This new version compares the final waypoint with the target's
obtruction shape and uses reachability checks to know if we will be in
range or not.
Differential Revision: https://code.wildfiregames.com/D1983
This was SVN commit r22430.
This is a partial revert of 4fda917f46, which skipped the
"MoveToTargetAttackingRange" in APPROACHING. I (incorrectly) assumed
that the original order was still perfectly fine, but in fact the
attacker's max range may have changed as that depends on the relative
elevation between attack and target - and so the original order might
never get us in range!
This was introduced originally in 8c74df2acd.
Add a comment to clarify this.
Further, this makes sure UnitMotion still is aware that it has a target
even if it is in range from the beginning, as that could lead
to stuckiness (and did when chasing sometimes). This was done in D1984
anyways.
Fixes#5478.
Differential Revision: https://code.wildfiregames.com/D2035
This was SVN commit r22429.
This makes it possible to merge the TryGoingStraightTo family of
function.
This means it is now called even for point goals and decreases the
likelihood of stuck units (as we take over from the long-range
pathfinder which brings units to navcell centers, which may not be in
the goal).
Differential Revision: https://code.wildfiregames.com/D1982
This was SVN commit r22426.
D981/c219ee54b2 changed IsInXChecks to use edge-to-edge distance instead
of centre-to-edge, which broke UnitMotion's min-range movement, which
assumed distance to the center.
Since units are represented as squares, the diagonal point may be closer
to the target than the "real" clearance by a factor √2, so the delta
between minimum range and maximum range should be at least `(√2 - 1) *
clearance` to be safe in all situations (this is generally not a problem
for regular units which have a clearance of 0.8, but could be one for
catapults or elephants).
Differential Revision: https://code.wildfiregames.com/D1969
This was SVN commit r22422.
Previously, unitMotion had no code that checked particularly if the
target was still in the world.
When the target moved out of the world, unitMotion would follow the path
to its last known position, then send a "MoveSucceeded" message once
there.
Following 98f609df1d, this message was no longer sent. Thus unit would
follow their path to its last waypoint and stay there, unable to carry
on or finish the order. UnitMotion now explcitly sends a "MoveFailed"
message.
This still changes behaviour from A23, requiring further revisions to
UnitAI (see D1992 for one such case).
Minohaka tested an earlier version of this change (which incorporated
D1992) and accepted it.
Differential Revision: https://code.wildfiregames.com/D1979
This was SVN commit r22415.
Due to an issue in 4a15cc3b9f, animals incorrectly tried going towards
the roaming point instead of away from it.
With that fixed, MovementUpdate still did not trigger as the min and max
range were the same.
Use -1 as max range (= infinite) instead since we want to move
arbitrarily away.
Having an infinite max range was broken in c219ee54b2, this
re-implements that.
Further, other calls with equal min and max range have been changed
likewise.
This does not entirely fix whales, which run in other problems because
of their large roaming range.
Differential Revision: https://code.wildfiregames.com/D1980
This was SVN commit r22413.
UnitMotion will now send hints that a move may be completed if:
- it is a formation member and the formation member has stopped moving.
- it isn't and it is in range from its target.
Differential Revision: https://code.wildfiregames.com/D1899
This was SVN commit r22366.
Units in 0 A.D. exhibited a "gliding" behaviour at the end of a
movement, e.g. they switched to the Idle animation and still moved a
little bit.
The reason for this behaviour is that entities check if they reached
their destination after moving. Other components (unitAI mostly) will
then possibly change animations and such, resulting in a movement over
the turn while the entity is possibly now in another animation state
than "move".
Instead, what should be done is checking if the entity has arrived
before moving, so if UnitAI calls StopMoving, then entity won't move at
all on the same turn, and the gliding effect vanishes.
The STATE_STOPPING state is made un-necessary by this change, since this
off-by-one mistake was the reason for its existence. It can be removed
(see downstream).
Differential Revision: https://code.wildfiregames.com/D1898
This was SVN commit r22365.
Move() is generally 4 parts:
- Moving
- Updating our state
- Handling obstructed moves
- Checking if we are at destination.
These can all be put in their own functions, clarifying logic and making
it harder to make mistakes.
Differential Revision: https://code.wildfiregames.com/D1897
This was SVN commit r22364.
In preparation for D1897, this logic isn't related to the "movement"
part of Move(), and can be called earlier.
Differential Revision: https://code.wildfiregames.com/D1896
This was SVN commit r22363.
The obstruction manager keeps a flag of moving units. This should be
updated in the same location as the entity's current speed, since they
have similar properties.
Differential Revision: https://code.wildfiregames.com/D1895
This was SVN commit r22362.
VisualActor refers to unit motion's `m_CurSpeed` to know what animation
to display in the special walk mode. However that variable was
inconsistently updated. By making sure we reach that codepath in one
place, we remove opportunities for issues.
This should fix most existing instances of moonwalking, another upstream
diff will fix the rest.
Differential Revision: https://code.wildfiregames.com/D1894
This was SVN commit r22361.
FacePointAfterMoving intends for the unit to face the destination once a
move is done. Since 4fda917f46, stopping is the responsibility of UnitAI
(through a call to StopMoving()). Thus we should move that code in that
function, as this ensures we don't forget to do it and removes
duplications.
Differential Revision: https://code.wildfiregames.com/D1889
This was SVN commit r22355.
The variable is not necessary since having a target is equivalent,
removing it thus reduces redundant state.
Differential Revision: https://code.wildfiregames.com/D1888
This was SVN commit r22354.
These variables together held the state for the target of UnitMotion, as
set by the MoveTo[X] family of functions.
Wrapping them in a struct reduces the chances that one will accidentally
forget to reset part of the state and makes it explicit in-code that
these are grouped together.
Calling StopMoving() resets this target, which wasn't before and left
the component in an incoherent state.
Differential Revision: https://code.wildfiregames.com/D1887
This was SVN commit r22352.
UnitAI is now solely in charge of moving and stopping, making UnitMotion
behaviour easier to predict, which will ultimately help with unitAI
development. It might temporarily make units more resilient than before
however.
UnitMotion also tells UnitAI that it's arrived with "MoveCompleted"
messages, but these actually could be wrong - unitAI could decide that
we didn't want to stop after all - so change the name for something less
misleading.
Differential Revision: https://code.wildfiregames.com/D1886
This was SVN commit r22351.
"MoveStarted" messages were sent by UnitMotion when the unit started
moving (sort of) or failed to do so. This was used by formations and
guarding but was not really necessary as this can be done in "enter" or
in a timer.
Differential Revision: https://code.wildfiregames.com/D1885
This was SVN commit r22350.
These functions were placed in UnitMotion, which had nothing to do with
range checks and made them available only to moving entities for no
particular reason.
This patch also adds support for square-square range checks and
shape-shape distance checks.
Modified from a patch by bb on top of work from wraitii.
Differential Revision: https://code.wildfiregames.com/D981
This was SVN commit r22345.
0c20afdfda had two issues:
- some of the unitAI code did not return true when switching states in
the "enter" phase
- missed a return false in unitMotion.cpp
This fixes the issue noticed by @bb.
Differential Revision: https://code.wildfiregames.com/D1947
This was SVN commit r22339.
This also moves the actual "moving" code to states instead of orders,
making states more self-contained and removing the change of errors when
cleaning up a state.
Differential Revision: https://code.wildfiregames.com/D1865
This was SVN commit r22313.
Having Async in the name was not really informative and made it awkward
to reuse for non-async code.
Reviewed By: Kuba386
Differential Revision: https://code.wildfiregames.com/D1854
This was SVN commit r22305.
It improves readability of the code when a function that seems like it
does a simple check actually only does a simple check and the caller is
the one changing state.
Differential Revision: https://code.wildfiregames.com/D1884
This was SVN commit r22299.
Following 809f297707, this decouples the hierarchical pathfinder and the
long pathfinder. The long pathfinder was the class owning the
hierarchical pathfinder, which didn't particularly make sense and
resulted in some interface awkwardness.
At the moment, the long pathfinder still needs to hierarchical
pathfinder to compute paths (to make sure they are reachable).
Differential Revision: https://code.wildfiregames.com/D1867
This was SVN commit r22278.
This:
- cleans up a code duplication and clarifies the intent.
- reorders things around for clarity
- improves variable names.
Commented By: elexis
Differential Revision: https://code.wildfiregames.com/D1840
This was SVN commit r22277.
The vertex pathfinder was implemented directly in CCmpPathfinder,
instead of being a separate helper like the hierarchical pathfinder or
the long pathfinder.
This moves it to its own helper VertexPathfinder, which gets us ready
for D14 and pathfinder threading. Some struct definitions need to be
moved around.
Differential Revision: https://code.wildfiregames.com/D1855
This was SVN commit r22253.
Because of the limited precision of our fixed-point numbers, the
timeLeft calculation could sometimes return results above the actual
time left, resulting in units moving a few fixed::epsilons farther than
they should be, which makes them switch to the running animation. This
was rather unstable however, so there was a constant 'flickering'
between walking and running.
If we divide last instead of first in the operation, the errors get
gobbled up by the division and we no longer have this issue.
Reported by: wowgetoffyourcellphone
Differential Revision: https://code.wildfiregames.com/D1856
This was SVN commit r22249.
FindPassableRegions intends to return all passable regions in a chunk,
but did not as it used the number of regions in that chunk. In fact,
regions can have individual IDs higher than the number of regions (as
shown by the test), therefore FindPassableRegions might miss some.
This only affected the JPS pathfinder, when starting on an impassable
cell, which called FindNearestPassableNavcell with then possibly
returned a sub-optimal navcell. This is a limited impact but upcoming
patches will rely on that function more.
Fixed using a vector to store IDs, which also makes for-range loops
usable.
Differential Revision: https://code.wildfiregames.com/D1832
This was SVN commit r22218.
Includes some style normalization.
Refs 60cfd2d16c.
Patch By: Stan
Reviewed By: vladislavbelov, Itms
Differential Revision: https://code.wildfiregames.com/D1845
This was SVN commit r22216.
This changes running speed into a running multiplier (of walk speed).
The advantage is that it simplifies code since you can setup a default
run multiplier at the template level and it'll work for all subsequent
templates, and technologies cannot forget to change it. It makes
specialised unit templates easier to maintain, too.
Formations have a 100 run multiplier which effectively sets their
maximal walking speed at 100
Reviewed By: bb, O2 JS Simulation
Differential Revision: https://code.wildfiregames.com/D438
This was SVN commit r22197.
This generally makes this code cleaner and easier to maintain, while
also being more modern and fully tested.
Looked at by: leper
Differential Revision: https://code.wildfiregames.com/D8
This was SVN commit r22029.
Use an enum to indicate the type of engine shutdown instead of three
bools.
State in the comments that the program is restarted within the same
process.
In preparation of introducing an IsQuitRequested function (which shall
not be named is_quit_requested as stressed by Vladislav):
Rename kill_mainloop to QuitEngine, restart_mainloop to RestartEngine,
restart_mainloop_in_atlas to StartAtlas to not break consistency.
Don't rename RestartInAtlas in JS just now.
Group declarations at the top of main.cpp.
This was SVN commit r21817.
Clean the code by removing the animal hardcoding in the Foundation
component and adding a flag DeleteUponConstruction to the Obstruction
component.
Have locked gates and upgraded entities equally delete entities when
transforming.
Add a workaround for trees inside walls on random maps.
Reviewed by: temple
Differential Revision: https://code.wildfiregames.com/D1415
Refs #4268
This was SVN commit r21624.
so that units don't try to gather forever while not being able to reach
it and
so that the sheep corpse can't be selected and seen anymore.
Differential Revision: https://code.wildfiregames.com/D21Fixes#4268
Based On Patch By: wraitii
Reviewed By: temple
Previously Reviewed By: Itms
This was SVN commit r21597.
This currently disabled by variants, and will be reenabled when sounds
have been chosen.
Reviewed by: Vladislav,
Comments by: Itms, elexis
Differential Revision: https://code.wildfiregames.com/1257
This was SVN commit r21359.
Add a button to toggle between normal and "diplomacy" colors.
Reviewed by: elexis
Comments by: leper, ffffffff, Stan
Fixes: #4747
This was SVN commit r21107.
Since the component initialization order is the alphabetic filename
order and
since the RangeOverlayRenderer < Player < Selectable, the new file in
that commit introduced the issue.
Similar to 36b315ca32, refs #4632.
UpdateColor function taken from temple's D754
This was SVN commit r20963.
Use the distance to the nearest point on the dropsite's obstruction
rather than the center of the dropsite.
Differential Revision: https://code.wildfiregames.com/D1160
Reviewed by: wraitii
Comments by: bb, elexis
Fixes: #4592
This was SVN commit r20950.
Removes the serialization of JSON files, shrinking savegame files and
rejoin states sent across the network, refs #3834, #4239, #3909,
f24523dc8f.
Removes the AI C++ code to read JSON files from e33d4a52e9 since the AI
can now use the globalscript.
Allows the AI to read Aura templates and removal of GUIInterface code to
improve performance.
Serialization of the JSON objects in other simulation components was
removed in 9c0e37f2c0 / D1109, a6f14f5631 / D1130.
Serialization removal planned by sanderd17
AI part proofread by mimo
Simulation part proofread by bb
Discussed with Itms on irc
Differential Revision: https://code.wildfiregames.com/D1108
This was SVN commit r20737.
This will enable us in the future to have technologies that change
projectiles.
This is also somewhat of a refactoring.
Patch By: Mate-86
Reviewed By: wraitii
Trac Tickets: #1909
Differential Revision: https://code.wildfiregames.com/D945
This was SVN commit r20676.
Original patch from Sandarac, reworked by elexis and finalized by mimo
Discussed with elexis
Differential Revision: https://code.wildfiregames.com/D1159
This was SVN commit r20671.
By adding a custom function in Grid, the code gets vectorised on both
gcc and clang, resulting in much faster code and faster update times,
sometimes substantially (on giant maps or when few chunks must be
updated).
Reviewed By: mimo
Differential Revision: https://code.wildfiregames.com/D73
This was SVN commit r20630.
This also allows non-selectable entities like building previews to cast
range visualizations.
Patch By: Sandarac
Differential Revision: https://code.wildfiregames.com/D555
Refs #3915, #4349, ee5bb1fd61 / D238.
Comments By: Vladislav
This was SVN commit r20622.
This abstraction allows calling that code again from other simulation
components, like the RangeOverlayRenderer in D555.
Differential Revision: https://code.wildfiregames.com/D1139
Refs #3915, D555
Comments By: Vladislav, echotangoecho
This was SVN commit r20621.
Removes the workaround copy of the resources JSON each turn in
GetSimulationState.
Refs #3934, #4868
Differential Revision: https://code.wildfiregames.com/D1119
Reviewed By: mimo
This was SVN commit r20600.
Optimizing the memory footprint requires likely different functions
following the new SM versions and a lot of testing,
so having these doesn't lower the effort to achieve optimization while
confusing the reader.
Differential Revision: https://code.wildfiregames.com/D1073
Reviewed By: Yves
This was SVN commit r20572.
Completes an existing TODO. This is a large speedup in some cases for
those test functions, particularly the "sunken ship lag" issue.
Patch By: temple
Reviewed By: elexis, mimo, bb, wraitii
Accepted By: bb, wraitii
Trac Tickets: #4713
Differential Revision: https://code.wildfiregames.com/D850
This was SVN commit r20438.
This removes FindAllPlaceableTemplates, replaces the few uses of it by
FindAllTemplates,
and makes that ignore all templates starting with special/ in addition
to those starting
with template_.
Now modders can use entirely different template organization schemes
(more folders, different
folders, etc) without having to edit a file that was never well
documented.
In conjunction with a few of the template moving patches preceding this
rubble/ and other/catafalque
are now placeable. The former now does not decay anymore and users that
want that should use the
decay| filter, the latter will be taken care of in #4762.
Return to making FindAllTemplates return all placeable templates again
(switch to unplaceable filter).
To reiterate the main point: Only templates starting with special/ or
template_ will not show up as
placeable in Atlas (or show up to code querying for all (placeable)
templates. If you want to add more
of those use one of these naming schemes (and possibly subfolders in
special/).
Reviewed By: fatherbushido
Differential Revision: https://code.wildfiregames.com/D935
This was SVN commit r20246.
This allows for amphibious units (swimming animals, tanks, etc).
The specified depth is the distance from the water surface where the
unit will start to float. Should most likely be a little less than
the height of the actor.
Reviewed By: fatherbushido
Differential Revision: https://code.wildfiregames.com/D842
This was SVN commit r20196.
Thus throw errors if a simulation test tries to alternate the template
instead of silently passing tests with wrong values (as happened tests
fixed by ff90bb8490).
Remove the two clones from simulation components that were only relevant
for that test.
Add test to test the test.
Differential Revision: https://code.wildfiregames.com/D871Fixes#4759
setup.js and setup_test.js Reviewed By: leper
This was SVN commit r20134.
Note that this does not fix the warnings in AtlasObjectXML, someone
interested
in fixing those should check whether using ICU would be a nicer
solution.
Reviewed By: echotangoecho
Differential Revision: https://code.wildfiregames.com/D740
This was SVN commit r20095.
Remove unused SGUIBaseSettings and GUI comment.
Fix indentation of a macro, refs D794.
Differential Revision: https://code.wildfiregames.com/D838
Review by: leper.
Itms came up with the same patch for the SpiderMonkey 45 update
independently.
This was SVN commit r20062.
This fixes an assertion failure in ScriptEngine that can occur when
closing
the game while in the loading screen.
Reviewed By: vladislavbelov, leper
Differential Revision: https://code.wildfiregames.com/D684
This was SVN commit r20035.
Properly support capturing in the RallyPointRenderer component, even
though RallyPoints are typically deleted upon capturing.
Differential Revision: https://code.wildfiregames.com/D776
Refs D687
Reviewed By: Stan
This was SVN commit r19976.
Remove all hacks while keeping most optimizations in memory management.
This fixes incomplete grid updates that could cause OOS on rejoin, fixes
#4596.
Tested by: elexis, ffffffff
Reviewers: leper, wraitii
Differential Revision: https://code.wildfiregames.com/D675
This was SVN commit r19916.
Let the JS GUI take care of hiding the GUI and silhouettes and remove
the according hardcoding in the engine following 89aef0b6eb.
Thereby fix some bugs (like not having hid the GUI if there was a
message box or different page shown while playing a path) and
fix these two hotkeys broken by 5d49e6c456.
Differential Revision: https://code.wildfiregames.com/D631Fixes#4633
Reviewed By: Vladislav
This was SVN commit r19797.
6aeb5c64de forgot to add a LosIsOffWorld check in ExploreTerritories
(aka UpdateTerritoriesLos) and thus marked tiles outside of the world as
explored.
f5e60157bf transformed the bug into a non-simulation desynchronization,
causing rejoined players to see a different score,
as they excluded off-world tiles when filling the cache in
ResetDerivedData upon rejoin.
4a0673e44e transformed the bug into an actual simulation OOS by
serializing that map exploration percentage based on that cache.
Also tiles at the map border in square maps are not rendered as
expected, so this commit hides refs #4267.
Differential Revision: https://code.wildfiregames.com/D630Fixes#4598
Proofread by: Itms
Tested By: Imarok
This was SVN commit r19790.
This allows automated testing of AIs without any GUI or sound (similar
to non-visual replays).
Differential Revision: https://code.wildfiregames.com/D379
This was SVN commit r19645.
Also show the GUI after the cinematics stopped playing in Atlas.
Differential Revision: https://code.wildfiregames.com/D411
This was SVN commit r19533.
Doesn't provide a way to add/delete nodes of paths yet.
Differential Revision: https://code.wildfiregames.com/D348
Patch By: Vladislav
This was SVN commit r19427.
The remaining graphics code should be moved or removed.
Differential Revision: https://code.wildfiregames.com/D324
Patch By: Vladislav
This was SVN commit r19414.
Expose the FindAllPlaceableTemplates function to the simulation to
remove the hardcoded directory checks as suggested by fatherbushido.
Add the new special template directory from D176 / cd6c31e76e.
Move the code from XML to JS.
Differential Revision: https://code.wildfiregames.com/D277
Reviewed By: Vladislav
This was SVN commit r19399.
The terrain grid can be renewed without proper deallocation, which
happens at least at the start of a game when MT_TerrainChanged is sent.
Reviewers: wraitii, Itms
Differential Revision: https://code.wildfiregames.com/D247
This was SVN commit r19388.
Move graphics code to smaller helper functions UpdateSessionVisibility,
UpdateSilhouettesVisibility and DrawPaths.
Remove the hotkey TODO code which should be implemented differently.
Mark voids as const.
Differential Revision: https://code.wildfiregames.com/D271
Patch By: Vladislav
This was SVN commit r19375.
Add a test for the GetNonGaiaEntities function of the RangeManager
broken by that commit that was fixed by f3e4e619bc.
Don't call one script function from another, but directly call into the
Selection helper like the others.
Don't make a loop around that RangeManager function for all players but
call the function once for all players.
The pointless virtual keywords were removed by 8827db201a.
Differential Revision: https://code.wildfiregames.com/D166
Reviewed By: leper
This was SVN commit r19344.
Use std::set's logarithmic find instead of a linear loop.
Early return and some whitespace for getParabolicRangeForm.
Reviewed By: elexis
Differential Revision: https://code.wildfiregames.com/D260
This was SVN commit r19342.
Summary: Following discussions in https://code.wildfiregames.com/D60 the
use of u8 to store each boolean has been cleanup up. Now only one u8 is
used, thus reducing the size of the struct.
Reviewed By: leper
Differential Revision: https://code.wildfiregames.com/D101
refs #3834
This was SVN commit r19242.
Template-ize CallFunctionVoid.
Changes CallFunction parameter order to make template parameter
deduction with
variadic parameters work nicely.
Reviewed By: Itms, wraitii, Yves
Differential Revision: https://code.wildfiregames.com/D77
This was SVN commit r19183.
Summary:
To fix tickets like #3335, having a shared vision at the entity level is
needed. This patch implements that in CCmpRangeManager, interfaced with
a new JS component VisionSharing to manage the shared entities visions.
As an example of use case (in addition to garrisoning in allied
structure without the sharedLos tech), we can think of bribing enemy
units: there is a part about it in the patch, but this one is very wip
and not intended to be committed but rather for test purposes of the
feature.
So when garrisoning a unit in another player's building, the vision of
the garrisonHolder is shared (part intended for review). In addition,
for tests of the feature, when clicking on the new bribe icon in the
diplomacy window, a random unit of the chosen player is bribed and share
its vision during 15 s.
Test Plan: Garrison a unit in an allied structure without the sharedLos
tech, or test the wip bribe feature from the diplomacy window.
Reviewers: Itms
Reviewed By: Itms
Subscribers: Stan, leper, O11 Templates, wraitii, elexis, fatherbushido,
Itms, Vulcan, O1 C++ Simulation
Differential Revision: https://code.wildfiregames.com/D60
This was SVN commit r19175.
Remove the "Net" prefix from the non-networked classes.
Use variadic macros and mark the client turnmanager as NONCOPYABLE.
Differential Revision: D16
Reviewed By: leper
This was SVN commit r19165.
since they are only useful for AI debugging and counterproductive in
multiplayer mode, refs #3551.
Instead, enable AI developers to exit the game from a new AI API
command,
allowing to batch simulate matches. Refs #2755.
Differential Revision: D65
Reviewed By: leper
Consulted: mimo
This was SVN commit r19155.
Templates should be the same for all players, techs can be used to
provide
different lists of templates if that is needed.
Reviewed By: Itms
Differential Revision: https://code.wildfiregames.com/D25
This was SVN commit r19142.
Summary: Refs #4419.
Test Plan: When running tests, make sure nothing is created.
Reviewers: leper
Reviewed By: leper
Subscribers: leper, Vulcan
Differential Revision: https://code.wildfiregames.com/D1
This was SVN commit r19062.
The previous code worked in case new mutual allies were added, but, when
some were removed, it was suboptimal in terms of performance and it made
some units visible in the FoW.
Fixes#4266
This was SVN commit r18994.
Serialize the mapsize in the pathfinder and the reveal shoreline flag in
the range manager.
Reload the rangemanager data after other components have been
deserialized.
Use the SerializeCommon pattern in the pathfinder to avoid code
duplication.
Move the shoreline logic from the Vision component to the range manager.
Remove unused interface mocks from the rangemanager test following
b05879e151.
This was SVN commit r18879.
The internal data of the component, which is serialized, should not
depend on the presence of a visual Unit.
Also remove a misleading comment about a simple initialization code.
Refs #4270
This was SVN commit r18824.
* Rewrite the Update function, add more details and information in
comments, and properly serialize everything that function needs.
* Fix the broken deserialization code by using a sane helper function.
* Fix the `SelectMovementAnimation` function.
Fixes#4270.
This was SVN commit r18823.
If it is needed at some point, this map should be handled and updated
the same way as the passability map and territory map are.
Refs #4191
This was SVN commit r18760.
This should have a noticeable impact on performance (in the good way!)
Thanks mimo for noticing something was off with the planning system!
This was SVN commit r17866.
Using references matches the C++ coding style better and should improve
performance a bit in theory. It avoids 2 copies of T in case of the
functions registered with RegisterFunction (mainy used in the GUI). It
should also avoid one or two copies in case of
DEFINE_INTERFACE_METHOD_X, which is used in the simulation, but I
haven't bothered to count it there exactly.
It is now predefined which types have to be passed by const reference
and which are passed by value. Note that references can't be used as
out-parameters (to return multiple values to JS). This hasn't worked
before either and probably never will.
This was SVN commit r17696.
While there don't check for null when delete-ing since that is a no-op,
use the proper type instead of auto, always swap when removing a single
element (instead of comparing .size() to 1), and clean up includes.
foobar
This was SVN commit r17641.
Rename PickFriendlyEntitiesInRect to PickPlayerEntitiesInRect.
Rename PickSimilarFriendlyEntities to PickSimilarPlayerEntities.
Add comment.
Refs #3215.
This was SVN commit r17622.
Based on existing code that was still around from an old(not working)
implementation.
Supports basic control from trigger scirpts (queue, start and stop
camera paths) and works in multiplayer.
This was SVN commit r17594.
Also stop trying too hard when we are close to the destination and our
only order is to move there.
This should result in a slight optimization of the behavior, as well as
slightly more sanity overall.
This was SVN commit r17226.
Make sure we do not treat as circles entities that we really should
treat as squares (such as trees). This fixes an issue reported by Stan.
Make sure we never forget about our destination if we are blocked by
unit obstructions. This makes sure that units in a group but not in
formation will not be blocked by the other units, and probably makes the
general behavior more sane. Helps following [17166]
Refs #3505, #3471, #3376
This was SVN commit r17191.
Also fix a potential issue I noticed in some cases, though that
particular fix implies scrapping waypoints, so if units seem to get
lost, please report so and I'll revert those changes.
I can't find any way to get units stuck with this patch.
This was SVN commit r17163.
Change the way the long-range pathfinder rasterisation works slightly so
that we have a better compatibility with the short-range pathfinder.
Should fix the "stuck units" issues, though I am not sure so I am not
marking them as fixed so far. Refs #3471, #3505, and possibly #3292.
Caveat: I am now using clearance of 0.8 for "default" class, which might
have side-effects: please report anything weird.
Also fix leftover style issues.
This was SVN commit r17161.
This resulted in ALL "units inside obstructions" issues.
Thanks to elexis for the testing.
Fixes#3532, #3450.
Refs #3538 (still OOSes), #3410 (unitmotion remains buggy for
formations, but this is only aesthethic.)
Probably affects #3471 and #3505, but those are not fixed.
This was SVN commit r17152.
Please note that this commit makes #3410 really visible. Formation
members might pass through buildings.
Refs #3410, #3337.
This was SVN commit r17028.
This typo didn't have any influence since dirty unit shapes are usually
discarded during the rasterization (they don't block long-range
pathfinding). Switching the order of a check in RasterizeHelper prevents
performance problems to a really small extent.
This was SVN commit r17014.
Now that units flee it's necessary to fix the unit chasing: this commit
reintroduces some code from 298115f4c5 that disappeared with the
committing of the new pathfinder. Refs #1537.
Also includes some style improvements to the UnitMotion code.
This was SVN commit r17013.
With this change, units will not check their movement against all
obstructions when moving: terrain and static obstructions are assumed to
be handled by the long-range pathfinder.
However, when static obstructions are changed, the paths have to be
invalidated. In order to minimize the performance impact, units will
check for obstructions when they move after a passability change. If
they collide with something, they will recompute a path that will take
into account the new passability map.
Also includes some code cleanup. This patch should not change
performance a lot: the lower number of checks should give a small
performance improvement while using the message broadcasting system
should hurt it a bit.
Fixes#3376, #3337, #1914.
This was SVN commit r16998.
Those shapes need to be taken into account when computing a short path,
but they need to be discarded when checking movements or they will
create long/short inconsistencies.
This was SVN commit r16981.
As a general rule, pathfinding-blocking shapes should not be tested
against for movements. Only the passability grid should be considered.
As the algorithm which tests paths on this grid was fixed in 1c9ea56800,
it can be safely used.
Fixes#3376.
This was SVN commit r16971.
The list of modified entities was thrown out on global visibility
updates (those happen on rejoin) but only in-world entities were
actually updated then, causing problems with garrisoning.
Now the list of modified entities can happen to be large, so replace the
hacky check for infinite loops by a real check.
This was SVN commit r16962.
However, using the terrain-only grid reveals one discrepancy between the
short pathfinder (which uses unit radii) and the long one (which uses
unit clearances). So I implemented the change proposed by sanderd17 in
#3294, which is removing unit radius and using only the pathfinder
clearance. Refs #3294
Now some tweaking has to be done in the templates, so that units get a
passability class suited to their apparent size. In the meantime the
unit motion is quite bugged.
This was SVN commit r16867.
Currently only preview entities are affected, and they are not
serialized, but it could break a lot of things in the future.
This was SVN commit r16858.
The game has to deal with situations such as: the visibility of an
entity changes, a mirage is created for it -> the mirage visibility is
updated -> the entity visibility is updated back.
All of this process now happens in the same turn, and all updates are
guaranteed to be performed. This fixes a source of serialization errors
and rejoin OOSes.
Fixes#3107
This was SVN commit r16857.
Also make the GetPassabilityClasses functions use references instead of
wild allocations. Use a reference when passing pass classes to the AI
worker.
This was SVN commit r16833.
This subdivision is faster but less precise, so range queries get more
entities and are a bit slower (up to 1ms approx.), but the overall gain
on a simulation update is always positive and can reach 10ms per frame.
For now, this new subdivision is only used by the range manager,
integrating it in the obstruction manager might be sensible.
Refs #2430
This was SVN commit r16540.
This improves performance quite a lot because it avoids a huge number of
calls from C++ to JS. Check the ticket for performance measurements.
Refs #2913
This was SVN commit r16337.
This fixes the problem where passing a non-ASCII string to
debug_printf(L"%hs", s) caused vswprintf_s to fail on Linux (because it
doesn't know what encoding the char* is meant to have). Now debug
messages will remain as UTF-8 until they reach the OS.
Fixes#3021.
This was SVN commit r16332.
It is necessary to rely on shared los masks, else some visibility
updates will be missing.
Refs #2913, see this ticket for a performance graph.
This was SVN commit r16328.
Fixes the animals hidden in the FoW, and adds the missing status bars
for mirages.
Also small cleanup of the code.
Refs #2913
This was SVN commit r16281.
This upgrade also introduces exact stack rooting (see to the wiki:
JSRootingGuide) and fixes problems with moving GC. This allows us to
enable generational garbage collection (GGC).
Measurements a few months ago have shown a performance improvement of a
non-visual replay of around 13.5%. This probably varies quite a bit, but
it should be somewhere between 5-20%. Memory usage has also been
improved. Check the forum thread for details.
Thanks to everyone from the team who helped with this directly or
indirectly (review, finding and fixing issues, the required C++11
upgrade, the new autobuilder etc.)! Also thanks to the SpiderMonkey
developers who helped on the #jsapi channel or elsewhere!
Fixes#2462, #2415, #2428, #2684, #1374
Refs #2973, #2669
This was SVN commit r16214.
Everything is char* now, so we don't need to mess around with different
string types.
Done with:
ag -ls 'LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR)' source | xargs perl
-pi -e'1 while
s/(LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR).*)%[hl]s/$1%s/g'
This was SVN commit r16187.
Done with:
ag -ls 'LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR)' source | xargs sed
-i 's/LOG\(MESSAGE\|MESSAGERENDER\|WARNING\|ERROR\)(L/LOG\1(/g'
This was SVN commit r16183.
First, do a ray intersection test with the bounding-sphere for all
entities on the map and then check the more detailed selection shape for
the remaining candidates. Do checks that require component lookups after
the ray intersection tests because these are relatively expensive.
The old method for figuring out which entities are below the mouse
cursor was incorrect because it does a 2D check to filter out the first
candidates which can lead to incorrect results with lower camera angles
and high buildings or buildings with a large footprint. Such problems
were avoided with quite a large radius for this 2D test and resulted in
a large number of candiate entities after this first test (200-500).
Also rename PickEntitiesAtPoint to PickEntityAtPoint and make it return
only one (the closest) match.
I've tested performance with the tracelogger by starting a map and then
moving the mouse in circles for one minute. The results were relatively
stable. I've compared the total time percentage of input.js:836, which
spends nearly all of the time in PickEntityAtPoint.
Ardennes Forest - Normal size: Original: 41.46% Patched: 31.6%
Ardennes Forest - Giant size: Original: 40.59% Patched: 51.55%
As we see, it's faster on normal map sizes but slower on giant maps with
a lot of entities.
This approach can be further improved with some kind of spatial
subdivision for the culling (like an octree), which would help the unit
renderer too. This way it should be possible to make it faster (and
still correct) on all map sizes and with a large total numbers of
entities.
This was SVN commit r16098.
On my ESR31 branch, I've made two measurements with different replays
(both around 15000 turns).
In the first, I got around 3% performance improvement and in the second
about 7.5%. It mainly depends on how often aura changes related to the
female citizen aura happen.
This was SVN commit r16055.
