If you've successfully run canopy.sample() but encountered error in the canopy.post(), most likely the error is due to not running the MCMC sampling long enough. It is STRONGLY recommended that the users keep the default sampling parameters in canopy.sample() or even set the number of sampling iterations even higher. A pdf file with plot of posterior likelihood and acceptance rate is generated by default and can be used to assess whether the sampling has converged or not and whether the sampling is stuck in a local mode. Two specific errors posted by previous users are listed below.
Error in 1:nrow(config.summary) : argument of length 0
The error was due to the fact that your MCMC has not converged yet. You need to let the chain run longer to get it converged. With the current thinning and binning that you have, the posterior tree space have too many configurations (because the chain is still sampling and accepting jumps) and the NULL cutoff tree configuration frequency is 0.05. As a result, all of your posterior configurations have frequencies < 0.05. A very quick fix to get a maximum-likelihood tree is to set post.config.cutoff = 0.01 but I would STRONGLY recommend that you sample longer.
Error in config.summary[i, 1] : subscript out of bounds
To speed up computation in MCMC and to avoid autocorrelation between neighboring trees that are samples consecutively, Canopy saves samples trees with a skip. The error that you see from the output you sent me seems to be due to the fact that with the writeskip, burnin, and thinning you don't have any posterior trees left. I would suggest running your trees longer to get more accurate posterior evaluations. Another option is you can change the burn in and thinning parameter (although not recommended). Either way I would suggest looking at the likelihood plot generated for sampling.