_schema)
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_CANNOT_VALIDATE,
"could not find the RelaxNG schema file");
}
else if (xmlTextReaderRelaxNGValidate (reader, rng_lang_schema))
{
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_CANNOT_VALIDATE,
"unable to load the RelaxNG schema '%s'",
rng_lang_schema);
}
}
parser_state = parser_state_new (language, engine,
defined_regexes, styles,
reader, loaded_lang_ids);
while (!tmp_error && (ret = xmlTextReaderRead (parser_state->reader)) == 1)
{
int type;
if (!xmlTextReaderIsValid (parser_state->reader))
{
/* TODO: get the error message from the
* xml parser */
g_set_error (&tmp_error,
PARSER_ERROR,
PARSER_ERROR_INVALID_DOC,
"invalid language file");
}
if (tmp_error)
break;
type = xmlTextReaderNodeType (parser_state->reader);
switch (type)
{
case XML_READER_TYPE_ELEMENT:
element_start (parser_state, &tmp_error);
break;
case XML_READER_TYPE_END_ELEMENT:
element_end (parser_state);
break;
}
}
parser_state_destroy (parser_state);
if (tmp_error)
{
g_propagate_error (error, tmp_error);
return FALSE;
}
return TRUE;
}
static ParserState *
parser_state_new (GtkSourceLanguage *language,
GtkSourceContextEngine *engine,
GHashTable *defined_regexes,
GHashTable *styles_mapping,
xmlTextReader *reader,
GSList **loaded_lang_ids)
{
ParserState *parser_state;
parser_state = g_new (ParserState, 1);
parser_state->language = language;
parser_state->engine = engine;
parser_state->current_lang_id = NULL;
parser_state->id_cookie = 0;
parser_state->regex_compile_flags = 0;
parser_state->reader = reader;
parser_state->defined_regexes = defined_regexes;
parser_state->styles_mapping = styles_mapping;
parser_state->loaded_lang_ids = loaded_lang_ids;
parser_state->curr_parents = g_queue_new ();
parser_state->opening_delimiter = g_strdup ("\\b");
parser_state->closing_delimiter = g_strdup ("\\b");
return parser_state;
}
static void
parser_state_destroy (ParserState *parser_state)
{
if (parser_state->reader != NULL)
xmlFreeTextReader (parser_state->reader);
g_queue_free (parser_state->curr_parents);
g_free (parser_state->current_lang_id);
g_free (parser_state->opening_delimiter);
g_free (parser_state->closing_delimiter);
g_free (parser_state);
}
static gboolean
steal_styles_mapping (char *style,
char *map_to,
GHashTable *styles)
{
if (strcmp (style, map_to))
{
g_hash_table_insert (styles, style, map_to);
return TRUE;
}
return FALSE;
}
gboolean
_gtk_source_language_file_parse_version2 (GtkSourceLanguage *language,
GtkSourceContextEngine *engine)
{
GHashTable *defined_regexes, *styles;
gboolean success;
GError *error = NULL;
gchar *filename;
GSList *loaded_lang_ids = NULL;
GSList *l;
g_return_val_if_fail (engine != NULL, FALSE);
filename = language->priv->lang_file_name;
/* TODO: tell the parser to validate the document while parsing
* (XML_PARSE_DTD_VALID or XML_PARSER_VALIDATE). */
/* TODO: as an optimization tell the parser to merge CDATA
* as text nodes (XML_PARSE_NOCDATA), and to ignore blank
* nodes (XML_PARSE_NOBLANKS), if it is possible with
* xmlTextReader. */
xmlKeepBlanksDefault (0);
xmlLineNumbersDefault (1);
xmlSubstituteEntitiesDefault (1);
DEBUG (xmlPedanticParserDefault (1));
defined_regexes = g_hash_table_new_full (g_str_hash, g_str_equal,
g_free, g_free);
styles = g_hash_table_new_full (g_str_hash, g_str_equal,
g_free, g_free);
success = file_parse (filename, language, engine,
defined_regexes, styles,
&loaded_lang_ids, &error);
if (success)
g_hash_table_foreach_steal (styles,
(GHRFunc) steal_styles_mapping,
language->priv->styles);
for (l = loaded_lang_ids; l != NULL; l = l->next)
{
xmlFree (l->data);
}
g_slist_free (loaded_lang_ids);
g_hash_table_destroy (defined_regexes);
g_hash_table_destroy (styles);
if (!success)
{
g_warning ("Failed to load '%s': %s",
filename, error->message);
g_error_free (error);
return FALSE;
}
return TRUE;
}
Reference in New Issue View Git Blame Copy Permalink