chiark / gitweb /
ian [Sun, 14 Sep 2008 15:18:23 +0000 (15:18 +0000)]
delete junking debugging
ian [Sun, 14 Sep 2008 15:17:40 +0000 (15:17 +0000)]
some debugging of fdet_nextseg's elapsed etc. being preserved in junk
ian [Sun, 24 Aug 2008 22:53:58 +0000 (22:53 +0000)]
use t->backwards not t->seg->tr_backwards (which may not yet be set) when doing need_polarise computation
ian [Sun, 24 Aug 2008 22:51:51 +0000 (22:51 +0000)]
more debugging of inversions
ian [Sun, 24 Aug 2008 22:50:53 +0000 (22:50 +0000)]
converging points bug
ian [Sat, 23 Aug 2008 11:02:55 +0000 (11:02 +0000)]
summary graph diagram
ian [Fri, 22 Aug 2008 20:51:46 +0000 (20:51 +0000)]
always show stastate
ian [Sun, 17 Aug 2008 16:48:35 +0000 (16:48 +0000)]
some documentation
ian [Sun, 17 Aug 2008 16:41:59 +0000 (16:41 +0000)]
todo tidy
ian [Wed, 30 Jul 2008 21:38:26 +0000 (21:38 +0000)]
Hopefully-this-time-correct arrangements for dealing with uncertainty>maxinto. safety.c r1.96 and r1.61 were definitely wrong ...
ian [Sun, 27 Jul 2008 23:06:05 +0000 (23:06 +0000)]
bug
ian [Sun, 27 Jul 2008 22:54:30 +0000 (22:54 +0000)]
MARGIN_TRAINLENGTH
ian [Sun, 27 Jul 2008 22:40:49 +0000 (22:40 +0000)]
do not print nuls in later bits of resolve ends
ian [Sun, 27 Jul 2008 22:19:29 +0000 (22:19 +0000)]
call predict() at end of resolution; fixes to resolution
ian [Sun, 27 Jul 2008 20:25:46 +0000 (20:25 +0000)]
New resolution arrangements; before move predict() call into resolve_motioncheck
ian [Sun, 27 Jul 2008 16:00:15 +0000 (16:00 +0000)]
Clearer definition of meaning of margins, maxinto, etc.
ian [Sun, 27 Jul 2008 15:17:15 +0000 (15:17 +0000)]
Do not increment lookahead before !!done_fdet (!)
ian [Sun, 27 Jul 2008 14:59:54 +0000 (14:59 +0000)]
bound neato images; this (neato) seems buggy but this works on liberator
ian [Sun, 27 Jul 2008 14:38:46 +0000 (14:38 +0000)]
show a # before event counter
ian [Sun, 27 Jul 2008 14:27:12 +0000 (14:27 +0000)]
Reuse old plan for direction setting; report train location on prediction entry
ian [Sun, 27 Jul 2008 14:26:40 +0000 (14:26 +0000)]
some debugging message improvements
ian [Sun, 27 Jul 2008 14:00:39 +0000 (14:00 +0000)]
Properly restore fdet->tr_backwards on failed direction change
ian [Sun, 27 Jul 2008 13:55:23 +0000 (13:55 +0000)]
increase headroom
ian [Sun, 27 Jul 2008 13:53:18 +0000 (13:53 +0000)]
abandon infinite loop in report_nextseg sooner
ian [Sun, 27 Jul 2008 13:48:51 +0000 (13:48 +0000)]
puts rather than lputs on realtime failure so we always report it
ian [Sat, 26 Jul 2008 23:13:52 +0000 (23:13 +0000)]
relay train on decel_done (untested)
ian [Sat, 26 Jul 2008 23:07:57 +0000 (23:07 +0000)]
Properly deal with MARGIN_SPEED
ian [Sat, 26 Jul 2008 22:59:22 +0000 (22:59 +0000)]
adjust (maximise) maxinto when we accelerate (untested)
ian [Sat, 26 Jul 2008 16:36:28 +0000 (16:36 +0000)]
Fix pretty much wholly buggy speed change logic
ian [Sat, 26 Jul 2008 16:35:58 +0000 (16:35 +0000)]
new macros MIN MAX ENSURE
ian [Sat, 26 Jul 2008 16:30:42 +0000 (16:30 +0000)]
declare ouprintfs __attribute__((format...))
ian [Sat, 26 Jul 2008 15:41:00 +0000 (15:41 +0000)]
Nicer interface between safety and speedmanager
ian [Sat, 26 Jul 2008 15:06:58 +0000 (15:06 +0000)]
Change default for predit to PREDF_NEWPLAN -
part 2, remove PREDF_NEWPLAN entirely
ian [Sat, 26 Jul 2008 15:05:44 +0000 (15:05 +0000)]
Change default for predit to PREDF_NEWPLAN -
part 1, change PREDF_ values and add PREDF_OLDPLAN
ian [Sun, 20 Jul 2008 22:50:05 +0000 (22:50 +0000)]
todo list items changed
ian [Sun, 20 Jul 2008 22:49:28 +0000 (22:49 +0000)]
todo list items changed
ian [Sun, 20 Jul 2008 18:03:44 +0000 (18:03 +0000)]
set direction from main keyboard too
ian [Sun, 20 Jul 2008 17:59:42 +0000 (17:59 +0000)]
can go backwards
ian [Sun, 20 Jul 2008 17:46:05 +0000 (17:46 +0000)]
more backwardsness bugfix
ian [Sun, 20 Jul 2008 17:37:18 +0000 (17:37 +0000)]
get directions right (for the benefit of setdirection)
ian [Sun, 20 Jul 2008 17:36:56 +0000 (17:36 +0000)]
produce debugging output when debugging with gdb
ian [Sun, 20 Jul 2008 17:14:43 +0000 (17:14 +0000)]
train direction command swap arguments
ian [Sun, 20 Jul 2008 17:14:32 +0000 (17:14 +0000)]
clear out old trains from saved replay
ian [Sun, 20 Jul 2008 16:38:59 +0000 (16:38 +0000)]
stub direction setter
ian [Sun, 20 Jul 2008 16:38:03 +0000 (16:38 +0000)]
speedw improvements
- static keyboard bindings available
- inhibition rationalised
- some useful restructuring
ian [Sun, 20 Jul 2008 15:17:25 +0000 (15:17 +0000)]
fix grammar in a comment
ian [Fri, 18 Jul 2008 18:17:30 +0000 (18:17 +0000)]
allow multiplex testmode to work again
with this rune:
./multiplex --testmode=1 --dev_railway=/dev/null
ian [Wed, 16 Jul 2008 20:54:44 +0000 (20:54 +0000)]
report train's new position after prediction succeeds
ian [Wed, 16 Jul 2008 20:51:49 +0000 (20:51 +0000)]
direction command
ian [Wed, 16 Jul 2008 20:43:10 +0000 (20:43 +0000)]
use -MMD
ian [Wed, 16 Jul 2008 20:37:14 +0000 (20:37 +0000)]
safety_setdirection compiles but untested
ian [Wed, 16 Jul 2008 17:38:44 +0000 (17:38 +0000)]
MUSTECR is not just for ps_...
ian [Tue, 15 Jul 2008 22:40:01 +0000 (22:40 +0000)]
wip obc buffering transparency debugging
testing with
-bessar:hostside> ./realtime -v2 shinkansen.speeds.record santafe.speeds.record homes.record
ian [Tue, 15 Jul 2008 20:38:30 +0000 (20:38 +0000)]
note that detection suppressed by README.commands
ian [Tue, 15 Jul 2008 20:37:56 +0000 (20:37 +0000)]
document detect-flap
ian [Tue, 15 Jul 2008 20:34:42 +0000 (20:34 +0000)]
coalesce flapping detection
ian [Tue, 15 Jul 2008 20:18:57 +0000 (20:18 +0000)]
publish total amount buffered in an obc
ian [Tue, 15 Jul 2008 20:17:07 +0000 (20:17 +0000)]
new OutBufferChain->empty feature
ian [Tue, 15 Jul 2008 00:27:41 +0000 (00:27 +0000)]
allow stastate to override picion on even if they come in different order (eg due to replay)
ian [Mon, 14 Jul 2008 23:50:11 +0000 (23:50 +0000)]
invoke gui through xterm so we get to see errors
ian [Mon, 14 Jul 2008 23:26:18 +0000 (23:26 +0000)]
turn lputs off by default
ian [Mon, 14 Jul 2008 23:06:39 +0000 (23:06 +0000)]
run things on bessar
ian [Mon, 14 Jul 2008 22:19:33 +0000 (22:19 +0000)]
get suppression of pings to lputs right
ian [Mon, 14 Jul 2008 22:17:27 +0000 (22:17 +0000)]
revert suppression of general ping since it's pointless unless we supress the timestamps too
ian [Mon, 14 Jul 2008 22:14:05 +0000 (22:14 +0000)]
fix trivial bug in suppression of general pings
ian [Mon, 14 Jul 2008 22:08:31 +0000 (22:08 +0000)]
suppress output of general pings to simulation log
ian [Mon, 14 Jul 2008 21:55:00 +0000 (21:55 +0000)]
remove segments with unknown movpos from resfin_done set so they don't end up owned by trains
ian [Mon, 14 Jul 2008 21:43:00 +0000 (21:43 +0000)]
minor debugging improvements
ian [Mon, 14 Jul 2008 21:42:48 +0000 (21:42 +0000)]
mark[0123] are supposed to be 1 bit each
ian [Mon, 14 Jul 2008 21:19:38 +0000 (21:19 +0000)]
get globals right for permanently open serchan
ian [Mon, 14 Jul 2008 21:16:54 +0000 (21:16 +0000)]
keep serchan open to defeat stupid kernel bug (EIO when opening a tty someone else is closing
ian [Mon, 14 Jul 2008 21:07:12 +0000 (21:07 +0000)]
do not log await-off ping/pong
ian [Mon, 14 Jul 2008 21:04:36 +0000 (21:04 +0000)]
do not lputs pings
ian [Mon, 14 Jul 2008 20:49:16 +0000 (20:49 +0000)]
use POSIX flags not r+ to open railway
ian [Mon, 14 Jul 2008 20:39:34 +0000 (20:39 +0000)]
treat low startup states as power off
ian [Mon, 14 Jul 2008 20:24:59 +0000 (20:24 +0000)]
can show unbound; sel train key on ebwikeb
ian [Mon, 14 Jul 2008 20:16:28 +0000 (20:16 +0000)]
select train with left button
ian [Mon, 14 Jul 2008 19:57:54 +0000 (19:57 +0000)]
fullscreen on bessar
ian [Sun, 13 Jul 2008 23:54:16 +0000 (23:54 +0000)]
fix it to work with empty config file
ian [Sun, 13 Jul 2008 23:36:15 +0000 (23:36 +0000)]
bessar gui seems ok now with mouse and keyboard
ian [Sun, 13 Jul 2008 23:35:49 +0000 (23:35 +0000)]
make multiplex not crash if !realtime auto and already running
ian [Sun, 13 Jul 2008 23:09:14 +0000 (23:09 +0000)]
initial ebuyer kbd code
ian [Sun, 13 Jul 2008 22:31:43 +0000 (22:31 +0000)]
use --redact rather than --dump
ian [Sun, 13 Jul 2008 22:31:06 +0000 (22:31 +0000)]
some signalling todos discovered
ian [Sun, 13 Jul 2008 21:51:42 +0000 (21:51 +0000)]
better bind-input command construction
ian [Sun, 13 Jul 2008 21:45:34 +0000 (21:45 +0000)]
decouple input-binding from speedw
ian [Sun, 13 Jul 2008 20:53:43 +0000 (20:53 +0000)]
wip bessar gui bindings
ian [Sun, 13 Jul 2008 20:53:22 +0000 (20:53 +0000)]
do not go wrong on !realtime stop
ian [Sun, 13 Jul 2008 20:52:59 +0000 (20:52 +0000)]
parse different config file on each host
ian [Sun, 13 Jul 2008 19:24:44 +0000 (19:24 +0000)]
identifiers here have - not _
ian [Sun, 13 Jul 2008 19:24:32 +0000 (19:24 +0000)]
get line number right in error msgs
ian [Sun, 13 Jul 2008 18:59:48 +0000 (18:59 +0000)]
support lookup by feature address too; specify feature commands
ian [Sat, 5 Jul 2008 13:38:29 +0000 (13:38 +0000)]
wip features command
ian [Mon, 30 Jun 2008 12:23:06 +0000 (12:23 +0000)]
parses and transmits features; compiled but not tested (including suppressed features like half speed)
ian [Thu, 26 Jun 2008 16:19:52 +0000 (16:19 +0000)]
Give relaxed xmissions of urgent messages >=50% share
ian [Thu, 26 Jun 2008 15:52:14 +0000 (15:52 +0000)]
features are named rather than lettered; define the features
ian [Tue, 24 Jun 2008 21:35:21 +0000 (21:35 +0000)]
Revert aborting feature changes as planned
ian [Tue, 24 Jun 2008 21:34:45 +0000 (21:34 +0000)]
Abortive change to feature arrangements.
Was going to
- change features to be named rather than lettered
- make realtime not deal with the names at all
But this is actually tedious. Instead if we need another level of
indirection for the UI we'll add one then.
So commit this now and immediately revert it.
ian [Tue, 24 Jun 2008 20:15:03 +0000 (20:15 +0000)]
fix comment for new speeds.record
ian [Mon, 16 Jun 2008 02:15:48 +0000 (02:15 +0000)]
new eventrun machinery for bessar