3544 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
maj00r	9343ebcb13	feat(track): frogs and crossings for every rail profile; grooved rails share the floor of their grooves where they cross and need no check rails, the V beside the rail going through is one block where the lofts of the rails overlap; the nose of an acute frog of flat rails is filled between its legs as far as their lofts overlap, like a cast insert	2026-10-17 23:28:41 +02:00
maj00r	59f1a5a144	fix(track): the lowest of two height lists also takes the points where they cross, so the top of an obtuse point rail is not cut narrower than its sides; the end of a short neighbour track joined at a crossing is the nearer one, not any within half a metre	2026-10-17 23:28:41 +02:00
maj00r	215f16e871	fix(editor): self-test frog takes the throat, nose, wing end and check rails from the sizes the rails are laid with, so the wing end is logged where it is laid; self-test crossings tells crossings at middle angles apart and counts them	2026-10-17 23:28:41 +02:00
maj00r	1a3a38e96d	refactor(track): one plan projection, cross product and meeting of two lines in plan for the crossings of rails, the frog of a switch and the check rails, in place of copies in each	2026-10-17 23:28:41 +02:00
maj00r	405081e81c	refactor(track): the rails of the two paths of a switch are laid by one loop for left and right switches alike, the side of the blade of a path and the blade itself come from one place, used when a switch is made, thrown and its frog found	2026-10-17 23:28:41 +02:00
maj00r	809b48c2ab	refactor(track): the blades of a thrown switch are drawn again with their cut faces by one helper instead of four copies; self-test command switches sets every switch to a path and moves its blades there at once	2026-10-17 23:28:41 +02:00
maj00r	e324eb10f9	perf(track): pieces of gauge lines whose bounding boxes do not meet are skipped before they are intersected; self-test railtime logs how much of the time goes to finding the crossings	2026-10-17 23:28:41 +02:00
maj00r	36fb68d8f9	perf(track): the gauge lines, crossings and check rails of a track are kept until a track is made, rebuilt or removed, so laying the rails of a track no longer recomputes its neighbours' crossings for each wing rail; self-test command railtime times planning the rails of all tracks	2026-10-17 23:28:41 +02:00
maj00r	929294af24	feat(editor): self-test command guards logs each guard rail and what each of its ends and breaks is cut along: the rail of the other track, how far off its gauge line and how skewed; a fault when a cut is not along a rail, or a glued end of the frame of a crossing at a middle angle is not on the face of a guard rail of the other track	2026-10-17 23:28:41 +02:00
maj00r	7088618d90	feat(editor): every switch laid in the editor gets the key frogs, a box in the Switch panel sets or clears it as one undo step; the editor rebuilds the tracks crossing a changed or removed one; self-test commands frog, crossings, frogkey and geomhash; Polish translation	2026-10-17 23:28:41 +02:00
maj00r	2159d682f2	feat(track): rails are cut where they cross other rails: a switch with the key frogs draws its frog, wing rails and check rails from its rails, tracks crossing other tracks get acute frogs with wing rails and check rails, obtuse points, guard rails along the quadrilateral of a crossing between 20 and 45 degrees and broken rails with guard rails at wider angles, the track going through chosen by name; sizes come from the rail profile, only for the profiles default, s60, uic60 and s39, others are laid whole; rail ends left open are closed with their cross-section; switches of older sceneries, with check rails in their includes and no key, are drawn as before	2026-10-17 23:28:41 +02:00
maj00r	7e1b2c553c	feat(renderer): a matte variant of a material for the parts of rails no wheel rolls on: it lights the same maps as weathered steel, with the weathering scaled per vertex by userdata, so a polished running surface fades into unused steel without a bright stripe; the cut faces of rails are drawn in it in an extra pass	2026-10-17 23:28:41 +02:00
jerrrrycho	c7f9796387	fix(ai): release horn when human driver takes over from AI (#210) ... * fix(ai): release what the AI was holding when the player takes over * fix(ai): keep the light selector lookup within its table, Sonar findings of the handover - the light selector size comes from the vehicle definition (LightsList Size) and isn't limited, while the table holds 18 positions; a bigger selector read past the table, and so did a table index other than the front or rear one. Only the positions the table holds are looked at now, and the selector stays where it is for a bad table index (cpp:S2259), - the lambdas of the handover capture what they use explicitly (cpp:S3608), - the chained assignments resetting the headlight state on the AI taking over are split (cpp:S1121).	2026-10-16 14:44:30 +02:00
jerrrrycho	b400255dae	Test/physics unit tests (#213) ... * build: doctest 2.4.11 in ref/ Single header test framework (MIT), vendored like the other libraries in ref/ so the tests need nothing extra installed on the CI runners. * refactor(physics): read vehicle definitions from text as well as files TMoverParameters::LoadFIZ(path) opens the file and hands the parser over to LoadFIZ_Parse(), which holds what LoadFIZ() did with it before. The new LoadFIZ_Text() reads the same format from a string, so a definition can be made up in memory, e.g. by a unit test. No change in behaviour for files. * build: the simulation core as a library the game links The vehicle physics (McZapkie), the vehicle definition parser and the helpers they use build as eu07_core, a static library the game links, so the unit tests to come can link the very same objects instead of compiling their own copy with their own settings. The game's build settings (include dirs, definitions, compile options, libraries) move from the executable to eu07_settings, an interface target the library links publicly, so the game, the library and later the tests all build with the same ones. What only the executable needs stays with it: /MAP, /LARGEADDRESSAWARE, output dirs, the precompiled header, runtime DLL copying, the crash reporter and the experimental renderer. No change in what the game builds or links. * test: unit tests of the simulation core, run as part of the build eu07_unit_tests: C++ tests (doctest) of the simulation, which set up the conditions, run the code and check the outcome. No game data, no window. - the code under test is the core library (eu07_core), the very objects the game links, built with the game's settings; engine_stubs.cpp stands in for the logs, settings and scenery layers the library reaches for. python gets linked with the settings but isn't needed to run the tests - test vehicles are JSON definitions in tests/unit/data, with the sections and keys of the game's vehicle definitions (tables as "rows", _keys are comments), changed by the tests where needed and loaded from memory by the game's own loader; an unknown section is an error rather than silently skipped - test::vehicle creates the mover the way TDynamicObject::Init() does and steps it the way the simulation loop does, on a straight track of given friction and gradient, optionally holding the brake pipe pressure the way a coupled locomotive would - the random numbers of the simulation are reseeded before every test case, so a test gives the same result alone (IDE) or among the others - building the game runs the tests at the end, once the game is linked, and prints a one line summary ("Unit tests: 11/11 passed (80 checks)"). A failure prints the full report and fails the build, in CI and locally alike, and removes the game binary so the next build reruns them. A change of the tests alone relinks the game so they run (Ninja and Makefile generators). -DWITH_TESTS=OFF leaves them out - first tests: definitions reach the vehicle, KE service application and release, deeper reduction brakes harder, every implemented distributor valve charges/applies/releases, service braking stops a moving wagon * test: stub basic_node::radius_() for GCC and Clang The Linux build failed to link the unit tests: undefined reference to scene::basic_node::radius_() from basic_node's vtable in engine_stubs.o. basic_node has no key function for the Itanium C++ ABI (tooltip() is declared plain but defined inline in the header), so GCC and Clang emit its vtable in every unit using scenenode.h, and the vtable needs radius_(), defined in scenenode.cpp, which the tests don't link. MSVC emits vtables differently, so the Windows build linked. * test: test vehicle definitions load with GCC and Clang All the tests reading a definition failed on Linux with "unknown vehicle definition section 0": the parsed definition was brace-initialised (auto const json{parse(...)}), which GCC and Clang take for the initializer list constructor of nlohmann::json, making an array holding the definition, while MSVC copies it. Initialise it with =, and require the definition to be a JSON object. * test: SonarCloud issues of the unit tests - test::definition_error for unusable test vehicle definitions instead of std::runtime_error (S112) - the rows of a table section written by a function of their own, less nesting in definition::text() (S134) - std::less<> for the section table (S6045) - comments in the empty methods of the reseeding listener and stub (S1186) - NOSONAR where the engine's headers fix the stubs: the extern globals and the member function signatures (S5421, S5817, S1238) - NOSONAR on LoadFIZ_Parse(): the cognitive complexity is the former body of LoadFIZ(), moved as it was (S3776) - a comment Sonar took for code reworded (S125)	2026-10-15 17:56:41 +02:00
maj00r	fb73a3e011	test(editor): self-test steps that keep a laying or connection ghost on screen under a set cursor, open the switch template editor, release a held placement and leave the top view, for screenshots of the tools at work	2026-10-14 23:55:04 +02:00
maj00r	aa20a88ed0	feat(editor): connections of tracks laid whole from the switch tool: a crossover with a straight insert or an aligning curve between switches of different angles or tracks that aren't parallel, a double crossover meeting on either track, a scissors crossover with the crossing of its inserts, and a switch ladder onto the parallel tracks of a fan or onto new ones at given spacings; press on a straight and drag onto the other track, Shift+wheel the kind, Ctrl+wheel the R of the aligning curve, Alt+wheel where a double crossover meets; the ghost shows the whole set with the spacing, the inserts and the distance of the mathematical points; switches get their drives and a //$r connection record; the crossover drawn from a diverging end uses the same solver; self-test connection	2026-10-14 23:38:31 +02:00
maj00r	52a289b15b	feat(editor): switch templates by their geometry and a template editor: the diverging track is a chain of straights, arcs and transition curves from the start of the switch, with the pre-blade stretch and the angle of the blade at its tip; own templates are kept in eu07_switch_templates.txt and listed in the switch tool; double switches Rpd and Rpj are laid as two switches joined by the main track and handled as one set by sliding, replacing, flipping, the gizmo and deleting, with drives for both switches; self-test switchtemplates	2026-10-14 23:17:30 +02:00
maj00r	ea341af27a	feat(editor): a switch path can be a chain of pieces: the attributes pieces1 and pieces2 keep the parts of each path after the classic bezier, which stays as the approximation older versions read and log the attributes as unknown; the simulation drives, draws and maps a path along its pieces by distance, pieces that don't join the ends of the path are reported and the bezier used; the catalogue Rz switches get the diverging path as the arc of their radius and the straight after it, moving, rotating and mirroring a switch carry the pieces, reshaping a path by hand drops them; self-test commands switchpieces and switchbreak build-feat-track-join	2026-10-12 18:32:23 +02:00
maj00r	46d41549c9	feat(editor): the profile takes the switch zones from the //$r records: the parts of a double slip and the paths of a crossing make one zone from its first part to its last, a crossing made of plain paths is a zone too, the zone is named by the number and the type of the switch; a switch with no record stays a zone named by its path; self-test vprofile logs the zones	2026-10-12 18:11:17 +02:00
maj00r	9b75e64c9b	feat(editor): digitizing takes the good parts of a line: a run whose design can't be read whole is split at the track the reading fails at and each part is digitized on its own, only single tracks that make no line are left out and named in the log; Digitize all reports the lines digitized and the paths left out, opening a line digitized in parts says so with the reason for the paths left out; a stored radius agrees with the one its design makes within 0.1 %	2026-10-12 18:00:26 +02:00
maj00r	95c3842748	feat(editor): digitizing takes a line whose tracks lie up to 0.5 m off the curves read from them: the record keeps how far as fit in its source line and counts the tracks within it as lying on it, so they aren't shown as departed; a record whose elements differ from the ones its design makes only by elements of no length is no longer reported as differing	2026-10-12 17:42:40 +02:00
maj00r	096122bcc3	feat(editor): digitizing reads more of the old tracks: curves of the same hand meeting through their transitions with no straight between are separate vertices, a run turning over 150 degrees is split into vertices at track joints, a curve made of two transitions with no arc is read with the radius at their joint; digitizing tries both transition shapes, then without the minimum arc length when a short curve would lose its transitions; the reason names the track left by how much at any distance	2026-10-12 03:45:12 +02:00
maj00r	f5895abfa8	feat(editor): tracks drawn without a design get digitized: the route design keeps the line it opens in the scenery when no design is kept for it, or says why it can't (straights meeting with no curve, the curves read leaving a path by how much), trying the other transition shape before giving up; Digitize all the tracks in the route design does it for every line of the scenery and reports the count; a kept design counts as departed only where no path lies along it any more, so a parallel track or a diverging one beside a digitized line no longer shows it as left by metres; self-test command digitizeall	2026-10-12 03:24:23 +02:00
maj00r	696168a132	feat(editor): fewer hints in Tracks: the bar in the view is one line with a short mode name and at most three keys, fading when the cursor comes near; the tool options show only the mode name with the full list under (?), the inspector says Nothing selected; the hints name what is dragged instead of its colour or shape, the break is press and drag, the S-curve its four clicks, the turntable and the signal with no type chosen have their own; Esc back to Select only in the full list; the gizmo settings are hidden in Plan; in Polish a track is tor everywhere, odcinek only for a piece being laid, an isolated section and a path of a switch	2026-10-12 02:59:30 +02:00
maj00r	f81698eaf1	feat(editor): the top view zooms out to 50 km for a whole line, from 0.5 m per pixel the rail tracks are drawn as thin lines and the hover reaches 16 px around them, only the tracks near the cursor are checked; handles drag beyond 200 m from the camera in Tracks; Plan works with the mouse alone and without the gizmo: the grips of a straight, the vertices and the radius of a curve drag directly, the break tool is press on the straight and drag, the panel tool stays chosen afterwards, the Q/W/E bar is hidden in Plan; self-test command ortho	2026-10-12 02:49:07 +02:00
maj00r	16c90eb674	feat(editor): Straight and Curve are one Plan tool (G, also C) right after Lay track, a straight opens with its grips, a curve or a line with its vertices; digits no longer type a length while extending from a free end; the scenery wizard chooses realistic or fictional, a realistic one gets //$g and //$o and opens with the orthophoto in the top view, a fictional one only the top view with the grid; the orthophoto settings are kept in the scenery as //$o (on, year, opacity, radius, height, drape, lift, hires, scene, origin) and survive the new formats; self-test command mapstate	2026-10-12 02:21:29 +02:00
maj00r	653d206a54	feat(editor): the vehicle tool places trainsets: what is placed on a named track is written to the scenery on save as trainset <timetable> <track> <offset> <speed>, with reversed when it faces against the track, listed until the save with x to leave one out; the timetable and the starting speed are set in the panel, a tick turns writing off for a vehicle only to drive; the starter WMB10 is only added when nothing was placed; self-test command trainset	2026-10-12 02:00:50 +02:00
maj00r	c9f381a9c9	feat(editor): the parallel track is a tool of its own (N) in the toolbar, its green preview only shows there, Shift+click on a straight in L opens it with that track; the norms are two buttons at the top of the Lay panel and of the Curve tool, shared by both, free hides the recommendations of the curve and its norm warnings, the duplicate transitions checkbox is gone; the Lay panel starts with the radius, the speed and the shape; the Tracks toolbar shows the tools that lay track, the ones that correct a line are under More (remembered); First steps ends with driving the track laid; self-test command tracktab	2026-10-12 01:52:28 +02:00
maj00r	7fea51ad06	feat(editor): norms of laying, free by default: free lays plain arcs with no transition curves and no cant and keeps quiet about the norms, PKP PLK gives the transitions and the cant for the design speed with the warnings, chosen at the top of the Lay panel; digits in Tracks only type values, never change the edit mode; in Lay a track under the cursor says the click starts on the ground beside it, Shift+click chooses it for a parallel track; Alt while laying turns the direction in steps of 15 degrees, with the azimuth by the cursor; K splits only within 20 px of the selected track, where a marker shows it; the environment, overhead line, transition curve, view and include labels translated, one name for laying track in Polish	2026-10-12 01:35:39 +02:00
maj00r	f28640dd44	feat(editor): laying track for a beginner: Ctrl+Z and Backspace during a chain of laid pieces go back to the end of the piece before instead of carrying on from a removed track (Ctrl+Y ends the chain); every warning of the laid shape and of a join is shown by the cursor for 4 s after the click, also when the radius chosen on the wheel didn't reach the cursor and the arc goes through it, the preview of a shape with a warning is orange as it is laid anyway; the switch type stays chosen after a switch is placed; the status says each click lays a piece; a "First steps" window (Help menu, opens with the editor until unticked) with eight steps ticked off as they are done; an empty scenery opens in Tracks with the camera over its origin and a grid where there is no ground under the cursor; F1 lists the laying gestures (Ctrl/Shift+wheel, drag from a free end, typed length); self-test command laychain; Polish translation	2026-10-12 01:21:50 +02:00
maj00r	866f626668	fix(editor): K splits the selected track where the cursor points on screen (as the hover highlight finds it), not at the delayed depth pick which hit the ground beside the rail	2026-10-12 00:44:54 +02:00
maj00r	704511fe5e	feat(editor): the plain straight a switch branch leads into gets its end moved to where the branch lands instead of a short straight being added: shortened when it reaches past, lengthened when it ends short; straights drawn with handles count (the trimmed end loses them), so a switch slid up to a track that ends close to the crossing now places	2026-10-12 00:36:13 +02:00
maj00r	b27d1772a7	feat(editor): a switch being placed slides along its straight first, so its branch lies on the line of the track it leads into (a straight between, or an arc of the switch radius right after it with the arc ending at the track's end); only when the slide leaves the straight it stays put and reverse curves fill in; the straight between reverse curves takes the speed of the switch's branch (R<300 40, <500 50, <760 60, <1200 80, <2500 100 km/h) instead of the track's	2026-10-12 00:25:01 +02:00
maj00r	d4d2a02018	feat(editor): the switch no longer slides to meet the track, only the straight and arc fill in: arc of the switch radius between two straights, and where the angles match but the tracks are apart, or the arc doesn't fit, reverse curves with no transitions and the largest radius that fits; the straight between them (6 m to V 60, 0.15 V to 100, 0.25 V above) only when 1/R1 + 1/R2 > 1/106.5 m (PLK ST-T1 9.3, table 9.4); new pieces take the style of a plain track, not of a switch	2026-10-12 00:09:49 +02:00
maj00r	0a73550c8d	feat(editor): the branch of a switch leads into the free end of a track: when the angles match the switch slides so its branch lies on the track's line, otherwise an arc of the switch radius right after the switch turns into it (with a straight between for a parallel track); works while placing a switch (drag toward the end) and when dragging from the branch end of a placed one, the track is shortened when it reaches past the branch; self-test switchlink / branchlink	2026-10-11 23:53:18 +02:00
maj00r	05ffb9058c	Merge pull request #212 from MaSzyna-EU07/fix/targetfps-default ... fix(gfx): targetfps off by default again	2026-10-11 23:27:54 +02:00
maj00r	eaf7e787b8	fix(gfx): targetfps off by default again, draw range stays at heightbase instead of being driven by measured fps (with vsync on a 60 Hz monitor it sank to half); targetfps in eu07.ini still turns the regulator on	2026-10-11 23:20:08 +02:00
maj00r	f1a3bde00f	refactor(editor): one declaration per statement (cpp:S1659) and no chained assignments (cpp:S1121) in the editor files of the branch; single statements under if/else/for that became two got braces	2026-10-11 22:28:35 +02:00
maj00r	d2111afe19	refactor(editor): alignment::join builds its candidates (straight, single arc, fitted curve, Bezier fallback) in separate functions, alignment_store::load reads each //$h line kind in its own function	2026-10-11 22:23:21 +02:00
maj00r	84cec61912	refactor(editor): the self-test runner dispatches its commands to seven groups (scene, bends and lineside, tracks and routes, switches, joining and gizmo, signals, terrain) instead of one function with 95 branches; deletetrack without a name deletes the selected path, which its second, unreachable branch meant to do	2026-10-11 22:17:22 +02:00
maj00r	3e9d5836f4	refactor(editor): the //$e header scripts share reading, decoding, the rewrite of the header, the backup and the zip in tools/editor/editor_headers.py instead of three copies; reports, zips and written files unchanged	2026-10-11 22:11:44 +02:00
maj00r	4030797b21	refactor(editor): the layouts laid from a free end (arc then straight, straight then arc, arc, straight) and the transition curve out of a curved neighbour in editor_track; joining and laying through points share the eased ends, the room check and the assembly of the pieces instead of two copies	2026-10-11 22:09:53 +02:00
maj00r	c5ac698702	refactor(editor): split the track mode into its gizmo, kept records, switch drive, switch slide and track bed files; path geometry shared in editorGeometry instead of four copies of the reversed path, track shape rules (stretching, rolls at joints, middle, switch axis, drive spot) in editor_track, fitting the track to the ground in ground_follow with the terrain query passed in, the record layer and the switch classification in the stores; long gizmo, record watch and switch slide functions broken up, out-parameters replaced by optional results	2026-10-11 22:03:59 +02:00
maj00r	f34056ff5a	fix(editor): the key launcher of a turntable reaches at most half way to the nearest other turntable (up to 150 m as before), so T and Shift+T turn only the turntable at hand where several stand close together, as on sk.scn	2026-10-11 21:33:28 +02:00
maj00r	777d1fbfe4	feat(editor): turntable_headers.py writes //$e descriptions into the turntable templates of the scenery that have none (eng/obrotnica21m.inc, linia61/cze_obrotnica.inc): category turntable, length from the obrotnica<N>m model, the fixed positions read from the torNrot events, so the editor places them like its own; --zip packs every described turntable template, the editor's obrotnica16m_e and obrotnica21m_e included	2026-10-11 21:33:28 +02:00
maj00r	33360eb4e2	fix(editor): drive_headers.py writes height 0.05 for the electric drives ele and ele1, which looked 5 cm too low on a switch laid in the editor	2026-10-11 21:33:28 +02:00
maj00r	080e308d5e	feat(editor): picking a drive or its side in the Switch mode panel changes the drive of the selected switch at once, no drive takes it away, each as one undo step; self-test command drivechoice; Polish translation	2026-10-11 21:33:28 +02:00
maj00r	b4a1cb03b3	fix(editor): drive_headers.py writes height 0 for every switch drive, as all drives of the digitized lines stand at the rail foot of the start of the switch	2026-10-11 21:33:28 +02:00
maj00r	71bdf8e284	fix(editor): a switch drive stands at the height given in its //$e description, switch: {height: ...} above the start point of the switch (rail foot): placed with a new switch, swapped for another drive or side, or following its switch; drive_headers.py writes height 0.15 for ele and ele1 and 0.05 for mech and mech1, whose models reach that far below the origin, so they stand on the sleepers; the drives with a counterweight (mech_*, mech2, 0.57 m down) keep 0, as all 1965 drives of the digitized lines do	2026-10-11 21:33:28 +02:00
maj00r	b103a561e0	fix(editor): the drive of a switch follows the switch: moved, turned, its grade or cant changed with the gizmo or any other change of the switch, the include gets the new position, direction and rot.x/rot.z of the start of the switch, in the same undo step; self-test drives prints the parameters of the drive	2026-10-11 21:33:28 +02:00