[R-sig-ME] Regarding mcmcsamp in the most recent alpha release

[Dave Warren]

Hi all,

    I've been trying out the most recent alpha release and keep running 
into the following exception when I try using mcmcsamp.

 > path.lmer = lmer( log( PathLength ) ~ Trial * Status + (Trial | 
Patient), data = path_agg )
 > summary( path.lmer )
Linear mixed model fit by REML
Formula: log(PathLength) ~ Trial * Status + (Trial | Patient)
   Data: path_agg
   AIC   BIC logLik deviance REMLdev
 640.4 674.1 -312.2    594.9   624.4
Random effects:
 Groups   Name        Variance   Std.Dev.  Corr  
 Patient  (Intercept) 1.3567e-03 0.0368331       
          Trial       2.1394e-05 0.0046254 -1.000
 Residual             1.9524e-01 0.4418560       
Number of obs: 502, groups: Patient, 10

Fixed effects:
                    Estimate Std. Error t value
(Intercept)         2.995689   0.060844   49.24
Trial              -0.001977   0.002406   -0.82
StatusNormal        0.245692   0.084061    2.92
Trial:StatusNormal  0.008709   0.003382    2.58

Correlation of Fixed Effects:
            (Intr) Trial  SttsNr
Trial       -0.661             
StatusNorml -0.724  0.479      
Trl:SttsNrm  0.470 -0.711 -0.660

 > a = mcmcsamp( path.lmer, 10000 )
Error in .local(object, n, verbose, ...) :
  crossproduct matrix 1 is not positive definite

    Have I broken something?  Is my model revealed to be invalid?  
Thanks for any thoughts,

Dave