Index: source/analysis/analysishelper.cxx
===================================================================
--- source/analysis/analysishelper.cxx	(revision 269817)
+++ source/analysis/analysishelper.cxx	(working copy)
@@ -669,209 +669,16 @@
 	return aRet;
 }
 
-/** Approximation algorithm for erf for 0 < x < 0.65. */
-void Erf0065( double x, double& fVal )
-{
-	static const double pn[] = {
-		1.12837916709551256,
-		1.35894887627277916E-1,
-		4.03259488531795274E-2,
-		1.20339380863079457E-3,
-		6.49254556481904354E-5
-	};
-	static const double qn[] = {
-		1.00000000000000000,
-		4.53767041780002545E-1,
-		8.69936222615385890E-2,
-		8.49717371168693357E-3,
-		3.64915280629351082E-4
-	};
-
-	double fPSum = 0.0;
-	double fQSum = 0.0;
-	double fXPow = 1.0;
-	for ( unsigned int i = 0; i <= 4; ++i )
-	{
-		fPSum += pn[i]*fXPow;
-		fQSum += qn[i]*fXPow;
-		fXPow *= x*x;
-	}
-	fVal = x * fPSum / fQSum;
-}
-
-/** Approximation algorithm for erfc for 0.65 < x < 6.0. */
-void Erfc0600( double x, double& fVal )
-{
-	double fPSum = 0.0;
-	double fQSum = 0.0;
-	double fXPow = 1.0;
-	const double *pn;
-	const double *qn;
-
-	if ( x < 2.2 )
-	{
-		static const double pn22[] = {
-			9.99999992049799098E-1,
-			1.33154163936765307,
-			8.78115804155881782E-1,
-			3.31899559578213215E-1,
-			7.14193832506776067E-2,
-			7.06940843763253131E-3
-		};
-		static const double qn22[] = {
-			1.00000000000000000,
-			2.45992070144245533,
-			2.65383972869775752,
-			1.61876655543871376,
-			5.94651311286481502E-1,
-			1.26579413030177940E-1,
-			1.25304936549413393E-2
-		};
-		pn = pn22;
-		qn = qn22;
-	}
-	else /* if ( x < 6.0 )  this is true, but the compiler does not know */
-	{
-		static const double pn60[] = {
-			9.99921140009714409E-1,
-			1.62356584489366647,
-			1.26739901455873222,
-			5.81528574177741135E-1,
-			1.57289620742838702E-1,
-			2.25716982919217555E-2
-		};
-		static const double qn60[] = {
-			1.00000000000000000,
-			2.75143870676376208,
-			3.37367334657284535,
-			2.38574194785344389,
-			1.05074004614827206,
-			2.78788439273628983E-1,
-			4.00072964526861362E-2
-		};
-		pn = pn60;
-		qn = qn60;
-	}
-
-	for ( unsigned int i = 0; i < 6; ++i )
-	{
-		fPSum += pn[i]*fXPow;
-		fQSum += qn[i]*fXPow;
-		fXPow *= x;
-	}
-	fQSum += qn[6]*fXPow;
-	fVal = exp( -1.0*x*x )* fPSum / fQSum;
-}
-
-/** Approximation algorithm for erfc for 6.0 < x < 26.54 (but used for all
-	x > 6.0). */
-void Erfc2654( double x, double& fVal )
-{
-	static const double pn[] = {
-		5.64189583547756078E-1,
-		8.80253746105525775,
-		3.84683103716117320E1,
-		4.77209965874436377E1,
-		8.08040729052301677
-	};
-	static const double qn[] = {
-		1.00000000000000000,
-		1.61020914205869003E1,
-		7.54843505665954743E1,
-		1.12123870801026015E2,
-		3.73997570145040850E1
-	};
-
-	double fPSum = 0.0;
-	double fQSum = 0.0;
-	double fXPow = 1.0;
-
-	for ( unsigned int i = 0; i <= 4; ++i )
-	{
-		fPSum += pn[i]*fXPow;
-		fQSum += qn[i]*fXPow;
-		fXPow /= x*x;
-	}
-	fVal = exp(-1.0*x*x)*fPSum / (x*fQSum);
-}
-
-double Erfc( double );
-
-/** Parent error function (erf) that calls different algorithms based on the
-	value of x.  It takes care of cases where x is negative as erf is an odd
-	function i.e. erf(-x) = -erf(x).
-
-	Kramer, W., and Blomquist, F., 2000, Algorithms with Guaranteed Error Bounds
-	for the Error Function and the Complementary Error Function
-
-	http://www.math.uni-wuppertal.de/wrswt/literatur_en.html
-
-	@author Kohei Yoshida <kohei@openoffice.org>
-
-	@see #i55735#
- */
+// calculation moved to module sal, see i97091
 double Erf( double x )
 {
-	if( x == 0.0 )
-		return 0.0;
-
-	bool bNegative = false;
-	if ( x < 0.0 )
-	{
-		x = fabs( x );
-		bNegative = true;
-	}
-
-	double fErf = 1.0;
-	if ( x < 1.0e-10 )
-		fErf = (double) (x*1.1283791670955125738961589031215452L);
-    else if ( x < 0.65 )
-		Erf0065( x, fErf );
-	else
-		fErf = 1.0 - Erfc( x );
-
-	if ( bNegative )
-		fErf *= -1.0;
-
-	return fErf;
+    return ::rtl::math::erf(x);
 }
 
-/** Parent complementary error function (erfc) that calls different algorithms
-	based on the value of x.  It takes care of cases where x is negative as erfc
-	satisfies relationship erfc(-x) = 2 - erfc(x).  See the comment for Erf(x)
-	for the source publication.
-
-	@author Kohei Yoshida <kohei@openoffice.org>
-
-	@see #i55735#
- */
+// calculation moved to module sal, see i97091
 double Erfc( double x )
 {
-	if ( x == 0.0 )
-		return 1.0;
-
-	bool bNegative = false;
-	if ( x < 0.0 )
-	{
-		x = fabs( x );
-		bNegative = true;
-	}
-
-	double fErfc = 0.0;
-    if ( x >= 0.65 )
-	{
-		if ( x < 6.0 )
-			Erfc0600( x, fErfc );
-		else
-			Erfc2654( x, fErfc );
-	}
-	else
-		fErfc = 1.0 - Erf( x );
-
-	if ( bNegative )
-		fErfc = 2.0 - fErfc;
-
-	return fErfc;
+    return ::rtl::math::erfc(x);
 }
 
 inline sal_Bool IsNum( sal_Unicode c )