I sincerely hope that this release cures the preferred direction problems associated with the Windows clipboard. It seemed like a good idea to include preferred directions at v2.9.0 but given the trouble they’ve caused now I’m not so sure.
One good thing is that one of the corrected bugs explains the error found by Daniel Gill at v2.9.0 in response to which the code was strengthened for v2.9.1, but it continued to bother me because I was unable (then) to reproduce it.
I have tested this release extensively but still feel quite apprehensive, my fingers and everything else are well crossed! Please let me know if you see any odd behaviour and send in any error reports if they are generated – thanks.
I have also to confess to another error, not a bug this time but a plain blunder on my part. I had noticed a while back that the performance log omitted exit counts – early, late and on time exits – so in an earlier release I included them, but forgot to initialise them all to zero! A schoolboy howler! I only noticed it recently but I’m sure others heve seen it too, with exit numbers listed in the hundreds of millions or billions! These have now been corrected.
I was ready to release the new version when Simon Banham sent me an image of a railway with a locked route and signals on it all green! That was clearly not right. On investigation it was found to occur when a segment of automatic signal route had a segment of non-automatic signal route behind it, and that route was locked. When a train in advance of the locked route passed a signal all signals behind it were reset, but the resetting function failed to take account of the locked route so signals were reset as if it didn’t exist. Fortunately it was quite an easy fix, so that shouldn’t happen again. Thanks are due to Simon for pointing this error out.
Download it here