Inconsistent error when using rstarts

[winterstat]

I recently read De Jonckere & Rosseel (2025)'s BRM paper on using random starting values. I want to use this option in a larger simulation study but I'm running into an error. The problem with the error is that it doesn't consistently show up under fixed conditions. Here is the error:

 ‘Error in x.noerror[sapply(x.rstarts, "attr", "converged")] : invalid subscript type 'list' ’

I'm also playing around with the bounds argument, and I think that is the cause of these errors. The example below, using that seed, should replicate the issue (note that I'm generating a random covariance matrix on purpose):

library(lavaan)
library(clusterGeneration)

set.seed(123)
p <- 24
temp_mat <- genPositiveDefMat(dim = 24, covMethod ="onion", eta=1, rangeVar=c(1,1))$Sigma
colnames(temp_mat) <- rownames(temp_mat) <- paste0("x", seq(1:p))

mod1 <- '
f1 =~ x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + 
x9 + x10 + x11 + x12 + x13 + x14 + x15 + x16 +
x17 + x18 + x19 + x20 + x21 + x22 + x23 + x24
'

mod1b.fit <- cfa(mod1, sample.cov = temp_mat, 
                 bounds = "pos.var",             
                 rstarts = 10,
                 sample.nobs = 1000)

Any ideas what's going on? If the answer is that random starts and variance bounds don't combine, then that's also fine!

Thanks!

[yrosseel]

Indeed. Thanks for reporting this. The rstarts= and bounds= options did not like each other. I believe this is now fixed in the github version of lavaan.