More crashes

[Junna]

https://www.mediafire.com/?4mq9kkwn83p9bpb

Should crash at 2:52:xx of current month. Same type of crash as previously exhibited.

On the upside, the reservation oddities and trains not being released from signals appears to be fixed.

[jamespetts]

I think that I have fixed these, but have not had a chance to test fully. Are you able to re-test?

[Junna]

That fixes the one crash, but non-easily reproducible ones seem to form. They seem to occur randomly; re-loading a game saved just prior will not exhibit the same crash. Reservation oddities are back, too.

https://www.mediafire.com/?74z42j59ce1txmg

Applechester is a good example of a place where spurious reservations appear.

[jamespetts]

Thank you for letting me know: I will have to look into this. These problems are, unfortunately, difficult to solve because changing the signalling code to allow reservations routinely beyond stations invalidated assumptions made by very large numbers of disparate parts of the code.

Incidentally, would you be able to recompile your pakset, or at least the ways, with the latest makeobj (compiled with the new monthly way wear feature) so that I can load your saved games with the latest builds? Thank you.

[Junna]

The ways are unchanged from regular pak.britain-ex (apart from some unrelated roads I have added which are from standard and don't have experimental parameters), you should be able to replace those with whatever you already have compiled? 

[jamespetts]

Can you indicate the file names of the .pak files for these extra roads?

Edit: I have fixed one more crash - do you care to re-test?

The reason that I ask for that information, incidentally, is that my .pak files are compiled differently to yours, using the Python script rather than the makefile, so I just have one .pak file for each type of object (ways, rail vehicles, road vehicles, stations, depots, etc.) rather than one for each object as you have.

[Junna]

Half-heights ways on github does not seem to have been updated with the new variable set_monthly_wear ?

[jamespetts]

I had set the default values (for an unused way to be fully worn out in 50 years with no usage except for waterways, which never wear out) to be what I wanted so that I did not need to alter the individual .dat files at this stage, until such time, at least, as balancing testing shows those defaults to be sub-optimal.