<?php

> /**
// security - hide paths
>  * Helper functions.
if (!defined('ADODB_DIR')) die();
>  *

>  * Less commonly used functions are placed here to reduce size of adodb.inc.php.
global $ADODB_INCLUDED_LIB;
>  *
$ADODB_INCLUDED_LIB = 1;
>  * This file is part of ADOdb, a Database Abstraction Layer library for PHP.

>  *
/*
>  * @package ADOdb
  @version   v5.21.0  2021-02-27
>  * @link https://adodb.org Project's web site and documentation
  @copyright (c) 2000-2013 John Lim (jlim#natsoft.com). All rights reserved.
>  * @link https://github.com/ADOdb/ADOdb Source code and issue tracker
  @copyright (c) 2014      Damien Regad, Mark Newnham and the ADOdb community
>  *
  Released under both BSD license and Lesser GPL library license.
>  * The ADOdb Library is dual-licensed, released under both the BSD 3-Clause
  Whenever there is any discrepancy between the two licenses,
>  * and the GNU Lesser General Public Licence (LGPL) v2.1 or, at your option,
  the BSD license will take precedence. See License.txt.
>  * any later version. This means you can use it in proprietary products.
  Set tabs to 4 for best viewing.
>  * See the LICENSE.md file distributed with this source code for details.

>  * @license BSD-3-Clause
  Less commonly used functions are placed here to reduce size of adodb.inc.php.
>  * @license LGPL-2.1-or-later
*/
>  *

>  * @copyright 2000-2013 John Lim
function adodb_strip_order_by($sql)
>  * @copyright 2014 Damien Regad, Mark Newnham and the ADOdb community
{
>  */
	$rez = preg_match_all('/(\sORDER\s+BY\s(?:[^)](?!LIMIT))*)/is', $sql, $arr);
> 



< /*
<   @version   v5.21.0  2021-02-27
<   @copyright (c) 2000-2013 John Lim (jlim#natsoft.com). All rights reserved.
<   @copyright (c) 2014      Damien Regad, Mark Newnham and the ADOdb community
<   Released under both BSD license and Lesser GPL library license.
<   Whenever there is any discrepancy between the two licenses,
<   the BSD license will take precedence. See License.txt.
<   Set tabs to 4 for best viewing.
< 
<   Less commonly used functions are placed here to reduce size of adodb.inc.php.



> /**
>  * Strip the ORDER BY clause from the outer SELECT.
>  *
>  * @param string $sql
>  *
>  * @return string



< 



< 	$rez = preg_match_all('/(\sORDER\s+BY\s(?:[^)](?!LIMIT))*)/is', $sql, $arr);
< 	if ($arr) 
< 	{
< 		$tmp = array_pop($arr);
< 		$arr = [1=>array_pop($tmp)];
< 	}
< 	if ($arr)
< 		if (strpos($arr[1], '(') !== false) {
< 			$at = strpos($sql, $arr[1]);
< 			$cntin = 0;
< 			for ($i=$at, $max=strlen($sql); $i < $max; $i++) {
< 				$ch = $sql[$i];
< 				if ($ch == '(') {
< 					$cntin += 1;
< 				} elseif($ch == ')') {
< 					$cntin -= 1;
< 					if ($cntin < 0) {
< 						break;



> 	$num = preg_match_all('/(\sORDER\s+BY\s(?:[^)](?!LIMIT))*)/is', $sql, $matches, PREG_OFFSET_CAPTURE);
> 	if ($num) {
> 		// Get the last match
> 		list($last_order_by, $offset) = array_pop($matches[1]);
> 
> 		// If we find a ')' after the last order by, then it belongs to a
> 		// sub-query, not the outer SQL statement and should not be stripped
> 		if (strpos($sql, ')', $offset) === false) {
> 			$sql = str_replace($last_order_by, '', $sql);



< 			}
< 			$sql = substr($sql,0,$at).substr($sql,$i);
< 		} else {
< 			$sql = str_replace($arr[1], '', $sql);
< 		}
< 



< if (false) {
< 	$sql = 'select * from (select a from b order by a(b),b(c) desc)';
< 	$sql = '(select * from abc order by 1)';
< 	die(adodb_strip_order_by($sql));
< }
< 
< function adodb_probetypes(&$array,&$types,$probe=8)



> function adodb_probetypes($array,&$types,$probe=8)


	for ($j=0;$j < $max; $j++) {
		$row = $array[$j];
		if (!$row) break;
		$i = -1;
		foreach($row as $v) {
			$i += 1;

			if (isset($types[$i]) && $types[$i]=='C') continue;

			//print " ($i ".$types[$i]. "$v) ";
			$v = trim($v);

			if (!preg_match('/^[+-]{0,1}[0-9\.]+$/',$v)) {
				$types[$i] = 'C'; // once C, always C

				continue;
			}
			if ($j == 0) {
			// If empty string, we presume is character
			// test for integer for 1st row only
			// after that it is up to testing other rows to prove
			// that it is not an integer
				if (strlen($v) == 0) $types[$i] = 'C';
				if (strpos($v,'.') !== false) $types[$i] = 'N';
				else  $types[$i] = 'I';
				continue;
			}

			if (strpos($v,'.') !== false) $types[$i] = 'N';

		}
	}

}


< function  adodb_transpose(&$arr, &$newarr, &$hdr, &$fobjs)



> function adodb_transpose(&$arr, &$newarr, &$hdr, $fobjs)


{
	$oldX = sizeof(reset($arr));
	$oldY = sizeof($arr);

	if ($hdr) {
		$startx = 1;
		$hdr = array('Fields');
		for ($y = 0; $y < $oldY; $y++) {
			$hdr[] = $arr[$y][0];
		}
	} else
		$startx = 0;

	for ($x = $startx; $x < $oldX; $x++) {
		if ($fobjs) {
			$o = $fobjs[$x];
			$newarr[] = array($o->name);
		} else
			$newarr[] = array();

		for ($y = 0; $y < $oldY; $y++) {
			$newarr[$x-$startx][] = $arr[$y][$x];
		}
	}
}



< function _adodb_replace(&$zthis, $table, $fieldArray, $keyCol, $autoQuote, $has_autoinc)



> function _adodb_replace($zthis, $table, $fieldArray, $keyCol, $autoQuote, $has_autoinc)


{
	// Add Quote around table name to support use of spaces / reserved keywords
	$table=sprintf('%s%s%s', $zthis->nameQuote,$table,$zthis->nameQuote);

	if (count($fieldArray) == 0) return 0;

	if (!is_array($keyCol)) {
		$keyCol = array($keyCol);
	}
	$uSet = '';
	foreach($fieldArray as $k => $v) {
		if ($v === null) {
			$v = 'NULL';
			$fieldArray[$k] = $v;
		} else if ($autoQuote && /*!is_numeric($v) /*and strncmp($v,"'",1) !== 0 -- sql injection risk*/ strcasecmp($v,$zthis->null2null)!=0) {
			$v = $zthis->qstr($v);
			$fieldArray[$k] = $v;
		}
		if (in_array($k,$keyCol)) continue; // skip UPDATE if is key

		// Add Quote around column name to support use of spaces / reserved keywords
		$uSet .= sprintf(',%s%s%s=%s',$zthis->nameQuote,$k,$zthis->nameQuote,$v);
	}
	$uSet = ltrim($uSet, ',');

	// Add Quote around column name in where clause
	$where = '';
	foreach ($keyCol as $v) {
		if (isset($fieldArray[$v])) {
			$where .= sprintf(' and %s%s%s=%s ', $zthis->nameQuote,$v,$zthis->nameQuote,$fieldArray[$v]);
		}
	}
	if ($where) {
		$where = substr($where, 5);
	}

	if ($uSet && $where) {
		$update = "UPDATE $table SET $uSet WHERE $where";
		$rs = $zthis->Execute($update);

		if ($rs) {
			if ($zthis->poorAffectedRows) {
				// The Select count(*) wipes out any errors that the update would have returned.
				// PHPLens Issue No: 5696
				if ($zthis->ErrorNo()<>0) return 0;

				// affected_rows == 0 if update field values identical to old values
				// for mysql - which is silly.
				$cnt = $zthis->GetOne("select count(*) from $table where $where");
				if ($cnt > 0) return 1; // record already exists
			} else {
				if (($zthis->Affected_Rows()>0)) return 1;
			}
		} else
			return 0;
	}

	$iCols = $iVals = '';
	foreach($fieldArray as $k => $v) {
		if ($has_autoinc && in_array($k,$keyCol)) continue; // skip autoinc col

		// Add Quote around Column Name
		$iCols .= sprintf(',%s%s%s',$zthis->nameQuote,$k,$zthis->nameQuote);
		$iVals .= ",$v";
	}
	$iCols = ltrim($iCols, ',');
	$iVals = ltrim($iVals, ',');

	$insert = "INSERT INTO $table ($iCols) VALUES ($iVals)";
	$rs = $zthis->Execute($insert);
	return ($rs) ? 2 : 0;
}


< function _adodb_getmenu(&$zthis, $name,$defstr='',$blank1stItem=true,$multiple=false,



> function _adodb_getmenu($zthis, $name,$defstr='',$blank1stItem=true,$multiple=false,


			$size=0, $selectAttr='',$compareFields0=true)
{
	global $ADODB_FETCH_MODE;

	$s = _adodb_getmenu_select($name, $defstr, $blank1stItem, $multiple, $size, $selectAttr);

	$hasvalue = $zthis->FieldCount() > 1;
	if (!$hasvalue) {
		$compareFields0 = true;
	}

	$value = '';
	while(!$zthis->EOF) {
		$zval = rtrim(reset($zthis->fields));

		if ($blank1stItem && $zval == "") {
			$zthis->MoveNext();
			continue;
		}

		if ($hasvalue) {
			if ($ADODB_FETCH_MODE == ADODB_FETCH_ASSOC) {
				// Get 2nd field's value regardless of its name
				$zval2 = current(array_slice($zthis->fields, 1, 1));
			} else {
				// With NUM or BOTH fetch modes, we have a numeric index
				$zval2 = $zthis->fields[1];
			}
			$zval2 = trim($zval2);
			$value = 'value="' . htmlspecialchars($zval2) . '"';
		}


> 		/** @noinspection PhpUndefinedVariableInspection */


		$s .= _adodb_getmenu_option($defstr, $compareFields0 ? $zval : $zval2, $value, $zval);

		$zthis->MoveNext();
	} // while

	return $s ."\n</select>\n";
}


< function _adodb_getmenu_gp(&$zthis, $name,$defstr='',$blank1stItem=true,$multiple=false,



> function _adodb_getmenu_gp($zthis, $name,$defstr='',$blank1stItem=true,$multiple=false,


			$size=0, $selectAttr='',$compareFields0=true)
{
	global $ADODB_FETCH_MODE;

	$s = _adodb_getmenu_select($name, $defstr, $blank1stItem, $multiple, $size, $selectAttr);

	$hasvalue = $zthis->FieldCount() > 1;
	$hasgroup = $zthis->FieldCount() > 2;
	if (!$hasvalue) {
		$compareFields0 = true;
	}

	$value = '';
	$optgroup = null;
	$firstgroup = true;
	while(!$zthis->EOF) {
		$zval = rtrim(reset($zthis->fields));
		$group = '';

		if ($blank1stItem && $zval=="") {
			$zthis->MoveNext();
			continue;
		}

		if ($hasvalue) {
			if ($ADODB_FETCH_MODE == ADODB_FETCH_ASSOC) {
				// Get 2nd field's value regardless of its name
				$fields = array_slice($zthis->fields, 1);
				$zval2 = current($fields);
				if ($hasgroup) {
					$group = trim(next($fields));
				}
			} else {
				// With NUM or BOTH fetch modes, we have a numeric index
				$zval2 = $zthis->fields[1];
				if ($hasgroup) {
					$group = trim($zthis->fields[2]);
				}
			}
			$zval2 = trim($zval2);
			$value = "value='".htmlspecialchars($zval2)."'";
		}

		if ($optgroup != $group) {
			$optgroup = $group;
			if ($firstgroup) {
				$firstgroup = false;
			} else {
				$s .="\n</optgroup>";
			}
			$s .="\n<optgroup label='". htmlspecialchars($group) ."'>";
		}


> 		/** @noinspection PhpUndefinedVariableInspection */


		$s .= _adodb_getmenu_option($defstr, $compareFields0 ? $zval : $zval2, $value, $zval);

		$zthis->MoveNext();
	} // while

	// closing last optgroup
	if($optgroup != null) {
		$s .= "\n</optgroup>";
	}
	return $s ."\n</select>\n";
}

/**
 * Generate the opening SELECT tag for getmenu functions.
 *
 * ADOdb internal function, used by _adodb_getmenu() and _adodb_getmenu_gp().
 *
 * @param string $name
 * @param string $defstr
 * @param bool   $blank1stItem
 * @param bool   $multiple
 * @param int    $size
 * @param string $selectAttr
 *
 * @return string HTML
 */
function _adodb_getmenu_select($name, $defstr = '', $blank1stItem = true,
							   $multiple = false, $size = 0, $selectAttr = '')
{
	if ($multiple || is_array($defstr)) {
		if ($size == 0 ) {
			$size = 5;
		}
		$attr = ' multiple size="' . $size . '"';
		if (!strpos($name,'[]')) {
			$name .= '[]';
		}
	} elseif ($size) {
		$attr = ' size="' . $size . '"';
	} else {
		$attr = '';
	}

	$html = '<select name="' . $name . '"' . $attr . ' ' . $selectAttr . '>';
	if ($blank1stItem) {
		if (is_string($blank1stItem))  {
			$barr = explode(':',$blank1stItem);
			if (sizeof($barr) == 1) {
				$barr[] = '';
			}
			$html .= "\n<option value=\"" . $barr[0] . "\">" . $barr[1] . "</option>";
		} else {
			$html .= "\n<option></option>";
		}
	}

	return $html;
}

/**
 * Print the OPTION tags for getmenu functions.
 *
 * ADOdb internal function, used by _adodb_getmenu() and _adodb_getmenu_gp().
 *
 * @param string $defstr  Default values
 * @param string $compare Value to compare against defaults
 * @param string $value   Ready-to-print `value="xxx"` (or empty) string
 * @param string $display Display value
 *
 * @return string HTML
 */
function _adodb_getmenu_option($defstr, $compare, $value, $display)
{
	if (   is_array($defstr) && in_array($compare, $defstr)
		|| !is_array($defstr) && strcasecmp($compare, $defstr) == 0
	) {
		$selected = ' selected="selected"';
	} else {
		$selected = '';
	}

	return "\n<option $value$selected>" . htmlspecialchars($display) . '</option>';
}

/*
	Count the number of records this sql statement will return by using
	query rewriting heuristics...

	Does not work with UNIONs, except with postgresql and oracle.

	Usage:

	$conn->Connect(...);
	$cnt = _adodb_getcount($conn, $sql);

*/

< function _adodb_getcount(&$zthis, $sql,$inputarr=false,$secs2cache=0)



> function _adodb_getcount($zthis, $sql,$inputarr=false,$secs2cache=0)


{
	$qryRecs = 0;


< 	 if (!empty($zthis->_nestedSQL) || preg_match("/^\s*SELECT\s+DISTINCT/is", $sql) ||
< 	 	preg_match('/\s+GROUP\s+BY\s+/is',$sql) ||
< 		preg_match('/\s+UNION\s+/is',$sql)) {



> 	/*
> 	* These databases require a "SELECT * FROM (SELECT" type
> 	* statement to have an alias for the result
> 	*/
> 	$requiresAlias = '';
> 	$requiresAliasArray = array('postgres9','postgres','mysql','mysqli','mssql','mssqlnative','sqlsrv');
> 	if (in_array($zthis->databaseType,$requiresAliasArray)
> 		|| in_array($zthis->dsnType,$requiresAliasArray)
> 	) {
> 		$requiresAlias = '_ADODB_ALIAS_';
> 	}




> 	if (!empty($zthis->_nestedSQL)
		$rewritesql = adodb_strip_order_by($sql);
> 		|| preg_match("/^\s*SELECT\s+DISTINCT/is", $sql)

> 		|| preg_match('/\s+GROUP\s+BY\s+/is',$sql)
		// ok, has SELECT DISTINCT or GROUP BY so see if we can use a table alias
> 		|| preg_match('/\s+UNION\s+/is',$sql)
		// but this is only supported by oracle and postgresql...
> 	) {


		if ($zthis->dataProvider == 'oci8') {
			// Allow Oracle hints to be used for query optimization, Chris Wrye
			if (preg_match('#/\\*+.*?\\*\\/#', $sql, $hint)) {
				$rewritesql = "SELECT ".$hint[0]." COUNT(*) FROM (".$rewritesql.")";
			} else
				$rewritesql = "SELECT COUNT(*) FROM (".$rewritesql.")";

< 
< 		} else if (strncmp($zthis->databaseType,'postgres',8) == 0
< 			|| strncmp($zthis->databaseType,'mysql',5) == 0
< 		|| strncmp($zthis->databaseType,'mssql',5) == 0
< 			|| strncmp($zthis->dsnType,'sqlsrv',5) == 0
< 			|| strncmp($zthis->dsnType,'mssql',5) == 0
< 		){
< 			$rewritesql = "SELECT COUNT(*) FROM ($rewritesql) _ADODB_ALIAS_";


		} else {

< 			$rewritesql = "SELECT COUNT(*) FROM ($rewritesql)";



> 			$rewritesql = "SELECT COUNT(*) FROM ($rewritesql) $requiresAlias";


		}

> 


	} else {

< 		// now replace SELECT ... FROM with SELECT COUNT(*) FROM
< 		if ( strpos($sql, '_ADODB_COUNT') !== FALSE ) {
< 			$rewritesql = preg_replace('/^\s*?SELECT\s+_ADODB_COUNT(.*)_ADODB_COUNT\s/is','SELECT COUNT(*) ',$sql);
< 		} else {
< 			$rewritesql = preg_replace('/^\s*SELECT\s.*\s+FROM\s/Uis','SELECT COUNT(*) FROM ',$sql);



> 		// Replace 'SELECT ... FROM' with 'SELECT COUNT(*) FROM'
> 		// Parse the query one char at a time starting after the SELECT
> 		// to find the FROM clause's position, ignoring any sub-queries.
> 		$start = stripos($sql, 'SELECT') + 7;
> 		if ($start === false) {
> 			// Not a SELECT statement - probably should trigger an exception here
> 			return 0;


		}

> 		$len = strlen($sql);
		// fix by alexander zhukov, alex#unipack.ru, because count(*) and 'order by' fails
> 		$numParentheses = 0;
		// with mssql, access and postgresql. Also a good speedup optimization - skips sorting!
> 		for ($pos = $start; $pos < $len; $pos++) {
		// also see PHPLens Issue No: 12752
> 			switch ($sql[$pos]) {
		$rewritesql = adodb_strip_order_by($rewritesql);
> 				case '(': $numParentheses++; continue 2;
	}
> 				case ')': $numParentheses--; continue 2;

> 			}
	if (isset($rewritesql) && $rewritesql != $sql) {
> 			// Ignore whatever is between parentheses (sub-queries)
		if (preg_match('/\sLIMIT\s+[0-9]+/i',$sql,$limitarr)) $rewritesql .= $limitarr[0];
> 			if ($numParentheses > 0) {

> 				continue;
		if ($secs2cache) {
> 			}
			// we only use half the time of secs2cache because the count can quickly
> 			// Exit loop if 'FROM' keyword was found
			// become inaccurate if new records are added
> 			if (strtoupper(substr($sql, $pos, 4)) == 'FROM') {
			$qryRecs = $zthis->CacheGetOne($secs2cache/2,$rewritesql,$inputarr);
> 				break;

> 			}
		} else {
> 		}
			$qryRecs = $zthis->GetOne($rewritesql,$inputarr);
> 		$rewritesql = 'SELECT COUNT(*) ' . substr($sql, $pos);
	  	}
> 



< 		if (preg_match('/\sLIMIT\s+[0-9]+/i',$sql,$limitarr)) $rewritesql .= $limitarr[0];



> 		if (preg_match('/\sLIMIT\s+[0-9]+/i',$sql,$limitarr)) {
> 			$rewritesql .= $limitarr[0];
> 		}


	}

> 


	//--------------------------------------------
	// query rewrite failed - so try slower way...


< 


	// strip off unneeded ORDER BY if no UNION

< 	if (preg_match('/\s*UNION\s*/is', $sql)) $rewritesql = $sql;
< 	else $rewritesql = $rewritesql = adodb_strip_order_by($sql);



> 	if (preg_match('/\s*UNION\s*/is', $sql)) {
> 		$rewritesql = $sql;
> 	} else {
> 		$rewritesql = adodb_strip_order_by($sql);
> 	}




< 	if (preg_match('/\sLIMIT\s+[0-9]+/i',$sql,$limitarr)) $rewritesql .= $limitarr[0];



> 	if (preg_match('/\sLIMIT\s+[0-9]+/i',$sql,$limitarr)) {
> 		$rewritesql .= $limitarr[0];
> 	}



	if ($secs2cache) {
		$rstest = $zthis->CacheExecute($secs2cache,$rewritesql,$inputarr);
		if (!$rstest) $rstest = $zthis->CacheExecute($secs2cache,$sql,$inputarr);
	} else {
		$rstest = $zthis->Execute($rewritesql,$inputarr);
		if (!$rstest) $rstest = $zthis->Execute($sql,$inputarr);
	}
	if ($rstest) {
		$qryRecs = $rstest->RecordCount();
		if ($qryRecs == -1) {
			// some databases will return -1 on MoveLast() - change to MoveNext()
			while(!$rstest->EOF) {
				$rstest->MoveNext();
			}
			$qryRecs = $rstest->_currentRow;
		}
		$rstest->Close();
		if ($qryRecs == -1) return 0;
	}
	return $qryRecs;
}

/*
 	Code originally from "Cornel G" <conyg@fx.ro>

	This code might not work with SQL that has UNION in it

	Also if you are using CachePageExecute(), there is a strong possibility that
	data will get out of synch. use CachePageExecute() only with tables that
	rarely change.
*/

< function _adodb_pageexecute_all_rows(&$zthis, $sql, $nrows, $page,



> function _adodb_pageexecute_all_rows($zthis, $sql, $nrows, $page,


						$inputarr=false, $secs2cache=0)
{
	$atfirstpage = false;
	$atlastpage = false;

< 	$lastpageno=1;



	// If an invalid nrows is supplied,
	// we assume a default value of 10 rows per page
	if (!isset($nrows) || $nrows <= 0) $nrows = 10;


< 	$qryRecs = false; //count records for no offset
< 


	$qryRecs = _adodb_getcount($zthis,$sql,$inputarr,$secs2cache);
	$lastpageno = (int) ceil($qryRecs / $nrows);
	$zthis->_maxRecordCount = $qryRecs;


< 
< 


	// ***** Here we check whether $page is the last page or
	// whether we are trying to retrieve
	// a page number greater than the last page number.
	if ($page >= $lastpageno) {
		$page = $lastpageno;
		$atlastpage = true;
	}

	// If page number <= 1, then we are at the first page
	if (empty($page) || $page <= 1) {
		$page = 1;
		$atfirstpage = true;
	}

	// We get the data we want
	$offset = $nrows * ($page-1);
	if ($secs2cache > 0)
		$rsreturn = $zthis->CacheSelectLimit($secs2cache, $sql, $nrows, $offset, $inputarr);
	else
		$rsreturn = $zthis->SelectLimit($sql, $nrows, $offset, $inputarr, $secs2cache);


	// Before returning the RecordSet, we set the pagination properties we need
	if ($rsreturn) {
		$rsreturn->_maxRecordCount = $qryRecs;
		$rsreturn->rowsPerPage = $nrows;
		$rsreturn->AbsolutePage($page);
		$rsreturn->AtFirstPage($atfirstpage);
		$rsreturn->AtLastPage($atlastpage);
		$rsreturn->LastPageNo($lastpageno);
	}
	return $rsreturn;
}

// Iván Oliva version

< function _adodb_pageexecute_no_last_page(&$zthis, $sql, $nrows, $page, $inputarr=false, $secs2cache=0)



> function _adodb_pageexecute_no_last_page($zthis, $sql, $nrows, $page, $inputarr=false, $secs2cache=0)


{

< 


	$atfirstpage = false;
	$atlastpage = false;

	if (!isset($page) || $page <= 1) {
		// If page number <= 1, then we are at the first page
		$page = 1;
		$atfirstpage = true;
	}
	if ($nrows <= 0) {
		// If an invalid nrows is supplied, we assume a default value of 10 rows per page
		$nrows = 10;
	}

	$pagecounteroffset = ($page * $nrows) - $nrows;

	// To find out if there are more pages of rows, simply increase the limit or
	// nrows by 1 and see if that number of records was returned. If it was,
	// then we know there is at least one more page left, otherwise we are on
	// the last page. Therefore allow non-Count() paging with single queries
	// rather than three queries as was done before.
	$test_nrows = $nrows + 1;
	if ($secs2cache > 0) {
		$rsreturn = $zthis->CacheSelectLimit($secs2cache, $sql, $nrows, $pagecounteroffset, $inputarr);
	} else {
		$rsreturn = $zthis->SelectLimit($sql, $test_nrows, $pagecounteroffset, $inputarr, $secs2cache);
	}

	// Now check to see if the number of rows returned was the higher value we asked for or not.
	if ( $rsreturn->_numOfRows == $test_nrows ) {
		// Still at least 1 more row, so we are not on last page yet...
		// Remove the last row from the RS.
		$rsreturn->_numOfRows = ( $rsreturn->_numOfRows - 1 );
	} elseif ( $rsreturn->_numOfRows == 0 && $page > 1 ) {
		// Likely requested a page that doesn't exist, so need to find the last
		// page and return it. Revert to original method and loop through pages
		// until we find some data...
		$pagecounter = $page + 1;

< 		$pagecounteroffset = ($pagecounter * $nrows) - $nrows;



		$rstest = $rsreturn;
		if ($rstest) {
			while ($rstest && $rstest->EOF && $pagecounter > 0) {
				$atlastpage = true;
				$pagecounter--;
				$pagecounteroffset = $nrows * ($pagecounter - 1);
				$rstest->Close();
				if ($secs2cache>0) {
					$rstest = $zthis->CacheSelectLimit($secs2cache, $sql, $nrows, $pagecounteroffset, $inputarr);
				}
				else {
					$rstest = $zthis->SelectLimit($sql, $nrows, $pagecounteroffset, $inputarr, $secs2cache);
				}
			}
			if ($rstest) $rstest->Close();
		}
		if ($atlastpage) {
			// If we are at the last page or beyond it, we are going to retrieve it
			$page = $pagecounter;
			if ($page == 1) {
				// We have to do this again in case the last page is the same as
				// the first page, that is, the recordset has only 1 page.
				$atfirstpage = true;
			}
		}
		// We get the data we want
		$offset = $nrows * ($page-1);
		if ($secs2cache > 0) {
			$rsreturn = $zthis->CacheSelectLimit($secs2cache, $sql, $nrows, $offset, $inputarr);
		}
		else {
			$rsreturn = $zthis->SelectLimit($sql, $nrows, $offset, $inputarr, $secs2cache);
		}
	} elseif ( $rsreturn->_numOfRows < $test_nrows ) {
		// Rows is less than what we asked for, so must be at the last page.
		$atlastpage = true;
	}

	// Before returning the RecordSet, we set the pagination properties we need
	if ($rsreturn) {
		$rsreturn->rowsPerPage = $nrows;
		$rsreturn->AbsolutePage($page);
		$rsreturn->AtFirstPage($atfirstpage);
		$rsreturn->AtLastPage($atlastpage);
	}
	return $rsreturn;
}


< function _adodb_getupdatesql(&$zthis, &$rs, $arrFields, $forceUpdate=false, $force=2)



> /**
>  * Performs case conversion and quoting of the given field name.
>  *
>  * See Global variable $ADODB_QUOTE_FIELDNAMES.
>  *
>  * @param ADOConnection $zthis
>  * @param string $fieldName
>  *
>  * @return string Quoted field name
>  */
> function _adodb_quote_fieldname($zthis, $fieldName)


{
	global $ADODB_QUOTE_FIELDNAMES;


> 	// Case conversion - defaults to UPPER
	if (!$rs) {
> 	$case = is_bool($ADODB_QUOTE_FIELDNAMES) ? 'UPPER' : $ADODB_QUOTE_FIELDNAMES;
		printf(ADODB_BAD_RS,'GetUpdateSQL');
> 	switch ($case) {
		return false;
> 		case 'LOWER':
	}
> 			$fieldName = strtolower($fieldName);

> 			break;
	$fieldUpdatedCount = 0;
> 		case 'NATIVE':
	if (is_array($arrFields))
> 			// Do nothing
		$arrFields = array_change_key_case($arrFields,CASE_UPPER);
> 			break;

> 		case 'UPPER':
	$hasnumeric = isset($rs->fields[0]);
> 		case 'BRACKETS':
	$setFields = '';
> 		default:

> 			$fieldName = strtoupper($fieldName);
	// Loop through all of the fields in the recordset
> 			break;
	for ($i=0, $max=$rs->fieldCount(); $i < $max; $i++) {
> 	}
		// Get the field from the recordset
> 
		$field = $rs->fetchField($i);
> 	// Quote field if requested, or necessary (field contains space)

> 	if ($ADODB_QUOTE_FIELDNAMES || strpos($fieldName, ' ') !== false ) {
		// If the recordset field is one
> 		if ($ADODB_QUOTE_FIELDNAMES === 'BRACKETS') {
		// of the fields passed in then process.
> 			return $zthis->leftBracket . $fieldName . $zthis->rightBracket;
		$upperfname = strtoupper($field->name);
> 		} else {
		if (adodb_key_exists($upperfname, $arrFields, $force)) {
> 			return $zthis->nameQuote . $fieldName . $zthis->nameQuote;

> 		}
			// If the existing field value in the recordset
> 	} else {
			// is different from the value passed in then
> 		return $fieldName;
			// go ahead and append the field name and new value to
> 	}
			// the update query.
> }

> 
			if ($hasnumeric) $val = $rs->fields[$i];
> function _adodb_getupdatesql(&$zthis, $rs, $arrFields, $forceUpdate=false, $force=2)
			else if (isset($rs->fields[$upperfname])) $val = $rs->fields[$upperfname];
> {


			else if (isset($rs->fields[$field->name])) $val =  $rs->fields[$field->name];
			else if (isset($rs->fields[strtolower($upperfname)])) $val =  $rs->fields[strtolower($upperfname)];
			else $val = '';


< 			if ($forceUpdate || strcmp($val, $arrFields[$upperfname])) {



> 			if ($forceUpdate || $val !== $arrFields[$upperfname]) {


				// Set the counter for the number of fields that will be updated.
				$fieldUpdatedCount++;

				// Based on the datatype of the field
				// Format the value properly for the database
				$type = $rs->metaType($field->type);

				if ($type == 'null') {
					$type = 'C';
				}


< 				if ((strpos($upperfname,' ') !== false) || ($ADODB_QUOTE_FIELDNAMES)) {
< 					switch ($ADODB_QUOTE_FIELDNAMES) {
< 						case 'BRACKETS':
< 							$fnameq = $zthis->leftBracket.$upperfname.$zthis->rightBracket;break;
< 						case 'LOWER':
< 							$fnameq = $zthis->nameQuote.strtolower($field->name).$zthis->nameQuote;break;
< 						case 'NATIVE':
< 							$fnameq = $zthis->nameQuote.$field->name.$zthis->nameQuote;break;
< 						case 'UPPER':
< 						default:
< 							$fnameq = $zthis->nameQuote.$upperfname.$zthis->nameQuote;break;
< 					}
< 				} else {
< 					$fnameq = $upperfname;
< 				}



> 				$fnameq = _adodb_quote_fieldname($zthis, $field->name);



				//********************************************************//
				if (is_null($arrFields[$upperfname])
					|| (empty($arrFields[$upperfname]) && strlen($arrFields[$upperfname]) == 0)
					|| $arrFields[$upperfname] === $zthis->null2null
					) {

					switch ($force) {

						//case 0:
						//	// Ignore empty values. This is already handled in "adodb_key_exists" function.
						//	break;

						case 1:
							// set null
							$setFields .= $fnameq . " = null, ";
							break;

						case 2:
							// set empty
							$arrFields[$upperfname] = "";
							$setFields .= _adodb_column_sql($zthis, 'U', $type, $upperfname, $fnameq, $arrFields);
							break;

						default:
						case 3:
							// set the value that was given in array, so you can give both null and empty values
							if (is_null($arrFields[$upperfname]) || $arrFields[$upperfname] === $zthis->null2null) {
								$setFields .= $fnameq . " = null, ";
							} else {
								$setFields .= _adodb_column_sql($zthis, 'U', $type, $upperfname, $fnameq, $arrFields);
							}
							break;

						case ADODB_FORCE_NULL_AND_ZERO:

							switch ($type) {
								case 'N':
								case 'I':
								case 'L':
									$setFields .= $fnameq . ' = 0, ';
									break;
								default:
									$setFields .= $fnameq . ' = null, ';
									break;
							}
							break;

					}
				//********************************************************//
				} else {
					// we do this so each driver can customize the sql for
					// DB specific column types.
					// Oracle needs BLOB types to be handled with a returning clause
					// postgres has special needs as well
					$setFields .= _adodb_column_sql($zthis, 'U', $type, $upperfname, $fnameq, $arrFields);
				}
			}
		}
	}

	// If there were any modified fields then build the rest of the update query.
	if ($fieldUpdatedCount > 0 || $forceUpdate) {
		// Get the table name from the existing query.
		if (!empty($rs->tableName)) {
			$tableName = $rs->tableName;
		} else {
			preg_match("/FROM\s+".ADODB_TABLE_REGEX."/is", $rs->sql, $tableName);
			$tableName = $tableName[1];
		}

		// Get the full where clause excluding the word "WHERE" from the existing query.
		preg_match('/\sWHERE\s(.*)/is', $rs->sql, $whereClause);

		$discard = false;
		// not a good hack, improvements?
		if ($whereClause) {

< 			#var_dump($whereClause);


			if (preg_match('/\s(ORDER\s.*)/is', $whereClause[1], $discard));
			else if (preg_match('/\s(LIMIT\s.*)/is', $whereClause[1], $discard));
			else if (preg_match('/\s(FOR UPDATE.*)/is', $whereClause[1], $discard));
			else preg_match('/\s.*(\) WHERE .*)/is', $whereClause[1], $discard); # see https://sourceforge.net/p/adodb/bugs/37/
		} else {
			$whereClause = array(false, false);
		}

		if ($discard) {
			$whereClause[1] = substr($whereClause[1], 0, strlen($whereClause[1]) - strlen($discard[1]));
		}

		$sql = 'UPDATE '.$tableName.' SET '.substr($setFields, 0, -2);
		if (strlen($whereClause[1]) > 0) {
			$sql .= ' WHERE '.$whereClause[1];
		}
		return $sql;
	} else {
		return false;
	}
}


< function adodb_key_exists($key, &$arr,$force=2)



> function adodb_key_exists($key, $arr,$force=2)


{
	if ($force<=0) {
		// the following is the old behaviour where null or empty fields are ignored
		return (!empty($arr[$key])) || (isset($arr[$key]) && strlen($arr[$key])>0);
	}

	if (isset($arr[$key])) 
		return true;
	## null check below
	return array_key_exists($key,$arr);
}

/**
 * There is a special case of this function for the oci8 driver.
 * The proper way to handle an insert w/ a blob in oracle requires
 * a returning clause with bind variables and a descriptor blob.
 *
 *
 */

< function _adodb_getinsertsql(&$zthis, &$rs, $arrFields, $force=2)



> function _adodb_getinsertsql(&$zthis, $rs, $arrFields, $force=2)


{
static $cacheRS = false;
static $cacheSig = 0;
static $cacheCols;

< 	global $ADODB_QUOTE_FIELDNAMES;



	$tableName = '';
	$values = '';
	$fields = '';

< 	$recordSet = null;


	if (is_array($arrFields))
		$arrFields = array_change_key_case($arrFields,CASE_UPPER);
	$fieldInsertedCount = 0;

	if (is_string($rs)) {
		//ok we have a table name
		//try and get the column info ourself.
		$tableName = $rs;

		//we need an object for the recordSet
		//because we have to call MetaType.
		//php can't do a $rsclass::MetaType()
		$rsclass = $zthis->rsPrefix.$zthis->databaseType;

< 		$recordSet = new $rsclass(-1,$zthis->fetchMode);



> 		$recordSet = new $rsclass(ADORecordSet::DUMMY_QUERY_ID, $zthis->fetchMode);


		$recordSet->connection = $zthis;

		if (is_string($cacheRS) && $cacheRS == $rs) {
			$columns = $cacheCols;
		} else {
			$columns = $zthis->MetaColumns( $tableName );
			$cacheRS = $tableName;
			$cacheCols = $columns;
		}
	} else if (is_subclass_of($rs, 'adorecordset')) {
		if (isset($rs->insertSig) && is_integer($cacheRS) && $cacheRS == $rs->insertSig) {
			$columns = $cacheCols;
		} else {

> 			$columns = [];


			for ($i=0, $max=$rs->FieldCount(); $i < $max; $i++)
				$columns[] = $rs->FetchField($i);
			$cacheRS = $cacheSig;
			$cacheCols = $columns;
			$rs->insertSig = $cacheSig++;
		}
		$recordSet = $rs;

	} else {
		printf(ADODB_BAD_RS,'GetInsertSQL');
		return false;
	}

	// Loop through all of the fields in the recordset
	foreach( $columns as $field ) {
		$upperfname = strtoupper($field->name);
		if (adodb_key_exists($upperfname,$arrFields,$force)) {
			$bad = false;

< 			if ((strpos($upperfname,' ') !== false) || ($ADODB_QUOTE_FIELDNAMES)) {
< 				switch ($ADODB_QUOTE_FIELDNAMES) {
< 				case 'BRACKETS':
< 					$fnameq = $zthis->leftBracket.$upperfname.$zthis->rightBracket;break;
< 				case 'LOWER':
< 					$fnameq = $zthis->nameQuote.strtolower($field->name).$zthis->nameQuote;break;
< 				case 'NATIVE':
< 					$fnameq = $zthis->nameQuote.$field->name.$zthis->nameQuote;break;
< 				case 'UPPER':
< 				default:
< 					$fnameq = $zthis->nameQuote.$upperfname.$zthis->nameQuote;break;
< 				}
< 			} else
< 				$fnameq = $upperfname;
< 



> 			$fnameq = _adodb_quote_fieldname($zthis, $field->name);


			$type = $recordSet->MetaType($field->type);

            /********************************************************/
            if (is_null($arrFields[$upperfname])
                || (empty($arrFields[$upperfname]) && strlen($arrFields[$upperfname]) == 0)
                || $arrFields[$upperfname] === $zthis->null2null

< 				)
<                {



> 			) {


                    switch ($force) {

                        case ADODB_FORCE_IGNORE: // we must always set null if missing
							$bad = true;
							break;

                        case ADODB_FORCE_NULL:
                            $values  .= "null, ";
                        break;

                        case ADODB_FORCE_EMPTY:
                            //Set empty
                            $arrFields[$upperfname] = "";
							$values .= _adodb_column_sql($zthis, 'I', $type, $upperfname, $fnameq, $arrFields);
                        break;

						default:
                        case ADODB_FORCE_VALUE:
                            //Set the value that was given in array, so you can give both null and empty values
							if (is_null($arrFields[$upperfname]) || $arrFields[$upperfname] === $zthis->null2null) {
								$values  .= "null, ";
							} else {
								$values .= _adodb_column_sql($zthis, 'I', $type, $upperfname, $fnameq, $arrFields);
             				}
              			break;

						case ADODB_FORCE_NULL_AND_ZERO:

< 							switch ($type)
< 							{



> 						switch ($type) {


								case 'N':
								case 'I':
								case 'L':
									$values .= '0, ';
									break;
								default:
									$values .= "null, ";
									break;
							}
						break;

             		} // switch

            /*********************************************************/
			} else {
				//we do this so each driver can customize the sql for
				//DB specific column types.
				//Oracle needs BLOB types to be handled with a returning clause
				//postgres has special needs as well
				$values .= _adodb_column_sql($zthis, 'I', $type, $upperfname, $fnameq, $arrFields);
			}


< 			if ($bad) continue;



> 			if ($bad) {
> 				continue;
> 			}


			// Set the counter for the number of fields that will be inserted.
			$fieldInsertedCount++;


< 


			// Get the name of the fields to insert
			$fields .= $fnameq . ", ";
		}
	}


	// If there were any inserted fields then build the rest of the insert query.
	if ($fieldInsertedCount <= 0)  return false;

	// Get the table name from the existing query.
	if (!$tableName) {
		if (!empty($rs->tableName)) $tableName = $rs->tableName;
		else if (preg_match("/FROM\s+".ADODB_TABLE_REGEX."/is", $rs->sql, $tableName))
			$tableName = $tableName[1];
		else
			return false;
	}

	// Strip off the comma and space on the end of both the fields
	// and their values.
	$fields = substr($fields, 0, -2);
	$values = substr($values, 0, -2);

	// Append the fields and their values to the insert query.
	return 'INSERT INTO '.$tableName.' ( '.$fields.' ) VALUES ( '.$values.' )';
}


/**
 * This private method is used to help construct
 * the update/sql which is generated by GetInsertSQL and GetUpdateSQL.
 * It handles the string construction of 1 column -> sql string based on
 * the column type.  We want to do 'safe' handling of BLOBs
 *
 * @param string the type of sql we are trying to create
 *                'I' or 'U'.
 * @param string column data type from the db::MetaType() method
 * @param string the column name
 * @param array the column value
 *
 * @return string
 *
 */
function _adodb_column_sql_oci8(&$zthis,$action, $type, $fname, $fnameq, $arrFields)
{

<     $sql = '';
< 


    // Based on the datatype of the field
    // Format the value properly for the database
    switch($type) {
    case 'B':
        //in order to handle Blobs correctly, we need
        //to do some magic for Oracle

        //we need to create a new descriptor to handle
        //this properly
        if (!empty($zthis->hasReturningInto)) {
            if ($action == 'I') {
                $sql = 'empty_blob(), ';
            } else {
                $sql = $fnameq. '=empty_blob(), ';
            }
            //add the variable to the returning clause array
            //so the user can build this later in
            //case they want to add more to it
            $zthis->_returningArray[$fname] = ':xx'.$fname.'xx';

<         } else if (empty($arrFields[$fname])){



> 			} else {
> 				if (empty($arrFields[$fname])) {


            if ($action == 'I') {
                $sql = 'empty_blob(), ';
            } else {
                $sql = $fnameq. '=empty_blob(), ';
            }
        } else {
            //this is to maintain compatibility
            //with older adodb versions.
			$sql = _adodb_column_sql($zthis, $action, $type, $fname, $fnameq, $arrFields, false);
        }

> 			}


        break;

    case "X":
        //we need to do some more magic here for long variables
        //to handle these correctly in oracle.

        //create a safe bind var name
        //to avoid conflicts w/ dupes.
       if (!empty($zthis->hasReturningInto)) {
            if ($action == 'I') {
                $sql = ':xx'.$fname.'xx, ';
            } else {
                $sql = $fnameq.'=:xx'.$fname.'xx, ';
            }
            //add the variable to the returning clause array
            //so the user can build this later in
            //case they want to add more to it
            $zthis->_returningArray[$fname] = ':xx'.$fname.'xx';
        } else {
            //this is to maintain compatibility
            //with older adodb versions.
			$sql = _adodb_column_sql($zthis, $action, $type, $fname, $fnameq, $arrFields, false);
        }
        break;

    default:
		$sql = _adodb_column_sql($zthis, $action, $type, $fname, $fnameq,  $arrFields, false);
        break;
    }

    return $sql;
}

function _adodb_column_sql(&$zthis, $action, $type, $fname, $fnameq, $arrFields, $recurse=true)
{

	if ($recurse) {
		switch($zthis->dataProvider)  {
		case 'postgres':
			if ($type == 'L') $type = 'C';
			break;
		case 'oci8':
			return _adodb_column_sql_oci8($zthis, $action, $type, $fname, $fnameq, $arrFields);

		}
	}

	switch($type) {
		case "C":
		case "X":
		case 'B':
			$val = $zthis->qstr($arrFields[$fname]);
			break;

		case "D":
			$val = $zthis->DBDate($arrFields[$fname]);
			break;

		case "T":
			$val = $zthis->DBTimeStamp($arrFields[$fname]);
			break;

		case "N":
		    $val = $arrFields[$fname];
			if (!is_numeric($val)) $val = str_replace(',', '.', (float)$val);
		    break;

		case "I":
		case "R":
		    $val = $arrFields[$fname];
			if (!is_numeric($val)) $val = (integer) $val;
		    break;

		default:
			$val = str_replace(array("'"," ","("),"",$arrFields[$fname]); // basic sql injection defence
			if (empty($val)) $val = '0';
			break;
	}

	if ($action == 'I') return $val . ", ";

	return $fnameq . "=" . $val  . ", ";
}



< 
< function _adodb_debug_execute(&$zthis, $sql, $inputarr)



> /**
> * Replaces standard _execute when debug mode is enabled
> *
> * @param ADOConnection   $zthis    An ADOConnection object
> * @param string|string[] $sql      A string or array of SQL statements
> * @param string[]|null   $inputarr An optional array of bind parameters
> *
> * @return  handle|void A handle to the executed query
> */
> function _adodb_debug_execute($zthis, $sql, $inputarr)


{

> 	// Unpack the bind parameters


	$ss = '';
	if ($inputarr) {
		foreach($inputarr as $kk=>$vv) {

< 			if (is_string($vv) && strlen($vv)>64) $vv = substr($vv,0,64).'...';
< 			if (is_null($vv)) $ss .= "($kk=>null) ";
< 			else $ss .= "($kk=>'$vv') ";



> 			if (is_string($vv) && strlen($vv) > 64) {
> 				$vv = substr($vv, 0, 64) . '...';
> 			}
> 			if (is_null($vv)) {
> 				$ss .= "($kk=>null) ";
> 			} else {
> 				if (is_array($vv)) {
> 					$vv = sprintf("Array Of Values: [%s]", implode(',', $vv));
> 				}
> 				$ss .= "($kk=>'$vv') ";
> 			}


		}
		$ss = "[ $ss ]";
	}

> 


	$sqlTxt = is_array($sql) ? $sql[0] : $sql;

< 	/*str_replace(', ','##1#__^LF',is_array($sql) ? $sql[0] : $sql);
< 	$sqlTxt = str_replace(',',', ',$sqlTxt);
< 	$sqlTxt = str_replace('##1#__^LF', ', ' ,$sqlTxt);
< 	*/



> 
> 	// Remove newlines and tabs, compress repeating spaces
> 	$sqlTxt = preg_replace('/\s+/', ' ', $sqlTxt);
> 


	// check if running from browser or command-line
	$inBrowser = isset($_SERVER['HTTP_USER_AGENT']);


< 	$dbt = $zthis->databaseType;
< 	if (isset($zthis->dsnType)) $dbt .= '-'.$zthis->dsnType;



> 	$myDatabaseType = $zthis->databaseType;
> 	if (!isset($zthis->dsnType)) {
> 		// Append the PDO driver name
> 		$myDatabaseType .= '-' . $zthis->dsnType;
> 	}
> 


	if ($inBrowser) {
		if ($ss) {

< 			$ss = '<code>'.htmlspecialchars($ss).'</code>';



> 			// Default formatting for passed parameter
> 			$ss = sprintf('<code class="adodb-debug">%s</code>', htmlspecialchars($ss));
> 		}
> 		if ($zthis->debug === -1) {
> 			$outString = "<br class='adodb-debug'>(%s):  %s &nbsp; %s<br class='adodb-debug'>";
> 			ADOConnection::outp(sprintf($outString, $myDatabaseType, htmlspecialchars($sqlTxt), $ss), false);
> 		} elseif ($zthis->debug !== -99) {
> 			$outString = "<hr class='adodb-debug'>(%s):  %s &nbsp; %s<hr class='adodb-debug'>";
> 			ADOConnection::outp(sprintf($outString, $myDatabaseType, htmlspecialchars($sqlTxt), $ss), false);
> 		}
> 	} else {
> 		// CLI output
> 		if ($zthis->debug !== -99) {
> 			$outString = sprintf("%s\n%s\n    %s %s \n%s\n", str_repeat('-', 78), $myDatabaseType, $sqlTxt, $ss, str_repeat('-', 78));
> 			ADOConnection::outp($outString, false);


		}

< 		if ($zthis->debug === -1)
< 			ADOConnection::outp( "<br>\n($dbt): ".htmlspecialchars($sqlTxt)." &nbsp; $ss\n<br>\n",false);
< 		else if ($zthis->debug !== -99)
< 			ADOConnection::outp( "<hr>\n($dbt): ".htmlspecialchars($sqlTxt)." &nbsp; $ss\n<hr>\n",false);
< 	} else {
< 		$ss = "\n   ".$ss;
< 		if ($zthis->debug !== -99)
< 			ADOConnection::outp("-----<hr>\n($dbt): ".$sqlTxt." $ss\n-----<hr>\n",false);


	}


> 	// Now execute the query


	$qID = $zthis->_query($sql,$inputarr);


< 	/*
< 		Alexios Fakios notes that ErrorMsg() must be called before ErrorNo() for mssql
< 		because ErrorNo() calls Execute('SELECT @ERROR'), causing recursion
< 	*/



> 	// Alexios Fakios notes that ErrorMsg() must be called before ErrorNo() for mssql
> 	// because ErrorNo() calls Execute('SELECT @ERROR'), causing recursion


	if ($zthis->databaseType == 'mssql') {

< 	// ErrorNo is a slow function call in mssql, and not reliable in PHP 4.0.6
< 



> 		// ErrorNo is a slow function call in mssql


		if($emsg = $zthis->ErrorMsg()) {
			if ($err = $zthis->ErrorNo()) {

< 				if ($zthis->debug === -99)
< 					ADOConnection::outp( "<hr>\n($dbt): ".htmlspecialchars($sqlTxt)." &nbsp; $ss\n<hr>\n",false);



> 				if ($zthis->debug === -99) {
> 					ADOConnection::outp("<hr>\n($myDatabaseType): " . htmlspecialchars($sqlTxt) . " &nbsp; $ss\n<hr>\n", false);
> 				}



				ADOConnection::outp($err.': '.$emsg);
			}
		}

< 	} else if (!$qID) {
< 
< 		if ($zthis->debug === -99)
< 				if ($inBrowser) ADOConnection::outp( "<hr>\n($dbt): ".htmlspecialchars($sqlTxt)." &nbsp; $ss\n<hr>\n",false);
< 				else ADOConnection::outp("-----<hr>\n($dbt): ".$sqlTxt."$ss\n-----<hr>\n",false);



> 	} else {
> 		if (!$qID) {
> 			// Statement execution has failed
> 			if ($zthis->debug === -99) {
> 				if ($inBrowser) {
> 					$outString = "<hr class='adodb-debug'>(%s):  %s &nbsp; %s<hr class='adodb-debug'>";
> 					ADOConnection::outp(sprintf($outString, $myDatabaseType, htmlspecialchars($sqlTxt), $ss), false);
> 				} else {
> 					$outString = sprintf("%s\n%s\n    %s %s \n%s\n",str_repeat('-',78),$myDatabaseType,$sqlTxt,$ss,str_repeat('-',78));
> 					ADOConnection::outp($outString, false);
> 				}
> 			}




< 		ADOConnection::outp($zthis->ErrorNo() .': '. $zthis->ErrorMsg());



> 			// Send last error to output
> 			$errno = $zthis->ErrorNo();
> 			if ($errno) {
> 				ADOConnection::outp($errno . ': ' . $zthis->ErrorMsg());
> 			}
> 		}


	}


< 	if ($zthis->debug === 99) _adodb_backtrace(true,9999,2);



> 	if ($qID === false || $zthis->debug === 99) {
> 		_adodb_backtrace();
> 	}


	return $qID;
}


< # pretty print the debug_backtrace function
< function _adodb_backtrace($printOrArr=true,$levels=9999,$skippy=0,$ishtml=null)



> /**
>  * Pretty print the debug_backtrace function
>  *
>  * @param string[]|bool $printOrArr       Whether to print the result directly or return the result
>  * @param int           $maximumDepth     The maximum depth of the array to traverse
>  * @param int           $elementsToIgnore The backtrace array indexes to ignore
>  * @param null|bool     $ishtml           True if we are in a CGI environment, false for CLI,
>  *                                        null to auto detect
>  *
>  * @return string Formatted backtrace
>  */
> function _adodb_backtrace($printOrArr=true, $maximumDepth=9999, $elementsToIgnore=0, $ishtml=null)


{

< 	if (!function_exists('debug_backtrace')) return '';



> 	if (!function_exists('debug_backtrace')) {
> 		return '';
> 	}




< 	if ($ishtml === null) $html =  (isset($_SERVER['HTTP_USER_AGENT']));
< 	else $html = $ishtml;



> 	if ($ishtml === null) {
> 		// Auto determine if we in a CGI enviroment
> 		$html = (isset($_SERVER['HTTP_USER_AGENT']));
> 	} else {
> 		$html = $ishtml;
> 	}




< 	$fmt =  ($html) ? "</font><font color=#808080 size=-1> %% line %4d, file: <a href=\"file:/%s\">%s</a></font>" : "%% line %4d, file: %s";



> 	$cgiString = "</font><font color=#808080 size=-1> %% line %4d, file: <a href=\"file:/%s\">%s</a></font>";
> 	$cliString = "%% line %4d, file: %s";
> 	$fmt = ($html) ? $cgiString : $cliString;



	$MAXSTRLEN = 128;

	$s = ($html) ? '<pre align=left>' : '';


< 	if (is_array($printOrArr)) $traceArr = $printOrArr;
< 	else $traceArr = debug_backtrace();



> 	if (is_array($printOrArr)) {
> 		$traceArr = $printOrArr;
> 	} else {
> 		$traceArr = debug_backtrace();
> 	}
> 
> 	// Remove first 2 elements that just show calls to adodb_backtrace


	array_shift($traceArr);
	array_shift($traceArr);

< 	$tabs = sizeof($traceArr)-2;



> 
> 	// We want last element to have no indent
> 	$tabs = sizeof($traceArr) - 1;



	foreach ($traceArr as $arr) {

< 		if ($skippy) {$skippy -= 1; continue;}
< 		$levels -= 1;
< 		if ($levels < 0) break;



> 		if ($elementsToIgnore) {
> 			// Ignore array element at start of array
> 			$elementsToIgnore--;
> 			$tabs--;
> 			continue;
> 		}
> 		$maximumDepth--;
> 		if ($maximumDepth < 0) {
> 			break;
> 		}



		$args = array();

< 		for ($i=0; $i < $tabs; $i++) $s .=  ($html) ? ' &nbsp; ' : "\t";
< 		$tabs -= 1;
< 		if ($html) $s .= '<font face="Courier New,Courier">';
< 		if (isset($arr['class'])) $s .= $arr['class'].'.';
< 		if (isset($arr['args']))



> 
> 		if ($tabs) {
> 			$s .= str_repeat($html ? ' &nbsp; ' : "\t", $tabs);
> 			$tabs--;
> 		}
> 		if ($html) {
> 			$s .= '<font face="Courier New,Courier">';
> 		}
> 
> 		if (isset($arr['class'])) {
> 			$s .= $arr['class'] . '.';
> 		}
> 
> 		if (isset($arr['args'])) {


		 foreach($arr['args'] as $v) {

< 			if (is_null($v)) $args[] = 'null';
< 			else if (is_array($v)) $args[] = 'Array['.sizeof($v).']';
< 			else if (is_object($v)) $args[] = 'Object:'.get_class($v);
< 			else if (is_bool($v)) $args[] = $v ? 'true' : 'false';
< 			else {



> 				if (is_null($v)) {
> 					$args[] = 'null';
> 				} elseif (is_array($v)) {
> 					$args[] = 'Array[' . sizeof($v) . ']';
> 				} elseif (is_object($v)) {
> 					$args[] = 'Object:' . get_class($v);
> 				} elseif (is_bool($v)) {
> 					$args[] = $v ? 'true' : 'false';
> 				} else {


				$v = (string) @$v;

< 				$str = htmlspecialchars(str_replace(array("\r","\n"),' ',substr($v,0,$MAXSTRLEN)));
< 				if (strlen($v) > $MAXSTRLEN) $str .= '...';



> 					// Truncate
> 					$v = substr($v, 0, $MAXSTRLEN);
> 					// Remove newlines and tabs, compress repeating spaces
> 					$v = preg_replace('/\s+/', ' ', $v);
> 					// Convert htmlchars (not sure why we do this in CLI)
> 					$str = htmlspecialchars($v);
> 
> 					if (strlen($v) > $MAXSTRLEN) {
> 						$str .= '...';
> 					}
> 


				$args[] = $str;
			}
		}

> 		}


		$s .= $arr['function'].'('.implode(', ',$args).')';

< 
< 


		$s .= @sprintf($fmt, $arr['line'],$arr['file'],basename($arr['file']));

< 


		$s .= "\n";
	}

< 	if ($html) $s .= '</pre>';
< 	if ($printOrArr) print $s;
< 
< 	return $s;
< }
< /*
< function _adodb_find_from($sql)
< {
< 
< 	$sql = str_replace(array("\n","\r"), ' ', $sql);
< 	$charCount = strlen($sql);
< 
< 	$inString = false;
< 	$quote = '';
< 	$parentheseCount = 0;
< 	$prevChars = '';
< 	$nextChars = '';
< 
< 
< 	for($i = 0; $i < $charCount; $i++) {
< 
<     	$char = substr($sql,$i,1);
< 	    $prevChars = substr($sql,0,$i);
<     	$nextChars = substr($sql,$i+1);
< 
< 		if((($char == "'" || $char == '"' || $char == '`') && substr($prevChars,-1,1) != '\\') && $inString === false) {
< 			$quote = $char;
< 			$inString = true;



> 	if ($html) {
> 		$s .= '</pre>';


		}

< 
< 		elseif((($char == "'" || $char == '"' || $char == '`') && substr($prevChars,-1,1) != '\\') && $inString === true && $quote == $char) {
< 			$quote = "";
< 			$inString = false;



> 	if ($printOrArr) {
> 		print $s;


		}


< 		elseif($char == "(" && $inString === false)
< 			$parentheseCount++;
< 
< 		elseif($char == ")" && $inString === false && $parentheseCount > 0)
< 			$parentheseCount--;
< 
< 		elseif($parentheseCount <= 0 && $inString === false && $char == " " && strtoupper(substr($prevChars,-5,5)) == " FROM")
< 			return $i;
< 
< 	}



> 	return $s;


}

< */