crash @2011-02-19 01:46:00 GMT info save-dump +dump.2011-02-19T01-46-00+0000 bavarian forwards from A5 to B7 but B7/J was set against it, so fault previous attempts via the UI to set B7/J weren't successful, but accelerating towards the junction showed it apparently going on through A5 to A6 but actually A5/J was reversed bessar hidraw so we can have gamepad bugs at converging points eg X8, can we make train go from (say) X6 fast enough that when we stop it it fouls the dynamic envelope for X4-X8 but doesn't enter X8 and thus signalling system allows crash ? gui bugs speedw-redisplay-speed can crash reading unset variable gcommanded when trying to switch away from current train after realtime crashed; presumably some state is not properly sorted out ? allegedly fixed, but untested current set of stuff on liberator liberator:trains> rsync -vPI bessar:/bessar/bessar/playground/+dump.2008-07-28T00-02-43+0100/* hostside/ * wrongly says santafe impeded by shinkansen when we try to reverse it (look from tail of log) * when we tell it to go it has a fit about A2 if train was abutting section boundary and is told to go, maxinto is not updated properly, so if it is then told to stop it stops owning the segment it may have just entered to reproduce: put shinkansen nose car at A6 (home location) and nothing else. command speed 1 then speed 0 should now own X12 too but does not bugs should expire dumps, eventually does not set serial port mode etc. when checking future polarity possibility, does not account for MARGIN_POLARISETIME - it just checks to see if extending front before reducing rear is sufficient this will not work if there are segments which are traversed more quickly than MARGIN_POLARISETIME movpos++ when under train does not work - is ignored and gives ok want to stop all trains command ? stop and turn off calmly when ^D why FAULT reported when power is turned off ? things not yet considered at all in safety code: min. curve specifications would be nice to do something with POINTED