The certification script (rcr_certification.do) returns errors because some results have changed. But these are the result of environmental changes and not bugs in the program:
[x] Stata's value of max_matsize has gone up, so "More than 25 control variables" case runs rather than creating an error message.
[x] Stata's default random number generator has changed, so "Up to 25 control variables" case (which uses random numbers) produces different results
[x] The graph files are not exactly identical, so the checksums are not the same. This happens all the time due to changes in computers, Stata versions, fonts, etc. Need to have a more robust way of dealing with that.
Update the certification script so that it runs cleanly.
The certification script (rcr_certification.do) returns errors because some results have changed. But these are the result of environmental changes and not bugs in the program:
Update the certification script so that it runs cleanly.