Implemented amb673 - pre-compiled regular expression support.

Also commented in the include file a bit better.
This commit is contained in:
Steve Dudenhoeffer 2007-08-01 17:36:10 +00:00
parent cf36abf7e6
commit 48022d3c5c
4 changed files with 222 additions and 17 deletions

View File

@ -43,12 +43,49 @@ bool RegEx::isFree(bool set, bool val)
}
}
int RegEx::Compile(const char *pattern)
int RegEx::Compile(const char *pattern, const char* flags)
{
if (!mFree)
Clear();
re = pcre_compile(pattern, 0, &mError, &mErrorOffset, NULL);
int iFlags = 0;
if (flags != NULL)
{
for ( ; *flags != 0; flags++)
{
switch (*flags)
{
case 'i':
{
iFlags |= PCRE_CASELESS;
break;
}
case 'm':
{
iFlags |= PCRE_MULTILINE;
break;
}
case 's':
{
iFlags |= PCRE_DOTALL;
break;
}
case 'x':
{
iFlags |= PCRE_EXTENDED;
break;
}
default:
{
break;
}
}
}
}
re = pcre_compile(pattern, iFlags, &mError, &mErrorOffset, NULL);
if (re == NULL)
{
@ -67,6 +104,8 @@ int RegEx::Match(const char *str)
if (mFree || re == NULL)
return -1;
this->ClearMatch();
//save str
subject = new char[strlen(str)+1];
strcpy(subject, str);
@ -88,6 +127,16 @@ int RegEx::Match(const char *str)
return 1;
}
void RegEx::ClearMatch()
{
// Clears match results
mErrorOffset = 0;
mError = NULL;
if (subject)
delete [] subject;
subject = NULL;
mSubStrings = 0;
}
const char *RegEx::GetSubstring(int s, char buffer[], int max)
{

View File

@ -9,8 +9,9 @@ public:
bool isFree(bool set=false, bool val=false);
void Clear();
int Compile(const char *pattern);
int Compile(const char *pattern, const char* flags = NULL);
int Match(const char *str);
void ClearMatch();
const char *GetSubstring(int s, char buffer[], int max);
public:
int mErrorOffset;

View File

@ -19,7 +19,28 @@ int GetPEL()
return (int)PEL.size() - 1;
}
// native Regex:regex_compile(const pattern[], &ret, error[], maxLen, const flags[]="");
static cell AMX_NATIVE_CALL regex_compile(AMX *amx, cell *params)
{
int len;
const char *regex = MF_GetAmxString(amx, params[1], 0, &len);
const char *flags = MF_GetAmxString(amx, params[5], 1, &len);
int id = GetPEL();
RegEx *x = PEL[id];
if (x->Compile(regex, flags) == 0)
{
cell *eOff = MF_GetAmxAddr(amx, params[2]);
const char *err = x->mError;
*eOff = x->mErrorOffset;
MF_SetAmxString(amx, params[3], err?err:"unknown", params[4]);
return -1;
}
return id+1;
}// 1.8 includes the last parameter
// Regex:regex_match(const string[], const pattern[], &ret, error[], maxLen, const flags[] = "");
static cell AMX_NATIVE_CALL regex_match(AMX *amx, cell *params)
{
int len;
@ -29,7 +50,14 @@ static cell AMX_NATIVE_CALL regex_match(AMX *amx, cell *params)
int id = GetPEL();
RegEx *x = PEL[id];
if (x->Compile(regex) == 0)
char* flags = NULL;
if ((params[0] / sizeof(cell)) >= 6) // compiled with 1.8's extra parameter
{
flags = MF_GetAmxString(amx, params[6], 2, &len);
}
if (x->Compile(regex, flags) == 0)
{
cell *eOff = MF_GetAmxAddr(amx, params[3]);
const char *err = x->mError;
@ -58,6 +86,44 @@ static cell AMX_NATIVE_CALL regex_match(AMX *amx, cell *params)
return id+1;
}
// native regex_match_c(const string[], Regex:id, &ret);
static cell AMX_NATIVE_CALL regex_match_c(AMX *amx, cell *params)
{
int len;
int id = params[2]-1;
const char *str = MF_GetAmxString(amx, params[1], 0, &len);
if (id >= (int)PEL.size() || id < 0 || PEL[id]->isFree())
{
MF_LogError(amx, AMX_ERR_NATIVE, "Invalid regex handle %d", id);
return 0;
}
RegEx *x = PEL[id];
int e = x->Match(str);
if (e == -1)
{
/* there was a match error. move on. */
cell *res = MF_GetAmxAddr(amx, params[3]);
*res = x->mErrorOffset;
/* only clear the match results, since the regex object
may still be referenced later */
x->ClearMatch();
return -2;
} else if (e == 0) {
cell *res = MF_GetAmxAddr(amx, params[3]);
*res = 0;
/* only clear the match results, since the regex object
may still be referenced later */
x->ClearMatch();
return 0;
} else {
cell *res = MF_GetAmxAddr(amx, params[3]);
*res = x->mSubStrings;
return x->mSubStrings;
}
}
static cell AMX_NATIVE_CALL regex_substr(AMX *amx, cell *params)
{
@ -101,7 +167,9 @@ static cell AMX_NATIVE_CALL regex_free(AMX *amx, cell *params)
}
AMX_NATIVE_INFO regex_Natives[] = {
{"regex_compile", regex_compile},
{"regex_match", regex_match},
{"regex_match_c", regex_match_c},
{"regex_substr", regex_substr},
{"regex_free", regex_free},
{NULL, NULL},

View File

@ -26,22 +26,109 @@ enum Regex
REGEX_OK
};
/* Return values:
-2 = Matching error (error code stored in ret)
-1 = Error in pattern (error message and offset # in error[] and ret)
0 = No match
>1 = Id for getting more info (you must call regex_free() later on)
(also note that ret will contain the number of substrings found)
/**
* Precompile a regular expression. Use this if you intend on using the
* same expression multiple times. Pass the regex handle returned here to
* regex_match_c to check for matches.
*
* @param pattern The regular expression pattern.
* @param errcode Error code encountered, if applicable.
* @param error Error message encountered, if applicable.
* @param maxLen Maximum string length of the error buffer.
* @param flags General flags for the regular expression.
* i = Ignore case
* m = Multilines (affects ^ and $ so that they match
* the start/end of a line rather than matching the
* start/end of the string).
* s = Single line (affects . so that it matches any character,
* even new line characters).
* x = Pattern extension (ignore whitespace and # comments).
*
* @return -1 on error in the pattern, > valid regex handle (> 0) on success.
*
* @note This handle is automatically freed on map change. However,
* if you are completely done with it before then, you should
* call regex_free on this handle.
*/
native Regex:regex_compile(const pattern[], &ret, error[], maxLen, const flags[]="");
native Regex:regex_match(const string[], const pattern[], &ret, error[], maxLen);
/**
* Matches a string against a pre-compiled regular expression pattern.
*
*
* @param pattern The regular expression pattern.
* @param string The string to check.
* @param ret Error code, if applicable, or number of results on success.
*
* @return -2 = Matching error (error code is stored in ret)
* 0 = No match.
* >1 = Number of results.
*
* @note You should free the returned handle (with regex_free())
* when you are done with this pattern.
*
* @note Use the regex handle passed to this function to extract
* matches with regex_substr().
*/
native regex_match_c(const string[], Regex:pattern, &ret);
/* Returns a matched substring from a regex handle
* substring ids start at 0 and end at ret-1, where ret is from the above function
/**
* Matches a string against a regular expression pattern.
*
* @note If you intend on using the same regular expression pattern
* multiple times, consider using regex_compile and regex_match_c
* instead of making this function reparse the expression each time.
*
* @param string The string to check.
* @param pattern The regular expression pattern.
* @param ret Error code, or result state of the match.
* @param error Error message, if applicable.
* @param maxLen Maximum length of the error buffer.
* @param flags General flags for the regular expression.
* i = Ignore case
* m = Multilines (affects ^ and $ so that they match
* the start/end of a line rather than matching the
* start/end of the string).
* s = Single line (affects . so that it matches any character,
* even new line characters).
* x = Pattern extension (ignore whitespace and # comments).
*
* @return -2 = Matching error (error code is stored in ret)
* -1 = Error in pattern (error message and offset # in error and ret)
* 0 = No match.
* >1 = Handle for getting more information (via regex_substr)
*
* @note Flags only exist in amxmodx 1.8 and later.
* @note You should free the returned handle (with regex_free())
* when you are done extracting all of the substrings.
*/
native Regex:regex_match(const string[], const pattern[], &ret, error[], maxLen, const flags[] = "");
/**
* Returns a matched substring from a regex handle.
* Substring ids start at 0 and end at ret-1, where ret is from the corresponding
* regex_match or regex_match_c function call.
*
* @param id The regex handle to extract data from.
* @param str_id The index of the expression to get - starts at 0, and ends at ret - 1.
* @param buffer The buffer to set to the matching substring.
* @param maxLen The maximum string length of the buffer.
*
*/
native regex_substr(Regex:id, str_id, buffer[], maxLen);
/* Frees the memory associated with a regex results and sets the handle to 0.
* You must do this if the handle >=1, once you're done.
/**
* Frees the memory associated with a regex result, and sets the handle to 0.
* This must be called on all results from regex_match() when you are done extracting
* the results with regex_substr().
* The results of regex_compile() (and subsequently, regex_match_c()) only need to be freed
* when you are done using the pattern.
*
*
* @param id The regex handle to free.
*
* @noreturn
*
* @note Do not use the handle again after freeing it!
*/
native regex_free(&Regex:id);