full_power() notched up as soon as the controlling half had closed its line
contactors. In the ET42 test the locomotive brake is still at about 2.0 bar
then, right at the pressure switch threshold, so whether the second half
starts as well depended on thousandths of a bar in its cylinder: a half whose
switch is on when the controller leaves zero doesn't close its line
contactors until the controller is back at zero. Any change to the air
brakes could tip it either way (with the corrected ESt3 the second half
stayed off and the train reached 88.7 instead of 106 km/h).
The helper now waits, as a driver does, until no pressure switch in the
train cuts the start off.
An electric and a diesel locomotive got their control coupled whenever the
scenario (and so the Starter) asked for it, since the scenery coupling ignores
ControlType, or when neither coupler named a control system, so a Vectron and
an SM42 ran in multiple unit.
TMoverParameters::Attach() now leaves the control connection out between an
electric (ElectricSeriesMotor, ElectricInductionMotor) and a diesel
(DieselEngine, DieselElectric/DumbDE) traction vehicle, for the scenery, the
AI, the coupling keys and the automatic couplers alike, and logs it. Vehicles
without own traction (cars, control trailers) are not affected, nor are
permanent couplings within a unit, nor couplers that both name the same
ControlType, which stays the vehicle authors' way to allow such a pair. The
AI coupling order expects the coupling the vehicles can take, so it doesn't
wait for the control connection forever.
Existing scenarios coupling such a pair still load, without the control
connection. "controlcoupling.enginecheck no" in eu07.ini turns the check off.
No SVN data change is needed. (eu07/maszyna#520)
New optional Cntrl. key ScndCtrlResistiveLock=Yes: IncScndCtrl() refuses
to step the shunt controller up while the master controller is at its
no-power position or on a resistive RList position (R >= 0.01), as the
interlock of the ET21, ET22 and EP09 does. Decreasing the shunt stays
allowed so it can't get stuck. The AI doesn't choose field shunting
while the lock applies. Without the key nothing changes.
Game data: the key is to be added to the Cntrl. line of the ET21
(et21_v2), ET22 (et22_v3, et22-2000_v2) and EP09 (ep09_v1, ep09_v2)
.fiz files that define it; until then they behave as before.
(eu07/maszyna#402)
The shunt mode of the diesel-electrics (ShuntMode=1: SM42, SP42, SU42) now
suspends their vigilance device (CA): no alerter while running, no alarm for the
button held too long, and an alarm already on goes out. The cab signalling (SHP)
stays as it was. Back in the line mode the CA works as before.
CurrentSwitch() asked MaxCurrentSwitch() first. Since 2e86b3a5 (2021) that one
takes any change of the motor overload relay threshold, on every engine type, and
reports it, so the first move of the switch only flipped the threshold (meaningless
on a diesel) and the shunt mode of the SM42/SP42 (and the 2Ls150 gear of the SM03)
took a second one, both on and off; the cab switch, which shows the shunt mode,
didn't move at the first click. The diesels with a shunt mode now set it alone.
The 6Dg sets its power demand in 5% steps, while the universal controller
(EIMCtrlType 3) ramped it continuously. A new optional UCList header key,
CtrlValStep (e.g. 0.05), makes the setting move in whole steps: one at once
when the controller is moved onto a raising or lowering position, then one
each time SpeedUp / SpeedDown covers a step, stopping at the position's set
value. Without the key (default 0) nothing changes.
Data (SVN, optional, after this): add CtrlValStep=0.05 to the UCList: line
of dynamic/pkp/6dg_v1/6dg.fiz (its variants include it).
(eu07/maszyna#412)
The mainctrlact: gauge showed MainCtrlActualPos, which nothing sets for
DieselEngine/DieselElectric without an EIM controller, so the ST43's marker
never moved. For those it now follows MainCtrlPos (vehicle/mainctrlact.h);
every other vehicle reads what it did before.
New optional gauge key scaletable: [ value scaled, ... ] in the block form
gives a non-uniform scale, interpolated linearly between its points, in the
units of value * scale + offset; without it gauges behave as before.
Assets: none needed. To get the true ST43 scale, the line
"mainctrlact: stery_1 rot -0.016 0 0.0" of dynamic/pkp/st43_v1/*.mmd is to
become a block with a scaletable: of the 25 positions read off the cab's
scale; until then the marker moves linearly with -0.016 turn per position.
The limiter slept off the rest of each frame with sleep_for() given a float
duration. MSVC adds that to steady_clock::now() in float nanoseconds, so after
a few hours of system uptime the wake-up time is rounded to tens of ms: at
95 h uptime fpslimit 60 had no effect at all and fpslimit 30 gave ~22 fps, in
the full and the legacy renderer alike.
frame_pacer now schedules frames on a fixed cadence in steady_clock ticks and
the loop waits with sleep_until; an overshooting wait is made up in the next
frame. The FPS shown and the motion blur use the new mainloop_period stopwatch
(frame start to frame start, wait included); the targetfps/fpsaverage
draw-range controllers keep reading the work time as before.
(eu07/maszyna#416)
The smoke of a diesel came out at full strength the moment its ignition
set Mains and started the spin up, together with the ignition sound, whose
first part is the starter cranking the engine.
The ignition now starts a countdown of the new optional .fiz parameter
SmokeDelay (Engine section of DieselEngine and DieselElectric vehicles,
seconds, default 2.0, 0 restores the old behaviour); the smoke source of
the vehicle spawns nothing until it runs out. Engines already running, or
switched on another way, smoke as before. No SVN data change is needed;
vehicle authors may tune SmokeDelay to the cranking part of their
ignition: sample.
main_off_bt declared with "type: toggle" in the cab's .mmd is now the
line breaker's hold coil lever of the ET22 and EP09: it stays where it is
put. Off, it opens the line breaker and keeps it from closing (also
cancelling a closing armed by a held main_on_bt); back on, it only feeds
the coil again. The single keyboard key puts the lever back on before
closing, and a line breaker closed by the AI or through the control
cable puts the lever back on to match.
TGauge now remembers whether type: was given, since every control
defaults to toggle. Without type: on main_off_bt (all vehicles today)
nothing changes.
Data (SVN, not part of this commit): add "type: toggle" to main_off_bt:
in dynamic/pkp/et22_v3/*.mmd and dynamic/pkp/ep09_v1, ep09_v2/*.mmd.
Older exes ignore it.
Starting uphill the AI released the train brake (and the locomotive
brake with it) at once, while the power came on only later: a train on
10 per mille rolled back 20-30 m (E186 with 19 container wagons), and an
E189 with 7 Z2 coaches rolled back for minutes without power, because
the brake applied again for the rollback kept the power off. Many
locomotives cannot pull while their own cylinder is under pressure
(asynchronous drives switch to braking, others have a control pressure
switch), so holding the train on the locomotive brake is no way out.
A train standing on an up gradient with the train brake applied now
starts as a driver of a heavy train does:
- the train brake stays applied, the locomotive brake is released and
the locomotive's own share of the train brake is let off with the
releaser;
- the power is applied against the braked train (Ready does not hold
it back meanwhile, and the rule that brakes a train slipping back
is off);
- the train brake is released only once the tractive force exceeds 1.1
times the gradient force, the train moves forward, or after 45 s if
the locomotive cannot hold the train on its own power.
Multiple units are not affected.
Changes to how the AI works the train brake on the way to a lower speed
or a stop, so that it uses few, early applications instead of a late
deep one:
- Approaching a lower speed, the deceleration may fall short of the
demand by only half of a step (fBrake_a1) before the brake is applied
harder. With a full step the demand crept up meanwhile and a long G
train got its next step many seconds too late.
- The shortfall is checked every 3 s. If it keeps growing at a handle
left alone for longer than the brake takes to respond, the AI adds a
small step at once instead of waiting until the margin calls for a
deep one. A shortfall that shrinks on its own (brake still filling,
blocks gripping harder at lower speed) does not trigger it.
- Releasing, a position short of the first one counts as giving the
minimum stage of the distributor, not less.
- Holding the speed on a descent, the predicted excess speed is taken
off over the time the brake takes to respond, not at once, so a slow
brake does not get a deep application that then takes as long to
release.
Uphill a long train often came to a stand with its brake on the way
off and rolled back several metres: the AI released the train brake
to 0 about 20 m short of the stop (the gradient alone would have
stopped the train short of the stopping window), applied only the
first step again at the last moment, and a G train cannot apply its
brake again within the few seconds the gradient takes to stop it.
Coming to a stand on an up gradient (also when only part of the train
is on it) the AI now:
- applies the brake before the train stops on the gradient alone, once
the time left to the stand is no longer than half of what the brake
takes to apply;
- does not let the train coast on or pull up while the brake is applied
and the gradient would stop the train sooner than the brake of the
consist can be released;
- does not release the train brake any further below 10 km/h on the
way to a stand (unless the train is to be driven on);
- puts the locomotive brake on below 3 km/h when the train will reach
the point it is to stop at (not when it has slowed down short of it
and is to pull up), and at once if the train rolls back.
SetLights() gives RaLightsSet() the lights for the vehicle's end towards
the head of the consist and the end towards its tail, whichever way the
occupied cab faces. The coupling that switches an end off was looked up
at the opposite end whenever the cab faced the tail of the consist
(e.g. 6Dg driven from its second cab position), so the free end went
dark and the end with the train was lit.
The per-vehicle choice moves to lights_programmer::vehicle_lights() in
vehicle/lightsprogrammer.h, which checks the coupling at the same end
it lights, with unit tests for both cabs, both orientations and the
train at either end. Driving from a cab facing the head of the consist
gives the same lights as before. No data change needed.
The config key from eu07/maszyna#541. By default (yes) a pass reported to
__test.result waits until every vehicle has stood still for 1 s; with no it
ends the test at once. The key is listed with the other test.* settings in the
log and in the test start line.
Scenario tests: new case test_mode_pass_while_moving reports the pass from
event2 of tor1k as the train runs onto it and checks the run ended with the
train still moving (new check ends_moving). The README lists the test.* keys
with their defaults and the input the test mode ignores.
With --test the gamepad, mouse picking, UART, ZeroMQ and motion telemetry are
off, the hidden window gets no key, mouse, scroll or char callbacks, and the
window procedure that takes WM_COPYDATA commands from other programs isn't
installed, so nothing but the scenery drives a test run.
The test manager is reached through simulation::test() instead of a global,
the result cell is owned by a unique_ptr until the memory table takes it, the
cell value is compared with a margin, strings are built with std::format and
the --test settings moved to apply_test_mode_settings(). The driver's signal
check is folded into UpdateSpeedPosition(), which updates a speed table entry
and reports a signal at stop the head of the train just passed. The test.*
keys are listed with the other settings at the start of the log.
run.py starts eu07 with --test (and --test-timeout <duration> for cases that
record a fixed time) and compares the result from the exit code with the
case's expect key, before the physicslog checks. The 14 cases take about 30 s
with -j 4 instead of minutes. --watch keeps the old windowed run.
New case keys: expect, test_config (lines of a config block), track_events
(lines added to a track of the base scenery, e.g. event0 test_pass). New cases
show each end of the test mode: test_mode_pass, _signal_passed_at_danger,
_timeout and _standstill. The slippery rail cases slide past the red signal,
which now ends a test, so they set test.fail_if_stop_overran no.
README documents the test mode for scenery authors.
eu07 --test -s <scenery> [--test-timeout <s>] runs the scenery in ghostview with
the null renderer (models loaded, no textures), a hidden window, no sound,
network or python, and in fixed steps of 0.03 s of simulation time per frame
(three 10 ms physics steps) as fast as it can. Without -seed it uses a fixed
seed, so two runs give byte-identical physicslogs.
The scenery reports its result to the __test.result memory cell, which the
simulator always creates: value 1 passes once every vehicle has stood still for
1 s, value 0 fails at once. The run ends with exit code
0 pass, 1 fail,
2 a train passed a signal at stop (SetVelocity 0 0, as seen in its driver's
speed table; SetVelocity 0 x on automatic block doesn't count),
3 timeout (test.timeout, default 3600 s of simulation time),
4 standstill (every vehicle still for test.standstill s, default 300, with no
result; a pass with test.pass_on_standstill yes),
5 error (no -s, the scenery didn't load or has no vehicles, an exception).
The test.* keys go in the scenery's config block; test.fail_if_stop_overran no
only logs a signal passed at danger. Outside the test mode a signal passed at
danger is logged.
The null renderer gets a ui backend which draws nothing, hands out the null
texture and material instead of throwing, and ends the ui frame. A GL context
is still made for the hidden window, without asking for a version, as parts
of the ui create textures with gl calls of their own. No crash dialog or
message box in test mode; the window code no longer needs a monitor.
- physicslog rows keep all columns (text ones as None), checks can read
the wheelslip flag and the new vehicle speed column
- slips: the lead vehicle's wheels slip at some point (makes sure a
slippery rail case does what it's meant to)
- slip_clears: once stopped, the wheels stop slipping within N seconds
- holds_still: once stopped, the vehicle doesn't creep, by vehicle speed
- td_red_signal_moving.scn: the red signal case starting already moving
- eu07_slippery_red_signal / eu07_slippery_push_red_signal: skid on
braking for the red signal with track friction 0.03; the pushing one
fails without the fix for eu07/maszyna#479 (stays skidding, creeps)
- sn61_td_timetable: stop hold 5 -> 2 s, the railcar arrives 5-7 s
before its departure time, so 5 s made the case flaky
The Velocity column is the wheel speed, which stays at 0 while the wheels
are locked in a skid, so a vehicle sliding on locked wheels looked as if
it stood still. Add the signed vehicle speed in km/h as the last column.
- eu07_td_red_signal: no route set, the train has to stop before the exit
signal and stay there (new checks stays_stopped and distance; distance
is integrated from the speed, the physicslog has no position)
- en57/ep09/et22/sm42/sn61/vectron/eu07_td_timetable share one fragment
template, td_timetable.scn; a case sets vars.consist and the lead vehicle
is named {{vehicle}} = autotest_<case>
- -j N runs cases in parallel: each case has its own physicslog and, with
multiplelogs, its own log files, which are moved to out/<case>/
- a run stops at once on Bad vehicle / Bad trainset instead of waiting for
the timeout, and fails after 120 s without physicslog rows
- the .sbt geometry cache of a test scenery is dropped when the base
scenery is newer
All 8 cases pass in about a minute with -j 4.
eu07_td_timetable now takes about 20 s instead of 4 min. Its speed profile
at 20 and 60 matched a real-time run within a few km/h and seconds; the
physicslog keeps its 0.05 s rows. The runner polls the log every 0.25 s so
it does not overshoot the case duration by a minute of simulation.
tests/scenarios/run.py builds a test scenery from a game install's scenery
(own trainsets removed, case fragment added), runs eu07 in ghostview with
physicslog on, and checks what the vehicle did: starts, max speed, brakes,
stops. First case: EU07 on td.scn driven by AI on the td/rozklad timetable,
with the Stacja1 exit route set by an event launcher instead of shift+4.
eu07 gets a -ini <file> option: an extra settings file read after eu07.ini,
so a test run can switch to a window and turn on physicslog without touching
the user's config.
Local only: scenery and vehicle data are not in the repo.
* docs: CLAUDE.md with the team's notes for Claude Code
The rules Claude Code sessions kept relearning, shared with the team:
- the code builds and behaves the same with GCC 13 (Linux CI and the
Sonar build) as with MSVC, with the MSVC leniencies to watch for;
- every SonarCloud remark on the changed code gets fixed without
changing behaviour, with the recipes for the rules that keep coming
up;
- building, the unit tests, scripted editor runs, the game data and
the profile folders, and GPU hangs;
- the game data in the asset SVN and backward compatibility: existing
files keep being read as before, new entries are optional, code
falls back without new assets and a change that uses them says so
in its commit message, what the editor saves stays readable by the
current release, new features can be turned off in eu07.ini and
don't change the look of existing data abruptly; StarterNG is the
official starter.
* docs: CLAUDE.md on the move to OpenGL 4.6 and Vulkan
betterRenderer/ is to be removed and the OpenGL 3.3 and legacy
renderers are being replaced by feat/gl46-migration (PR #216, a draft
with milestone 27.04). Work on that code and on the shaders is kept to
fixes and to what a feature on master needs to be drawn, with the PR
saying where the branch has the counterpart.
---------
Co-authored-by: maj00r <maj00r@host>
The rule "wyluzuj lokomotywe, to szybciej ruszymy" released the
locomotive's own brake whenever power was asked for and the locomotive
cylinder held more than 0.4 bar. It had no speed condition, so it also
fired in motion, at 50-75 km/h, right after an application. On a
locomotive with the EStED distributor (E186) this vented the control
reservoir; with the pipe not yet recharged the locomotive brake then
faded to nothing at the next application.
The rule is meant for starting off, so it now applies only below
1 km/h.
The rule "wyluzuj lokomotywe, to szybciej ruszymy" released the
locomotive's own brake whenever power was asked for and the locomotive
cylinder held more than 0.4 bar. It had no speed condition, so it also
fired in motion, at 50-75 km/h, right after an application. On a
locomotive with the EStED distributor (E186) this vented the control
reservoir; with the pipe not yet recharged the locomotive brake then
faded to nothing at the next application.
The rule is meant for starting off, so it now applies only below
1 km/h.
On the electric locomotives the lamp of the open line contactors tells the start can be made, so it goes out with the
start cut off by the pressure switch. The diesel-electric SU45 and SU46 (the only diesel-electrics with the lamp) use
it just to show the contactors are open, but it went out the same way: with too low a pressure in the main pipe the
contactors were open and the lamp out. Once the pressure came back with the master controller past the first
position, the contactors stayed open, as they should, and the lamp lit up on a power position, which looked like
the lamp working the other way round.
On the diesel-electrics TMoverParameters::LineContactorsLampLit() now follows the contactors alone; the electric
locomotives and the other main circuit lamps are unchanged.
Tests: tests/unit/test_line_contactors.cpp, on a locomotive made after the SU46.
The SM03 had no way to stop its engine. Its fuel pump starts on its own, so the fuel pump switch is ignored; the
engine runs the converter feeding the low voltage circuit, so the battery switch doesn't starve the fuel pump either;
and its cab has only the starter button, for which TTrain::OnCommand_linebreakeropen() kept a legacy branch which
animated the button and returned without opening the circuit. MainSwitch(false) itself stops the engine fine.
The command now opens the circuit on diesels whose cab has only the starter button
(TMoverParameters::MainCircuitOpensWithStarterOnly()), without animating the button, which held down would start
the engine again. Other vehicles with only the button closing the circuit keep ignoring the command.
Tests: tests/unit/test_engine_shutdown.cpp, on a shunter made after the SM03.
* test: real vehicles and golden traces in the unit tests
The game's own vehicles (dynamic/ of the release used for the tests) as
JSON of what the game's vehicle loader reads of their .fiz files, includes
expanded: test_fiz.cpp converts them on demand and checks the loader reads
the same sections, values and rows as of the game's file (fiz.h). Chosen
to cover every engine type, brake valve, brake handle, local brake,
coupler type, door type and brake delay set the game data uses.
Golden traces (golden.h) characterise the simulation, compared within a
relative tolerance or exactly, re-recorded with
EU07_UNIT_TESTS_UPDATE_GOLDEN=1:
- the parameters of every real vehicle as loaded,
- the brakes of every real vehicle through charge, service steps,
release, overcharge and emergency, on every brake delay setting,
- every position of every driver's and local brake handle,
- every powered vehicle started up the way the AI does it, driven to
full power and braked to a stop (drive.h),
- the real electric units and a locomotive with wagons driven whole, as
trains of coupled vehicles stepped the way the game steps them
(consist.h),
- scenarios of the controls (scenario.h): doors, reversing and cab
change, slip and sand, emergency stops, releasers, the ep and the
electrodynamic brake, heating, loading, a collision, and frames of
several physics iterations.
* test: commands passed between vehicles
Every command RunCommand knows run on every real vehicle and recorded,
commands travelling through a train with a vehicle facing backwards, a
coupling without the control cable, and loading and unloading by commands.
* test: every switch of the cab, sent to the next vehicle
Each switch sending its state to the train, each way and to each range,
with the controllers and direction among them, on every real vehicle
driving another of its kind, recorded for both.
* test: the switches of the cab worked on a pair under the catenary
The same switches with two seconds for the vehicles to act on each, so
the pantographs rise, the main switch closes and the auxiliaries start.
* test: gauges of trains in multiple and the damage descriptions
The ammeters and the brake force the AI expects, recorded on the electric
units and locomotives in multiple driven, the damage described on the
driver's panel and ammeters showing the current of controlled vehicles.
* test: give load commands their place, as the game does
The loading test set its commands with SetInternalCommand, which leaves
TCommand::Location as it is: uninitialised in a fresh vehicle, so the
distance check of Load= and UnLoad= read garbage and the test failed
now and then. The game gives them through PutCommand with the place of
the platform, and so do the tests now.
* test: brake valves and handles no vehicle of the game data has
Fitted to a wagon and to an EU07: the valve variants and handles the
loader knows that no real vehicle uses, the CV1 ones and the test handle
among them, and a name the loader doesn't know.
* test: braking from speed with the automatic ED brake, and a draisine
Trains with the EP09, the E6ACT and the EW58 braked by ep, light and full
service applications from speed, and the draisine pumped from rest and
rolling. Only the E6ACT brings its electrodynamic brake in here.
* test: the EP09's electrodynamic brake in service braking held to a stop
start_up() took a vehicle for ready once its line breaker was closed and
its reservoir full, which a train made already moving is at once: the
EP09's manually started converter was never switched on, so there was
no 110 V and no electrodynamic brake. It now waits for the converter.
The EP09 with wagons is braked from 120 km/h by light, medium, high and
full service applications held to a stop: the motors brake, the air of
its cylinders is let out and given back as the train slows down. The
flag of the electrodynamic brake stays on at a standstill below 2.1 bar.
C strings given to CAPTURE are shown as text.
* test: the converter switched on before a train is ready, the EW58's ep-staged electrodynamic brake
start_up() called a train ready as soon as its line breaker closed and the air was there, so the converter
of a unit having it in another car than the cab (the EW58's middle car) was never switched on, and its
electrodynamic brake, gated by the 110 V, never came in. The converter and the compressor now always get
switched on first. The EW58 brakes electrically by the stages of the ep brake, not by the pipe alone.
* test: the ET42 and the 214M railbus among the real vehicles
The ET42's halves (its own series motor branch: the motors in series at the high current setting) and the
214MA, a diesel railbus with the hydraulic torque converter, the retarder and the controller of an induction
motor drive, converted from the game data and recorded in every sweep of the real vehicles.
* test: the wagons' hand brakes released with the train's brakes; the 754, SM31 and SU46
release_brakes() released the hand brake of the driver's vehicle only, so every wagon set up braked kept its
hand brake on and got dragged along. It now releases the hand and spring brakes of the whole train, as
TController::AutoRewident() does when the AI takes it over, and the traces of the trains with wagons are
recorded anew.
The 754, the SM31 and the SU46 join the real vehicles for the kinds of the diesel-electrics' field shunting
relays the others lack (RelayType 0, 1 and 46).
* test: traction on a lost and too high catenary voltage, the ET42, the diesel-electrics' shunting relays
The line breaker opened by the voltage relays of a series motor locomotive and by the limits of an induction
motor one. The ET42 at the high current setting, which keeps its halves' motors in series, and its switched
electrodynamic brake, whose current trips the overload relay from the top speed of the low setting. The field
shunting of the diesel-electrics through each kind of their relays, at full power and notched back.
* test: the EW58's electrodynamic brake stages, the EU47's speed control and braking at the voltage's limit
The stages of the EW58's motor brake set by the positions of its ep handle, the first one alone over
100 km/h. The EU47's second controller holding, raising and lowering the speed set, its local brake setting
the flag of the electrodynamic brake without the motors braking, and its brake feeding back less current into
a catenary over the drive's MaxV.
* test: diesel engines too cold and too hot to start, the 214M railbus's controller and retarder
The heat of the diesel engines: at -10 degrees the ones with the lowest water temperature set don't start
until the heater warms them, at 88 degrees the ones with a second water circuit don't start while it cools,
and the SM42 run hot at full power is kept under its highest temperature by its fans and shutters. The 214M
railbus's controller from the air brake through the retarder and neutral to power, its torque converter's
lock-up and its gearbox.
The controller helpers of the traction tests (full power, waiting for the line contactors, and notching back)
move to test::scenario.
* test: couplers coupled, uncoupled, broken and colliding; loads assigned, loaded and unloaded
Coupling by hand touching or apart and of the same kind or not, with the sounds of what changes; uncoupling a
screw coupler pressed together, stretched, with the hoses alone or broken; the automatic couplers of electric
units coupling on contact, without the control cable between two kinds; a coupler broken with its brake
hose; collisions over 15 km/h derailing. The pantographs set up by the pantstate load, a load of one kind at
a time up to the capacity, and the steps of loading and unloading.
* test: the battery drained and charged, the pantographs' tank and pressure switch, the 214M's speed control
The EU07's battery drained without the converter by the lights and the small compressor until it opens the
line breaker and nothing starts, charged by the converter, and switched off. The pantographs on their own
leaking tank until the pressure switch opens the line breaker, and the small compressor pumping it up and
priming the switch. The 214M railbus's speed control set by its buttons, steps and power limit, braking down
with the retarder, and standing by.
* test: the MHZ handles' buttons, the valves at speed, every brake material, the NESt3's relays
The overload and the high pressure buttons of the MHZ handles. The KE and the LSt at speed in the R and P
settings, under the anti-slip request and the releaser button. Every material of the brake shoes and disks
the loader takes, braking a wagon from speed, the ones no game vehicle has among them. The CV1_R and the
Oerlikon valve's relays set by its name (-ED, AL2 without ESt3) fitted to a wagon.
* test: the loader's lists and lines: the 218M, SA134 and 22WEc, a compressor list, Pmax by speed, skipped lines
The 218M (V2NList), the SA134 (HTCList) and the Elf 22WEc (ffBrakeList) join the real vehicles for the lists
the others lack. Definitions made by the tests for what no game file has: a compressor list, whose table of
whole numbers reads a factor of 0.9 as 0; the compressor's limits apart for the rear cab; the highest power
of an induction motor drive by the speed, extrapolated past the table; and the lines the loader skips or
trims (comments, a hash anywhere, leading spaces, carriage returns).
* test: what the cab and the AI call on a vehicle: gauges of the train, the split ED brake, spring brake, cut-off
The engine rotation of up to three vehicles in multiple and the lamps of the overload relay and the starting
resistors read through the control cable; the engine's idle speed and the ET42's second controller shown in
its dynamic brake; the controller of the electrodynamic brake apart from the local brake, which no game
vehicle has yet; the spring brake shut off and released, the ep brake's fuse and the handle stepped by the AI;
damaged motors cut off; the cab activated by itself; a road vehicle moved across a road, for which the tests'
track gets a width; and the lamps and the buzzer of the security system left unattended.
* test: the EP05 and its camshaft among the real vehicles, the headlights' colour and brightness
The EP05, whose camshaft closes the line breaker and the line contactors by the camshaft's own position,
converted from the game data and recorded in every sweep of the real vehicles. The Headlights section, which
only the 214M's own file has, in a definition made by the test.
* test: what the unit tests reach of the simulation core, what's dead or left, and what they found
COVERAGE.md: the blocks of Mover.cpp, hamulce.cpp, Oerlikon_ESt.cpp and friction.cpp the tests run, the code
nothing in the game runs, the code only data no game vehicle has would run, what's reachable and left, and
the behaviour the tests found that may deserve a look.
* test: check invariants of the physics on every step of every golden trace
Rules any vehicle keeps whatever a test does (no NaN, cylinders and pipes
within their limits, no current with the line breaker or contactors open,
no force without power, distance as the speed runs, no acceleration over
5 g), checked on every step, with no comparison to a recorded file. The
traces where a test or a game vehicle breaks one are excused with the
reason; the vehicles' cases are in the findings of COVERAGE.md.
* test: key figures of every golden trace, per part of the run
Duration, speeds, distance, time to stop, the brake pipe's lowest
pressure and time to recover, the peaks of the cylinders, the motor
current and the tractive force, for each part between two events, in
data/golden/<name>.figures. Compared within 0.5% or the last decimal, so
they hold on any compiler, and an intended change of behaviour shows in
a few readable lines. Traces sampled by hand keep none.
* test: keep the golden samples out of the repository, recorded on the code before a change
The samples of the golden traces are 29 MB and only mean something against
the code before the change they check, so they're recorded locally before it
(EU07_UNIT_TESTS_UPDATE_GOLDEN=1) and compared after it, and ignored by git.
The key figures and the invariants stay, and are what the CI checks.
Without samples recorded a trace compares only its key figures, but with
EU07_UNIT_TESTS_GOLDEN_EXACT=1, the check of a change which shouldn't alter
anything, it fails, so the check can't pass by having nothing to compare.
* fix: draw the noise of the control pipe pressure in one order on every compiler
The order of the two Random(10) calls in 'Random(10) - Random(10)' is
unspecified: GCC draws them the other way round than MSVC Release, which
mirrored the control pipe pressure around 5 bar and failed 32 unit tests
on the Linux CI. The draws are sequenced in the order of MSVC Release,
which the golden traces were recorded with.
* test: count the start-up rounds with an integer, not a double (SonarCloud S2193)
* test: behaviour in CHECKs, the key figures out of the repository
The key figures of the golden traces are recorded with their samples on the code
before a change, and compared after it, instead of being kept: a trace without
them checks the invariants and its test's CHECKs, which is what the CI runs.
The scenarios of the controls state what they show in CHECKs, and the runs of
every real vehicle what holds for each of them. Writing them showed scenarios
which didn't do what they were named for: the cab change stopped in the engine
room, the slip runs put the controller to full power at once and never got the
line contactors closed, and the train stopped by the alarm chain never got its
pipe charged again. They do it now; the rest went to the findings.
* test: the SonarCloud findings in the unit tests
std::format for the texts put together, emplace_back, const references, string_view parameters, std::array and std::vector for the C arrays, the ranges algorithms, names no longer shadowing the coupler ends, comments in the empty loops; start_up(), trace::check() and the converter of the definitions split up. BrakeForceR() is const, which it always was in effect, so the gauges test needs no const_cast. The golden traces compare bit for bit.
* Text displays: accept characters above 127, in windows-1250 or utf-8
A text display (sub-model of type "text") looks up the sub-model of each character
in a 256 entry table, and indexed it with a plain char. With char being signed, any
character above 127 gave a negative index: reading the text went outside the table,
and so did filling the table if a character sub-model had such a name. In effect the
displays could only show ASCII, and a text with national characters could crash.
The lookup is now keyed by unicode code point. Both the displayed text and the names
of the character sub-models are decoded as utf-8 if they are valid utf-8, and as
windows-1250 otherwise, so legacy (ansi) sceneries and models work next to utf-8
ones, and the two can be mixed. A short ansi text can be valid utf-8 by accident
(two upper case accented letters in a row will do); when both readings are possible
the one which finds more of its characters in the model is used.
The rest is unchanged: the first character of the text still isn't displayed, the
characters are still drawn in the translucent pass only. A missing text pointer no
longer gets dereferenced.
* Text displays: draw opaque characters in the opaque pass
The characters of a text display were only laid out in the translucent pass, so a
display had to be made of translucent sub-models to show anything. Those don't write
depth: overhead wires and other translucent geometry drawn later showed through the
display, and translucent geometry in front of it could end up painted over.
The opaque pass now lays the characters out the same way, so a display built from
opaque sub-models is drawn with depth like any other opaque geometry (and casts
shadows). Translucent characters keep working as before.
A model with a text display shows the text of its own instance, so such models are
excluded from the GPU-instanced opaque path.
* Text displays: prepare the characters once for each text
The sequence of character sub-models was cached for the last text only, so
displays of several instances drawn in turn decoded their texts again on
every draw. It is now kept for each text the display shows; the texts of the
instances don't change, so a draw costs a single lookup.
Also the comments added in the renderers are now in English.
* Text displays: move the text decoding to utilities
utf8_to_utf32() and win1250_to_utf32(), with the windows-1250 table, now sit
next to win1250_to_utf8(). The existing function maps only the Polish letters
and passes other bytes above 127 through, so it can't serve the displays.
* Text displays: address the SonarCloud findings
std::array instead of the C-style arrays, the utf-8 sequence length worked
out in a function of its own instead of nested conditional operators,
std::byte for the bit operations, std::ranges::count_if and contains().
The character loop of the renderers moved to Render_Alpha_text(), which
takes the nesting back to the limit, and the static pasText is accessed
through the class.
* Text displays: move the opaque character loop to Render_text()
The same as for the translucent pass: the loop sits in a function of its
own and the static pasText is accessed through the class, which keeps the
nesting within the limit.
* Text displays: access the static iAlpha through the class in the opaque pass
* Text displays: transparent lookup of the text layouts, init-statement for the lookup result
Two SonarCloud findings left in the code of #198: the map of the text layouts
gets a transparent hasher with std::equal_to<>, and the result of the lookup
is declared in the if statement.
* perf(instancing): skip a cell's buckets lying wholly beyond the range of their instances
* perf(instancing): go only through the instances drawn one by one in the per-node pass
* perf(instancing): leave out of the color pass the cells holding only instances too far to be drawn
* refactor(instancing): keep the caches of the instances of a cell without mutable fields, set by the renderer when the cell is queued
The editor writes the turntable controls on each save to <scenery>_obrotnice.scm and includes it. A copy of a
scenery, saved under another name, still included the file of the scenery it came from. Its save wrote a file of
its own and included it too, so each turntable had its events twice.
Now the save leaves out of the scenery the other files of turntable controls it includes, told by the line the
editor writes at their start. Their files stay on the disk, for the scenery they were made for. A file of
turntables made by hand, without that line, stays included.
Co-authored-by: maj00r <maj00r@host>