1berry/wireshark
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

extcap: Rewrite the tokenizer to use regexps

Browse Source 
Change the tokenizer to two regular expressions, which
make the parsing of the sentence strings a lot safer and faster.

Change-Id: I444adb8db10b689dd387c0caa951981ba28be917
Reviewed-on: https://code.wireshark.org/review/13040
Reviewed-by: Roland Knall <rknall@gmail.com>
Reviewed-by: Stig Bjørlykke <stig@bjorlykke.org>
This commit is contained in:
Roland Knall 2016-01-04 14:19:55 +01:00 committed by Stig Bjørlykke
parent be0a7273f7
commit e1c97dd440
1 changed files with 71 additions and 93 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

164
extcap_parser.c
View File

@ -270,9 +270,10 @@ void extcap_free_tokenized_sentence_list(extcap_token_sentence *f) {
}
extcap_token_sentence *extcap_tokenize_sentence(const gchar *s) {
gchar *b, *e, *eq;
extcap_token_param *tv = NULL;
GRegex * regex = NULL;
GMatchInfo * match_info = NULL;
GError * error = NULL;
extcap_token_sentence *rs = g_new(extcap_token_sentence, 1);
@ -280,106 +281,83 @@ extcap_token_sentence *extcap_tokenize_sentence(const gchar *s) {
rs->next_sentence = NULL;
rs->param_list = NULL;
if ((b = g_strstr_len(s, -1, " ")) == NULL) {
/* Regex for catching just the allowed values for sentences */
if ( ( regex = g_regex_new ( "^[\\t| ]*(arg|value|interface|dlt)(?=[\\t| ]+\\{)",
(GRegexCompileFlags) G_REGEX_CASELESS, (GRegexMatchFlags) 0, NULL ) ) != NULL ) {
g_regex_match ( regex, s, (GRegexMatchFlags) 0, &match_info );
if ( g_match_info_matches ( match_info ) )
rs->sentence = g_match_info_fetch(match_info, 0);
g_match_info_free ( match_info );
g_regex_unref ( regex );
}
/* No valid sentence found, exiting here */
if ( rs->sentence == NULL ) {
extcap_free_tokenized_sentence(rs);
return NULL ;
return NULL;
}
rs->sentence = g_strndup(s, b - s);
/* Capture the argument and the value of the list. This will ensure,
* that regex patterns given to {validation=} are parsed correctly,
* as long as }{ does not occur within the pattern */
regex = g_regex_new ( "\\{([a-zA-Z_-]*?)\\=(.*?)\\}(?=\\{|$|\\s)",
(GRegexCompileFlags) G_REGEX_CASELESS, (GRegexMatchFlags) 0, NULL );
if ( regex != NULL ) {
g_regex_match_full(regex, s, -1, 0, (GRegexMatchFlags) 0, &match_info, &error );
while(g_match_info_matches(match_info)) {
gchar * arg = g_match_info_fetch ( match_info, 1 );
if ((b = g_strstr_len(s, -1, "{")) == NULL) {
/* printf("debug - tokenizer - sentence with no values\n"); */
extcap_free_tokenized_sentence(rs);
return NULL ;
}
if ( arg == NULL )
break;
while (b != NULL ) {
if ((e = g_strstr_len(b, -1, "}")) == NULL) {
/* printf("debug - tokenizer - invalid, missing }\n"); */
extcap_free_tokenized_sentence(rs);
return NULL ;
}
tv = g_new(extcap_token_param, 1);
tv->arg = arg;
tv->value = g_match_info_fetch ( match_info, 2 );
/* caught a regex quantifier end bracket and not the end of the line.
* let's find the correct end bracket */
if ( *(e+1) != '{' && strlen ( e ) > 1 ) {
gchar *f = (e + 1);
while ( ( f = g_strstr_len(f, -1, "}") ) != NULL) {
if ( strlen ( f ) <= 1 || *(f+1) == '{' )
break;
f++;
if (g_ascii_strcasecmp(tv->arg, "number") == 0) {
tv->param_type = EXTCAP_PARAM_ARGNUM;
} else if (g_ascii_strcasecmp(tv->arg, "call") == 0) {
tv->param_type = EXTCAP_PARAM_CALL;
} else if (g_ascii_strcasecmp(tv->arg, "display") == 0) {
tv->param_type = EXTCAP_PARAM_DISPLAY;
} else if (g_ascii_strcasecmp(tv->arg, "type") == 0) {
tv->param_type = EXTCAP_PARAM_TYPE;
} else if (g_ascii_strcasecmp(tv->arg, "arg") == 0) {
tv->param_type = EXTCAP_PARAM_ARG;
} else if (g_ascii_strcasecmp(tv->arg, "default") == 0) {
tv->param_type = EXTCAP_PARAM_DEFAULT;
} else if (g_ascii_strcasecmp(tv->arg, "value") == 0) {
tv->param_type = EXTCAP_PARAM_VALUE;
} else if (g_ascii_strcasecmp(tv->arg, "range") == 0) {
tv->param_type = EXTCAP_PARAM_RANGE;
} else if (g_ascii_strcasecmp(tv->arg, "tooltip") == 0) {
tv->param_type = EXTCAP_PARAM_TOOLTIP;
} else if (g_ascii_strcasecmp(tv->arg, "mustexist") == 0) {
tv->param_type = EXTCAP_PARAM_FILE_MUSTEXIST;
} else if (g_ascii_strcasecmp(tv->arg, "fileext") == 0) {
tv->param_type = EXTCAP_PARAM_FILE_EXTENSION;
} else if (g_ascii_strcasecmp(tv->arg, "name") == 0) {
tv->param_type = EXTCAP_PARAM_NAME;
} else if (g_ascii_strcasecmp(tv->arg, "enabled") == 0) {
tv->param_type = EXTCAP_PARAM_ENABLED;
} else if (g_ascii_strcasecmp(tv->arg, "parent") == 0) {
tv->param_type = EXTCAP_PARAM_PARENT;
} else if (g_ascii_strcasecmp(tv->arg, "required") == 0) {
tv->param_type = EXTCAP_PARAM_REQUIRED;
} else if (g_ascii_strcasecmp(tv->arg, "validation") == 0) {
tv->param_type = EXTCAP_PARAM_VALIDATION;
} else {
tv->param_type = EXTCAP_PARAM_UNKNOWN;
}
if ( f != NULL )
e = f;
tv->next_token = rs->param_list;
rs->param_list = tv;
g_match_info_next(match_info, &error);
}
if ((eq = g_strstr_len(b, -1, "=")) == NULL) {
/* printf("debug - tokenizer - invalid, missing =\n"); */
extcap_free_tokenized_sentence(rs);
return NULL ;
}
b++;
e--;
if (b >= eq || e <= eq) {
/* printf("debug - tokenizer - invalid, missing arg or value in {}\n"); */
extcap_free_tokenized_sentence(rs);
return NULL ;
}
tv = g_new(extcap_token_param, 1);
tv->arg = g_strndup(b, eq - b);
tv->value = g_strndup(eq + 1, e - eq);
if (g_ascii_strcasecmp(tv->arg, "number") == 0) {
tv->param_type = EXTCAP_PARAM_ARGNUM;
} else if (g_ascii_strcasecmp(tv->arg, "call") == 0) {
tv->param_type = EXTCAP_PARAM_CALL;
} else if (g_ascii_strcasecmp(tv->arg, "display") == 0) {
tv->param_type = EXTCAP_PARAM_DISPLAY;
} else if (g_ascii_strcasecmp(tv->arg, "type") == 0) {
tv->param_type = EXTCAP_PARAM_TYPE;
} else if (g_ascii_strcasecmp(tv->arg, "arg") == 0) {
tv->param_type = EXTCAP_PARAM_ARG;
} else if (g_ascii_strcasecmp(tv->arg, "default") == 0) {
tv->param_type = EXTCAP_PARAM_DEFAULT;
} else if (g_ascii_strcasecmp(tv->arg, "value") == 0) {
tv->param_type = EXTCAP_PARAM_VALUE;
} else if (g_ascii_strcasecmp(tv->arg, "range") == 0) {
tv->param_type = EXTCAP_PARAM_RANGE;
} else if (g_ascii_strcasecmp(tv->arg, "tooltip") == 0) {
tv->param_type = EXTCAP_PARAM_TOOLTIP;
} else if (g_ascii_strcasecmp(tv->arg, "mustexist") == 0) {
tv->param_type = EXTCAP_PARAM_FILE_MUSTEXIST;
} else if (g_ascii_strcasecmp(tv->arg, "fileext") == 0) {
tv->param_type = EXTCAP_PARAM_FILE_EXTENSION;
} else if (g_ascii_strcasecmp(tv->arg, "name") == 0) {
tv->param_type = EXTCAP_PARAM_NAME;
} else if (g_ascii_strcasecmp(tv->arg, "enabled") == 0) {
tv->param_type = EXTCAP_PARAM_ENABLED;
} else if (g_ascii_strcasecmp(tv->arg, "parent") == 0) {
tv->param_type = EXTCAP_PARAM_PARENT;
} else if (g_ascii_strcasecmp(tv->arg, "required") == 0) {
tv->param_type = EXTCAP_PARAM_REQUIRED;
} else if (g_ascii_strcasecmp(tv->arg, "validation") == 0) {
tv->param_type = EXTCAP_PARAM_VALIDATION;
} else {
tv->param_type = EXTCAP_PARAM_UNKNOWN;
}
tv->next_token = rs->param_list;
rs->param_list = tv;
/* printf("debug - tokenizer - got '%s' = '%s'\n", tv->arg, tv->value); */
b = e + 1;
if ((size_t) (b - s) > strlen(s))
break;
b = g_strstr_len(b, -1, "{");
g_match_info_free(match_info);
g_regex_unref(regex);
}
return rs;