covmat was only correct in upper triangle!

src/bk.f

@@ -252,6 +252,7 @@ c     determine search neighbourhood
       CALL SRCHNB(LON0,LAT0,DO0,INDDO,N0,NDO,LON,LAT,N,NS,
      .            RSEARCH,NSEARCH,NSMIN,NSMAX,
      .            INDSRT,INDSNB,INDSNW,DIST)
+      write(*,*)"SEARCHNB is:",NS
       if (usesbbt.ne.0) then
          do 1212 i=1,n
             if (indsnb(i).ne.0) then
@@ -415,10 +416,10 @@ c     prepare the right hand side(s)
       
 c     solve the system
 c     LAMBDA = (C0VEC+FWORK*PHIPR*F0WORK)*(COVMAT+FWORK*PHIPR*FWORK')^-1
-c      name="kwork\0"
-c      call matpr(name,kwork,ns,ns,LDkwrk,dbglvl)
-c      name="rhswork\0"
-c      call matpr(name,rhswork,ns,NDO,LDkWrk,dbglvl)
+      name="kwork\0"
+      call matpr(name,kwork,ns,ns,LDkwrk,dbglvl)
+      name="rhswork\0"
+      call matpr(name,rhswork,ns,NDO,LDkWrk,dbglvl)
 C     BETTER USE DSYSVX ?
 
       CALL DGESV(NS,NDO,KWORK,LDKWRK,IPIV,RHSWORK,LDKWRK,INFO)
@@ -447,6 +448,7 @@ c     extract optimal weights
          DO 50 I=1,NS
 c            write (*,*)i,indsnb(i),j
             LAMBDA(INDSNB(I),INDDO(J))=RHSWORK(I,J)
+c            write (*,*)LAMBDA(INDSNB(I),INDDO(J))
  50      CONTINUE
          
  400  CONTINUE

src/bk_point.c

@@ -127,6 +127,8 @@ void bk_point(double *xp,
 	for(j=i;j<(*n);j++){
 	    dst = sqrt((lon[i]-lon[j])*(lon[i]-lon[j])+(lat[i]-lat[j])*(lat[i]-lat[j]));
 	    covmat[i+j*(*n)]=F77_CALL(covfn)(covtype,covpar,&dst);
+	    covmat[j+i*(*n)]=covmat[i+j*(*n)];
+		
 	}
     } /* else { error !! } */