UNLOZE/sourcemod
10
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix/Update regex extension and docs (, )

Browse Source
This commit is contained in:
Dr!fter 2018-02-12 18:59:28 -05:00
parent 93e9a29353
commit cdb9851da6
5 changed files with 443 additions and 43 deletions

79
extensions/regex/CRegEx.cpp
View File

@ -33,22 +33,23 @@
#include "pcre.h"
#include "CRegEx.h"
#include <sh_string.h>
#include "extension.h"
RegEx::RegEx()
{
mErrorOffset = 0;
mErrorCode = 0;
mError = NULL;
re = NULL;
mFree = true;
subject = NULL;
mSubStrings = 0;
mMatchCount = 0;
}
void RegEx::Clear ()
{
mErrorOffset = 0;
mErrorCode = 0;
mError = NULL;
if (re)
pcre_free(re);
@ -57,7 +58,7 @@ void RegEx::Clear ()
if (subject)
delete [] subject;
subject = NULL;
mSubStrings = 0;
mMatchCount = 0;
}
RegEx::~RegEx()
@ -81,7 +82,7 @@ int RegEx::Compile(const char *pattern, int iFlags)
if (!mFree)
Clear();
re = pcre_compile(pattern, iFlags, &mError, &mErrorOffset, NULL);
re = pcre_compile2(pattern, iFlags, &mErrorCode, &mError, &mErrorOffset, NULL);
if (re == NULL)
{
@ -93,7 +94,7 @@ int RegEx::Compile(const char *pattern, int iFlags)
return 1;
}
int RegEx::Match(const char *str)
int RegEx::Match(const char *str, unsigned int offset)
{
int rc = 0;
@ -106,7 +107,9 @@ int RegEx::Match(const char *str)
subject = new char[strlen(str)+1];
strcpy(subject, str);
rc = pcre_exec(re, NULL, subject, (int)strlen(subject), 0, 0, ovector, 30);
unsigned int len = strlen(subject);
rc = pcre_exec(re, NULL, subject, len, offset, 0, mMatches[0].mVector, MAX_CAPTURES);
if (rc < 0)
{
@ -114,34 +117,80 @@ int RegEx::Match(const char *str)
{
return 0;
} else {
mErrorOffset = rc;
mErrorCode = rc;
return -1;
}
}
mSubStrings = rc;
mMatches[0].mSubStringCount = rc;
mMatchCount = 1;
return 1;
}
int RegEx::MatchAll(const char *str)
{
int rc = 0;
if (mFree || re == NULL)
return -1;
this->ClearMatch();
//save str
subject = new char[strlen(str) + 1];
strcpy(subject, str);
unsigned int offset = 0;
unsigned int len = strlen(subject);
unsigned int matches = 0;
while (matches < MAX_MATCHES && offset < len && (rc = pcre_exec(re, 0, subject, len, offset, 0, mMatches[matches].mVector, MAX_CAPTURES)) >= 0)
{
offset = mMatches[matches].mVector[1];
mMatches[matches].mSubStringCount = rc;
matches++;
}
if (rc < PCRE_ERROR_NOMATCH || (rc == PCRE_ERROR_NOMATCH && matches == 0))
{
if (rc == PCRE_ERROR_NOMATCH)
{
return 0;
}
else {
mErrorCode = rc;
return -1;
}
}
mMatchCount = matches;
return 1;
}
void RegEx::ClearMatch()
{
// Clears match results
mErrorOffset = 0;
mErrorCode = 0;
mError = NULL;
if (subject)
delete [] subject;
subject = NULL;
mSubStrings = 0;
mMatchCount = 0;
}
const char *RegEx::GetSubstring(int s, char buffer[], int max)
bool RegEx::GetSubstring(int s, char buffer[], int max, int match)
{
int i = 0;
if (s >= mSubStrings || s < 0)
return NULL;
char *substr_a = subject + ovector[2*s];
int substr_l = ovector[2*s+1] - ovector[2*s];
if (s >= mMatches[match].mSubStringCount || s < 0)
return false;
char *substr_a = subject + mMatches[match].mVector[2 * s];
int substr_l = mMatches[match].mVector[2 * s + 1] - mMatches[match].mVector[2 * s];
for (i = 0; i<substr_l; i++)
{
@ -152,6 +201,6 @@ const char *RegEx::GetSubstring(int s, char buffer[], int max)
buffer[i] = '\0';
return buffer;
return true;
}

20
extensions/regex/CRegEx.h
View File

@ -28,10 +28,20 @@
*
* Version: $Id$
*/
#include <am-string.h>
#ifndef _INCLUDE_CREGEX_H
#define _INCLUDE_CREGEX_H
#define MAX_MATCHES 20
#define MAX_CAPTURES MAX_MATCHES*3
struct RegexMatch
{
int mSubStringCount;
int mVector[MAX_CAPTURES];
};
class RegEx
{
public:
@ -41,17 +51,19 @@ public:
void Clear();
int Compile(const char *pattern, int iFlags);
int Match(const char *str);
int Match(const char *str, unsigned int offset);
int MatchAll(const char *str);
void ClearMatch();
const char *GetSubstring(int s, char buffer[], int max);
bool GetSubstring(int s, char buffer[], int max, int match);
public:
int mErrorOffset;
int mErrorCode;
const char *mError;
int mSubStrings;
int mMatchCount;
RegexMatch mMatches[MAX_MATCHES];
private:
pcre *re;
bool mFree;
int ovector[30];
char *subject;
};

173
extensions/regex/extension.cpp
View File

@ -33,6 +33,7 @@
#include "extension.h"
#include <sh_string.h>
#include "pcre.h"
#include "posix_map.h"
#include "CRegEx.h"
using namespace SourceHook;
@ -82,10 +83,11 @@ static cell_t CompileRegex(IPluginContext *pCtx, const cell_t *params)
if (x->Compile(regex, params[2]) == 0)
{
cell_t *eOff;
pCtx->LocalToPhysAddr(params[5], &eOff);
cell_t *eError;
pCtx->LocalToPhysAddr(params[5], &eError);
const char *err = x->mError;
*eOff = x->mErrorOffset;
// Convert error code to posix error code but use pcre's error string since it is more detailed.
*eError = pcre_posix_compile_error_map[x->mErrorCode];
pCtx->StringToLocal(params[3], params[4], err ? err:"unknown");
delete x;
return 0;
@ -112,6 +114,13 @@ static cell_t MatchRegex(IPluginContext *pCtx, const cell_t *params)
sec.pOwner = NULL;
sec.pIdentity = myself->GetIdentity();
unsigned int offset = 0;
if (params[0] >= 4)
{
offset = (unsigned int)params[4];
}
RegEx *x;
if ((err=g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
@ -129,14 +138,17 @@ static cell_t MatchRegex(IPluginContext *pCtx, const cell_t *params)
char *str;
pCtx->LocalToString(params[2], &str);
int e = x->Match(str);
if(offset >= strlen(str))
return pCtx->ThrowNativeError("Invalid string index\n");
int e = x->Match(str, offset);
if (e == -1)
{
/* there was a match error. move on. */
cell_t *res;
pCtx->LocalToPhysAddr(params[3], &res);
*res = x->mErrorOffset;
*res = x->mErrorCode;
/* only clear the match results, since the regex object
may still be referenced later */
x->ClearMatch();
@ -153,7 +165,60 @@ static cell_t MatchRegex(IPluginContext *pCtx, const cell_t *params)
}
else
{
return x->mSubStrings;
return x->mMatches[0].mSubStringCount;
}
}
static cell_t MatchRegexAll(IPluginContext *pCtx, const cell_t *params)
{
Handle_t hndl = static_cast<Handle_t>(params[1]);
HandleError err;
HandleSecurity sec;
sec.pOwner = NULL;
sec.pIdentity = myself->GetIdentity();
RegEx *x;
if ((err = g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
{
return pCtx->ThrowNativeError("Invalid regex handle %x (error %d)", hndl, err);
}
if (!x)
{
pCtx->ThrowNativeError("Regex data not found\n");
return 0;
}
char *str;
pCtx->LocalToString(params[2], &str);
int e = x->MatchAll(str);
if (e == -1)
{
/* there was a match error. move on. */
cell_t *res;
pCtx->LocalToPhysAddr(params[3], &res);
*res = x->mErrorCode;
/* only clear the match results, since the regex object
may still be referenced later */
x->ClearMatch();
return -1;
}
else if (e == 0)
{
/* only clear the match results, since the regex object
may still be referenced later */
x->ClearMatch();
return 0;
}
else
{
return x->mMatchCount;
}
}
@ -165,6 +230,8 @@ static cell_t GetRegexSubString(IPluginContext *pCtx, const cell_t *params)
sec.pOwner=NULL;
sec.pIdentity=myself->GetIdentity();
int match = 0;
RegEx *x;
if ((err=g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
@ -178,17 +245,93 @@ static cell_t GetRegexSubString(IPluginContext *pCtx, const cell_t *params)
return 0;
}
static char buffer[4096];
const char *ret=x->GetSubstring(params[2], buffer, sizeof(buffer));
if(!ret)
if (params[0] >= 5)
{
return 0;
match = params[5];
}
pCtx->StringToLocalUTF8(params[3], params[4], ret, NULL);
if(match >= x->mMatchCount || match < 0)
return pCtx->ThrowNativeError("Invalid match index passed.\n");
return 1;
char *buffer;
pCtx->LocalToString(params[3], &buffer);
return x->GetSubstring(params[2], buffer, params[4], match);
}
static cell_t GetRegexMatchCount(IPluginContext *pCtx, const cell_t *params)
{
Handle_t hndl = static_cast<Handle_t>(params[1]);
HandleError err;
HandleSecurity sec;
sec.pOwner = NULL;
sec.pIdentity = myself->GetIdentity();
RegEx *x;
if ((err = g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
{
return pCtx->ThrowNativeError("Invalid regex handle %x (error %d)", hndl, err);
}
if (!x)
{
return pCtx->ThrowNativeError("Regex data not found\n");
}
return x->mMatchCount;
}
static cell_t GetRegexCaptureCount(IPluginContext *pCtx, const cell_t *params)
{
Handle_t hndl = static_cast<Handle_t>(params[1]);
HandleError err;
HandleSecurity sec;
sec.pOwner = NULL;
sec.pIdentity = myself->GetIdentity();
RegEx *x;
if ((err = g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
{
return pCtx->ThrowNativeError("Invalid regex handle %x (error %d)", hndl, err);
}
if (!x)
{
return pCtx->ThrowNativeError("Regex data not found\n");
}
if (params[2] >= x->mMatchCount || params[2] < 0)
return pCtx->ThrowNativeError("Invalid match index passed.\n");
return x->mMatches[params[2]].mSubStringCount;
}
static cell_t GetRegexOffset(IPluginContext *pCtx, const cell_t *params)
{
Handle_t hndl = static_cast<Handle_t>(params[1]);
HandleError err;
HandleSecurity sec;
sec.pOwner = NULL;
sec.pIdentity = myself->GetIdentity();
RegEx *x;
if ((err = g_pHandleSys->ReadHandle(hndl, g_RegexHandle, &sec, (void **)&x)) != HandleError_None)
{
return pCtx->ThrowNativeError("Invalid regex handle %x (error %d)", hndl, err);
}
if (!x)
{
return pCtx->ThrowNativeError("Regex data not found\n");
}
if (params[2] >= x->mMatchCount || params[2] < 0)
return pCtx->ThrowNativeError("Invalid match index passed.\n");
return x->mMatches[params[2]].mVector[1];
}
void RegexHandler::OnHandleDestroy(HandleType_t type, void *object)
@ -209,5 +352,9 @@ const sp_nativeinfo_t regex_natives[] =
{"Regex.GetSubString", GetRegexSubString},
{"Regex.Match", MatchRegex},
{"Regex.Regex", CompileRegex},
{"Regex.MatchAll", MatchRegexAll},
{"Regex.MatchCount", GetRegexMatchCount},
{"Regex.CaptureCount", GetRegexCaptureCount},
{"Regex.MatchOffset", GetRegexOffset},
{NULL, NULL},
};

136
extensions/regex/posix_map.h Normal file
View File

@ -0,0 +1,136 @@
#pragma once
/*
Maps pcre_compile2 error codes to posix error codes.
From pcreposix.c and pcreposix.h
*/
// posix error codes
enum {
REG_ASSERT = 1, /* internal error ? */
REG_BADBR, /* invalid repeat counts in {} */
REG_BADPAT, /* pattern error */
REG_BADRPT, /* ? * + invalid */
REG_EBRACE, /* unbalanced {} */
REG_EBRACK, /* unbalanced [] */
REG_ECOLLATE, /* collation error - not relevant */
REG_ECTYPE, /* bad class */
REG_EESCAPE, /* bad escape sequence */
REG_EMPTY, /* empty expression */
REG_EPAREN, /* unbalanced () */
REG_ERANGE, /* bad range inside [] */
REG_ESIZE, /* expression too big */
REG_ESPACE, /* failed to get memory */
REG_ESUBREG, /* bad back reference */
REG_INVARG, /* bad argument */
// This isnt used below since it is not a compile error. So we remove it as to not conflict.
//REG_NOMATCH /* match failed */
};
// pcre compile error -> posix compile error
const int pcre_posix_compile_error_map[] = {
0, /* no error */
REG_EESCAPE, /* \ at end of pattern */
REG_EESCAPE, /* \c at end of pattern */
REG_EESCAPE, /* unrecognized character follows \ */
REG_BADBR, /* numbers out of order in {} quantifier */
/* 5 */
REG_BADBR, /* number too big in {} quantifier */
REG_EBRACK, /* missing terminating ] for character class */
REG_ECTYPE, /* invalid escape sequence in character class */
REG_ERANGE, /* range out of order in character class */
REG_BADRPT, /* nothing to repeat */
/* 10 */
REG_BADRPT, /* operand of unlimited repeat could match the empty string */
REG_ASSERT, /* internal error: unexpected repeat */
REG_BADPAT, /* unrecognized character after (? */
REG_BADPAT, /* POSIX named classes are supported only within a class */
REG_EPAREN, /* missing ) */
/* 15 */
REG_ESUBREG, /* reference to non-existent subpattern */
REG_INVARG, /* erroffset passed as NULL */
REG_INVARG, /* unknown option bit(s) set */
REG_EPAREN, /* missing ) after comment */
REG_ESIZE, /* parentheses nested too deeply */
/* 20 */
REG_ESIZE, /* regular expression too large */
REG_ESPACE, /* failed to get memory */
REG_EPAREN, /* unmatched parentheses */
REG_ASSERT, /* internal error: code overflow */
REG_BADPAT, /* unrecognized character after (?< */
/* 25 */
REG_BADPAT, /* lookbehind assertion is not fixed length */
REG_BADPAT, /* malformed number or name after (?( */
REG_BADPAT, /* conditional group contains more than two branches */
REG_BADPAT, /* assertion expected after (?( */
REG_BADPAT, /* (?R or (?[+-]digits must be followed by ) */
/* 30 */
REG_ECTYPE, /* unknown POSIX class name */
REG_BADPAT, /* POSIX collating elements are not supported */
REG_INVARG, /* this version of PCRE is not compiled with PCRE_UTF8 support */
REG_BADPAT, /* spare error */
REG_BADPAT, /* character value in \x{} or \o{} is too large */
/* 35 */
REG_BADPAT, /* invalid condition (?(0) */
REG_BADPAT, /* \C not allowed in lookbehind assertion */
REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */
REG_BADPAT, /* number after (?C is > 255 */
REG_BADPAT, /* closing ) for (?C expected */
/* 40 */
REG_BADPAT, /* recursive call could loop indefinitely */
REG_BADPAT, /* unrecognized character after (?P */
REG_BADPAT, /* syntax error in subpattern name (missing terminator) */
REG_BADPAT, /* two named subpatterns have the same name */
REG_BADPAT, /* invalid UTF-8 string */
/* 45 */
REG_BADPAT, /* support for \P, \p, and \X has not been compiled */
REG_BADPAT, /* malformed \P or \p sequence */
REG_BADPAT, /* unknown property name after \P or \p */
REG_BADPAT, /* subpattern name is too long (maximum 32 characters) */
REG_BADPAT, /* too many named subpatterns (maximum 10,000) */
/* 50 */
REG_BADPAT, /* repeated subpattern is too long */
REG_BADPAT, /* octal value is greater than \377 (not in UTF-8 mode) */
REG_BADPAT, /* internal error: overran compiling workspace */
REG_BADPAT, /* internal error: previously-checked referenced subpattern not found */
REG_BADPAT, /* DEFINE group contains more than one branch */
/* 55 */
REG_BADPAT, /* repeating a DEFINE group is not allowed */
REG_INVARG, /* inconsistent NEWLINE options */
REG_BADPAT, /* \g is not followed followed by an (optionally braced) non-zero number */
REG_BADPAT, /* a numbered reference must not be zero */
REG_BADPAT, /* an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT) */
/* 60 */
REG_BADPAT, /* (*VERB) not recognized */
REG_BADPAT, /* number is too big */
REG_BADPAT, /* subpattern name expected */
REG_BADPAT, /* digit expected after (?+ */
REG_BADPAT, /* ] is an invalid data character in JavaScript compatibility mode */
/* 65 */
REG_BADPAT, /* different names for subpatterns of the same number are not allowed */
REG_BADPAT, /* (*MARK) must have an argument */
REG_INVARG, /* this version of PCRE is not compiled with PCRE_UCP support */
REG_BADPAT, /* \c must be followed by an ASCII character */
REG_BADPAT, /* \k is not followed by a braced, angle-bracketed, or quoted name */
/* 70 */
REG_BADPAT, /* internal error: unknown opcode in find_fixedlength() */
REG_BADPAT, /* \N is not supported in a class */
REG_BADPAT, /* too many forward references */
REG_BADPAT, /* disallowed UTF-8/16/32 code point (>= 0xd800 && <= 0xdfff) */
REG_BADPAT, /* invalid UTF-16 string (should not occur) */
/* 75 */
REG_BADPAT, /* overlong MARK name */
REG_BADPAT, /* character value in \u.... sequence is too large */
REG_BADPAT, /* invalid UTF-32 string (should not occur) */
REG_BADPAT, /* setting UTF is disabled by the application */
REG_BADPAT, /* non-hex character in \\x{} (closing brace missing?) */
/* 80 */
REG_BADPAT, /* non-octal character in \o{} (closing brace missing?) */
REG_BADPAT, /* missing opening brace after \o */
REG_BADPAT, /* parentheses too deeply nested */
REG_BADPAT, /* invalid range in character class */
REG_BADPAT, /* group name must start with a non-digit */
/* 85 */
REG_BADPAT /* parentheses too deeply nested (stack check) */
};

78
plugins/include/regex.inc
View File

@ -58,6 +58,24 @@
enum RegexError
{
REGEX_ERROR_NONE = 0, /* No error */
REGEX_ERROR_ASSERT = 1, /* internal error ? */
REGEX_ERROR_BADBR, /* invalid repeat counts in {} */
REGEX_ERROR_BADPAT, /* pattern error */
REGEX_ERROR_BADRPT, /* ? * + invalid */
REGEX_ERROR_EBRACE, /* unbalanced {} */
REGEX_ERROR_EBRACK, /* unbalanced [] */
REGEX_ERROR_ECOLLATE, /* collation error - not relevant */
REGEX_ERROR_ECTYPE, /* bad class */
REGEX_ERROR_EESCAPE, /* bad escape sequence */
REGEX_ERROR_EMPTY, /* empty expression */
REGEX_ERROR_EPAREN, /* unbalanced () */
REGEX_ERROR_ERANGE, /* bad range inside [] */
REGEX_ERROR_ESIZE, /* expression too big */
REGEX_ERROR_ESPACE, /* failed to get memory */
REGEX_ERROR_ESUBREG, /* bad back reference */
REGEX_ERROR_INVARG, /* bad argument */
REGEX_ERROR_NOMATCH = -1, /* No match was found */
REGEX_ERROR_NULL = -2,
REGEX_ERROR_BADOPTION = -3,
@ -110,23 +128,58 @@ methodmap Regex < Handle
// @param str The string to check.
// @param regex Regex Handle from CompileRegex()
// @param ret Error code, if applicable.
// @return Number of substrings found or -1 on failure.
// @param offset Offset in the string to start searching from. MatchOffset returns the offset of the match.
// @return Number of captures found or -1 on failure.
//
// @note Use the regex handle passed to this function to extract
// matches with GetRegexSubString().
public native int Match(const char[] str, RegexError &ret = REGEX_ERROR_NONE);
// matches with GetSubString().
public native int Match(const char[] str, RegexError &ret = REGEX_ERROR_NONE, int offset = 0);
// Gets all matches from a string against a pre-compiled regular expression pattern.
//
// @param str The string to check.
// @param regex Regex Handle from CompileRegex()
// @param ret Error code, if applicable.
// @return Number of matches found or -1 on failure.
//
// @note Use GetSubString() and loop from 0 -> totalmatches - 1.
public native int MatchAll(const char[] str, RegexError &ret = REGEX_ERROR_NONE);
// Returns a matched substring from a regex handle.
//
// Substring ids start at 0 and end at substrings-1, where substrings is the
// number returned by Regex.Match.
// Substring ids start at 0 and end at captures-1, where captures is the
// number returned by Regex.Match or Regex.CaptureCount.
//
// @param regex The regex handle to extract data from.
// @param str_id The index of the expression to get - starts at 0, and ends at substrings - 1.
// @param str_id The index of the expression to get - starts at 0, and ends at captures - 1.
// @param buffer The buffer to set to the matching substring.
// @param maxlen The maximum string length of the buffer.
// @param match Match to get the captures for - starts at 0, and ends at MatchCount() -1
// @return True if a substring was found, False on fail/error
public native bool GetSubString(int str_id, char[] buffer, int maxlen);
//
// @note str_id = 0 is the full captured string, anything else is the capture group index.
// if Regex.Match is used match can only be 0
public native bool GetSubString(int str_id, char[] buffer, int maxlen, int match = 0);
// Returns number of matches
//
// When using Match this is always 1 or 0 (unless an error occured)
// @return Total number of matches found.
public native int MatchCount();
// Returns number of captures for a match
//
// @param match Match to get the number of captures for. Match starts at 0, and ends at MatchCount() -1
// @return Number of captures in the match.
//
// @note Use GetSubString() and loop from 1 -> captures -1 for str_id to get all captures
public native int CaptureCount(int match = 0);
// Returns the string offset of a match.
//
// @param match Match to get the offset of. Match starts at 0, and ends at MatchCount() -1
// @return Offset of the match in the string.
public native int MatchOffset(int match = 0)
};
/**
@ -149,7 +202,7 @@ native Regex CompileRegex(const char[] pattern, int flags = 0, char[] error="",
* @param str The string to check.
* @param regex Regex Handle from CompileRegex()
* @param ret Error code, if applicable.
* @return Number of substrings found or -1 on failure.
* @return Number of captures found or -1 on failure.
*
* @note Use the regex handle passed to this function to extract
* matches with GetRegexSubString().
@ -158,14 +211,17 @@ native int MatchRegex(Handle regex, const char[] str, RegexError &ret = REGEX_ER
/**
* Returns a matched substring from a regex handle.
* Substring ids start at 0 and end at substrings-1, where substrings is the number returned
* by MatchRegex
* Substring ids start at 0 and end at captures-1, where captures is the number returned
* by MatchRegex.
*
* @param regex The regex handle to extract data from.
* @param str_id The index of the expression to get - starts at 0, and ends at substrings - 1.
* @param str_id The index of the expression to get - starts at 0, and ends at captures - 1.
* @param buffer The buffer to set to the matching substring.
* @param maxlen The maximum string length of the buffer.
* @return True if a substring was found, False on fail/error
*
* @note str_id = 0 is the full captured string, anything else is the capture group index.
*
*/
native bool GetRegexSubString(Handle regex, int str_id, char[] buffer, int maxlen);