Because the page stack is a vector, if during PushPage, a new page is
pushed, the vector may re-allocate. This 'pulls the rug out' from
underneath the code stack that originally pushed, which then crashes.
To fix this, use a deque, since push/pop won't invalidate references.
Reported by: Imarok.
Based on a patch by: Imarok (tests are his)
Differential Revision: https://code.wildfiregames.com/D4037
This was SVN commit r25616.
Variants can now have limited or no diversity. This can occasionally
speed rendering slightly (5-10% FPS increase was reported on Combat Demo
Huge, which is very variant-heavy).
Reported by: bb
Based on a patch by: bb
Fixes#5831
Differential Revision: https://code.wildfiregames.com/D3035
This was SVN commit r25613.
Bit operations can be done via std::bitset. The current implementation
is too specific to have it just for a future. Refs 479f59e386,
5ca66fc757.
This was SVN commit r25569.
For modern standards base64 costs a bit less space and might be
implemented on demand. base32 usage was introduced in e0dfbe719d and
removed in 93cffe9deb.
Refs a34b759720, 317f98a6c0.
This was SVN commit r25568.
We already have a single place to access arguments, so we don't need
a platform dependent place for that.
Refs 3f58951808, 8c1bd05ab9.
This was SVN commit r25567.
Following D3108 / 876f6d5e50, sounds are attenuated by their actual
distance. However, as noted by players on A24, the dropoff is stark and,
when zoomed out, can easily result in not hearing things that are
happening in the middle of the screen.
The new default settings almost double the max-range, change the minimum
range to have greater dynamic range, and reduce the stereo depth
slightly to better match the default camera FOV.
These are stored per sound-grounp, possibly allowing future tweaks on a
per-soundgroup basis, and are configurable.
Tested by: Imarok
Discussed over mail with: Samulis, Porru
Differential Revision: https://code.wildfiregames.com/D3612
This was SVN commit r25547.
Parses mod.json data not only in temporary JS values, but in a proper
C++ struct.
This will ultimately make it more convenient to pass more than just the
version to JS in D3968, and it enforces the schema a bit more.
Differential Revision: https://code.wildfiregames.com/D3988
This was SVN commit r25546.
Supersedes D1674 and D3520.
This still generates some false positives (particularly the
singular-plural check), but in general it will suffice to better check
bundled languages.
Refs #4250
Differential Revision: https://code.wildfiregames.com/D3926
This was SVN commit r25538.
As noted by Angen in a26535d023, refs 3bcf360107
When starting with incompatible mods, the non-incompatible mods were no
longer shown. This fixes that.
I think the code looks cleaner from this new control flow that
explicitly only mounts "mod".
Tested by: Angen
Differential Revision: https://code.wildfiregames.com/D3996
This was SVN commit r25510.
Reduces the 'clumpiness' of units, particularly when ordered to move to
a single point. The minimum pushing force was increased to compensate.
Differential Revision: https://code.wildfiregames.com/D3978
This was SVN commit r25479.
I didn't quite understand what was supposed to happen here. This makes
things work as before.
While at it, fix a conversion bug & do some more tweaks.
Differential Revision: https://code.wildfiregames.com/D3990
This was SVN commit r25475.
- Non-visual replays now automatically try to load the replay mods. This
removes the annoyance that -mod=public usually had to be passed.
- MountMods is no longer called in InitVfs but later, making it possible
to load the game in one pass & simplifying things considerably.
- Explicitly ignore duplicates when loading mods
- Interface cleanup: failed mods and incompatible mods were redundant,
only incompatible mods is kept.
- Interface cleanup: `AreModsCompatible`becomes
`CheckForIncompatibleMods`, which becomes a private pure function,
simplifying the control flow somewhat.
- Interface cleanup: `CheckAndEnableMods` is just `EnableMods`, which
explicitly updates loaded & incompatible mods.
- `CacheEnabledModVersions` becomes private and is called on behalf of
the user, removing the need to be careful about updating that (fixes my
concern at 44ec2e324e)
Overall, the logic around mounting & enabled mods should be easier to
understand.
Differential Revision: https://code.wildfiregames.com/D3982
This was SVN commit r25474.
`Mod` currently modifies 4 global arrays, including some that are
'public'.
The logic flow is easier to understand if this becomes a real class with
const-correctness and public/private methods.
Differential Revision: https://code.wildfiregames.com/D3981
This was SVN commit r25469.
The purpose of our client-side hashing for lobby game passwords is to
prevent malicious hosts from getting valuable passwords from clients
(e.g. accidentally typing their lobby password instead of the game, or
even their email password, etc).
However, the hashing was deterministic (and rather simple), making it
possible to compute rainbow tables and recover user passwords anyways.
By adding more variation, including some that cannot so easily be
controlled by the host (the client name), this becomes impractical. The
password hashing function used is rather fast, but given the base low
probability of mistypes, this seems fine.
Differential Revision: https://code.wildfiregames.com/D3459
This was SVN commit r25459.
This makes incremental recompilation faster when changing static
interned strings.
Differential Revision: https://code.wildfiregames.com/D3975
This was SVN commit r25457.
Instead of using platform-specific sockets, use enet_socket* functions
(which ends up doing the same).
Clean up some confusing APIs, removing the distinction between finding
the public IP for the host/join.
Fix endianness support & use simpler code.
Refs D364 / 61261d14fc (and some subsequent fixing diffs).
Differential Revision: https://code.wildfiregames.com/D3970
This was SVN commit r25453.
This allows joining a lobby game hosted on the same network (behind the
same NAT gateway).
This is relatively primitive to keep things simple: if the server and
the client have the same public IP, it is assumed that they are on the
same network and the client instead requests the local IP.
Differential Revision: https://code.wildfiregames.com/D3944
This was SVN commit r25448.
ScriptComponent does not use ComponentManager, and its derived classes
(most ICMP*.cpp files) do not either. Therefore, we can skip a lot of
transitive inclusions and speed up recompiling by not including it
there.
Differential Revision: https://code.wildfiregames.com/D3962
This was SVN commit r25447.
Refs 64bfa089af and 44ec2e324e
When a .zip file is encountered by the VFS population, it reads the info
for all files in the archives. This is quite slow for the public archive
(400-500ms on my computer), which means calling GetEngineInfo()
repeatedly is impossible.
By only opening the external mod.json, we skip most of the work. The
archive can still be opened if needed as fallback.
Differential Revision: https://code.wildfiregames.com/D3216
This was SVN commit r25446.
This removes usage of CmptPrivate outside of ScriptInterface.
ScriptRequest can now be used to safely recover the scriptInterface from
a JSContext instead of going through ScriptInterface, which allows more
code cleanup.
Follows 34b1920e7b
Differential Revision: https://code.wildfiregames.com/D3963
This was SVN commit r25442.
Follows 34b1920e7b.
JSON functions and ToString are movec to their own headers.
Also clean out a few PersistentRooted usage to use the 2-phase init to
clean up scriptInterface usage.
With these functions split off, we can finally clean out headers and
remove ScriptInterface.h from most of them, in favour of smaller and
more precise headers.
Take the opportunity to clarify some comments regarding Mutability.
Differential Revision: https://code.wildfiregames.com/D3961
This was SVN commit r25434.
The one in ScriptInterface_Impl is actually used.
Introduced in e9e05f4efc and from what I can tell never actually used.
Differential Revision: https://code.wildfiregames.com/D3960
This was SVN commit r25433.
Follows 34b1920e7b.
This splits off the object-related functions, such as
[Set/Get/Has]Property, CreateObject, CreateArray, FreezeObject.
It also puts the definitions in the header itself, which might end up
with faster code here & there, though perhaps slower compilation time
(somewhat doubtful since we already included most things anyways).
Differential Revision: https://code.wildfiregames.com/D3956
This was SVN commit r25430.
All ToJSVal/FromJSVal definitions are put in a separate header.
Remove AssignOr[To/From]JSVal duplication.
The functions were already static so this is rather straightforward.
Follows 34b1920e7b and 2bae30c454
Differential Revision: https://code.wildfiregames.com/D3953
This was SVN commit r25428.
Follows 34b1920e7b.
This separates StructuredClone & DeepCopy logic into its own header,
reducing the size of the monolithic ScriptInterface header.
Differential Revision: https://code.wildfiregames.com/D3922
This was SVN commit r25419.
sizeof(const char*) returns the pointer size and not the string size,
and 'texture:' happend to be 8 characters long, which is the same as a
64-bit pointer but not a 32-bit pointer.
Reported by: Stan
Differential Revision: https://code.wildfiregames.com/D3947
This was SVN commit r25415.
make linter happy
add missing break to loop if correct comparator has been found, reported
by @Langbart
Differential revision: D3945
Tested by: @Langbart
This was SVN commit r25414.
Fixing following problems:
Issue number one:
Enable mod with a23 compatibility in a23b.
Save configuration.
Start a24.
Better result:
Mod will be enabled and invisible in mod selection screen producing
various errors.
Worse result:
Game will crash and refuse to start.
Issue number two:
Mods can silently set loaded mods without restarting the engine, so mods
can unlist themselves from compatibility detection.
Solution:
Enable necessary mods instead if running with gui and open mod page.
Open information window on top of mod page to infom why mod page is
showing up.
On mod page show mods which failed in compatibility check and color the
resposnible ones.
Disable start button without enabled mods.
Show non existed mods if they failed in compatibility check.
Else just log to mainlog and close.
Another fixes:
Display in enabled mods really enabled mods as current logic confuses
players about which mods they have enabled and is not helpful (ref
#4881)
Note:
this will not solve issue with mods claiming being compatible with
engine version while in fact being incompatible.
Comments by: @vladislavbelov, @Stan, @Imarok
Tested by: @wraitii
Differential revision: D3592
Fixes: #6044#4881
This was SVN commit r25410.
This allows GUI elements to use a texture as their "mouse event mask",
making it possible to have arbitrarily-shaped GUI elements with regards
to mouse interactions.
Used for the minimap idle worker button as a proof-of-concept.
Differential Revision: https://code.wildfiregames.com/D3814
This was SVN commit r25408.
XMPP JID has a concept of 'resources', which can be used to
differentiate multiple clients of the same account.
We currently hardcode this 'resource' to '0ad' in two places:
- The 0 A.D. client always uses '0ad'
- The network code expects a host resource to be '0ad' when connecting.
As noted in 0fd8aa2a77#31215, it is less effort to store the JI
D directly. This patch does that. It also makes 0 A.D. use a different
resource each time.
Note that resources ought not contain particular information, as the
XMPP server is free to
clobber it. I keep '0ad-' here for debug purposes.
This allows:
- multiple 0 A.D. instances to log on the lobby at the same time (not
massively useful, but good for debugging sometimes)
- hosting a game with a custom resource, which will potentially make it
easier to have dedi
cated servers on one account.
Note that hosting multiple games on one account is currently not
supported and will have weird behaviour on the lobbybots side. They
should be upgraded independently of this.
Refs #3556
Differential Revision: https://code.wildfiregames.com/D3500
This was SVN commit r25407.
GCC < 5 used to reserve the ebx register for PIC (position-independent
code) metadata. This meant that we needed to save the state of ebx
before calling cpuid (fixed in #2675)
However, the original patch from 03eaf9b461 did not force a particular
register to store this value in. Following the GCC 5 upgrade, GCC
stopped reserving ebx, and that register silently got used instead. The
code became non-sensical, and our ASM __cpuidex started returning random
garbage in edx.
Since we now only support GCC7 and above, the PIC-specific branch is no
longer necessary and is removed.
Fixes#6028. The assertion was a result of random data in ebx.
Refs #2675 / reverts 6334ee3f8b and reverts 03eaf9b461.
Patch by: nwtour
Comments by: vladislavbelov
Differential Revision: https://code.wildfiregames.com/D3575
This was SVN commit r25405.
Completes work starting in D2313 / a33fd55e81:
- Make C++ mistakes where values can be changed without messages being
sent easier to notice / harder to do.
- Make the IGUISetting interface more flexible, allowing custom settings
implementing their own logic. This is used to clean up hotkey code
introduced in 33af6da5e1.
Side effects:
- FromJSVal_Vector clears the vector being passed in. I have some vague
memory of not doing that in D24 / 2bae30c454 as an optimisation, but it
seems more like a footgun to me.
- Most usage of SetSettingFromString is replaced by direct method calls,
as we can generally cast to the proper GUI object type. Where we can't,
it is kept as a poor's man virtual dispatch.
- It moves a few member variables elsewhere, e.g. TextOwner now gets its
own member variable settings.
Differential Revision: https://code.wildfiregames.com/D3892
This was SVN commit r25392.
I last-minuted a "X = X * Y" into a "X *= Y" but that turns out to map
to "X = Y * X" with our matrix overloads.
Reported by: langbart
This was SVN commit r25391.
- We only need an XMBFile, not a CXeromyces to parse GUI
objects/scripts/... This makes the interface more consistent with other
usage, notably ParamNode.
- Rename IGUIObject::AddChild to RegisterChild and flesh out the
interface a little.
Differential Revision: https://code.wildfiregames.com/D3908
This was SVN commit r25378.
The dropdown converted the numbers to string, and the code expected
numbers.
Differential Revision: https://code.wildfiregames.com/D3924
This was SVN commit r25376.
XMB format is bumped to 4, invalidating all cached files. The
differences are:
- element/attribute names are stored after the elements themselves, and
not before. This allows writing XMB data in one pass instead of two.
- names themselves becomes offsets (instead of arbitrary integers),
making getting the string from the int name much more efficient.
XMBFile is renamed to XMBData to clarify that it does not, in fact,
refer to a file on disk.
XMBData::GetElementString is also changed to return a const char*, thus
not creating an std::string. A string_view version is added where
convenient.
The XML->XMB and JS->XMB conversion functions and the corresponding
storage are moved to `ps/XMB`, since that format doesn't particularly
relate to XML. CXeromyces becomes lighter and more focused as a result.
The XML->XMB conversion also benefits from the above streamlining.
Note that in a few cases, string_view gets printed to CLogger via
data(), which is generally not legal, but we know that the strings are
null-terminated here. Our libfmt (version 4) doesn't support
string_view, that would be v5.
Differential Revision: https://code.wildfiregames.com/D3909
This was SVN commit r25375.
We often only need to include ScriptRequest.h and not the full
ScriptInterface.h
Differential Revision: https://code.wildfiregames.com/D3920
This was SVN commit r25366.
This prints out a stack trace, which is quite helpful when debugging.
Also fixes an issue with stack trace not always being reported.
Differential Revision: https://code.wildfiregames.com/D3210
This was SVN commit r25365.
TERRAIN_TILE_SIZE is now only used in relevant places in the simulation.
This also makes it mostly possible to use CFixed in constexpr contexts.
Refs #5566
Differential Revision: https://code.wildfiregames.com/D3078
This was SVN commit r25360.
- 25332f9b86 added 'override' to a method in ShaderProgram.cpp. This
adds it to the rest.
- BinarySerializer.cpp had custom JS includes, because putting those in
ScriptExtraHeaders failed. The cause is that gloox leaves "#define
lookup" up. This has been reported upstream, and fixed here.
- VS17 complains about an uninitialised variable, though the usage is
safe.
Differential Revision: https://code.wildfiregames.com/D3916
This was SVN commit r25359.
Finishes work started in f3aedf88a6.
This removes the boost-CPP function wrappers entirely, in favour of pure
templated code in FunctionWrapper.h
The Call* functions were already heavily templated, so there is nothing
really new here. I just use tag dispatch to reduce the number of
overloads slightly.
The new functions do not need the script interface, only the script
request.
Differential Revision: https://code.wildfiregames.com/D3912
This was SVN commit r25354.
We can currently specify a custom tooltip for an icon. This extends that
to a generic tag `[tooltip='something']...[/tooltip]` for any particular
piece of text, and removes the iconTooltip hardcoding by instead using a
virtual `GetTooltipText` call since all IGUIObject descendants have a
tooltip. This is both cleaner & more efficient.
Removes the ability to set a custom tooltip style (that doesn't seem
particularly useful), it'll reuse the style of the CText itself.
Differential Revision: https://code.wildfiregames.com/D3895
This was SVN commit r25353.
The main benefits are more convenient code in the component interfaces
TU and no limit on the # of arguments that can be passed.
Differential Revision: https://code.wildfiregames.com/D3910
This was SVN commit r25350.
By making all GUI proxy objects share the same class, we can easily
confirm if a given JS object is one, and from there convert to the
appropriate types.
Differential Revision: https://code.wildfiregames.com/D3907
This was SVN commit r25347.
Usage of GL_VERTEX_ARRAY client state apparently prevented OpenGL from
using hardware acceleration on my computer (Macos 10.14 with an Intel
GPU).
It is unclear how many people would be affected by this issue, but on my
end it makes GLSL performance competitive with ARB again.
Reviewed By: vladislavbelov
Differential Revision: https://code.wildfiregames.com/D3900
This was SVN commit r25343.
CConfigDB is no longer a singleton, and its members are no longer
static.
This makes it possible to have several instances of it, useful for
testing of if we ever want to have different config profiles.
Also updates hooks to delete themselves on destruction, to fit a more
natural RAII pattern. Long-lived objects must take care to destroy hooks
before ConfigDB is reset.
Differential Revision: https://code.wildfiregames.com/D3866
This was SVN commit r25326.
It's currently possible to trigger ENSUREs in the code if loading empty
groups, which is easy enough with LOD and shouldn't happen.
Recover gracefully from these errors instead.
Differential Revision: https://code.wildfiregames.com/D3888
This was SVN commit r25320.
This patch is increasing contrast of player colours used on minimap in
minimap component instead of changing player colours themselves to not
mess with other cases, when they are used.
Differential revision: D3242
Comments by: @vladislavbelov, @ValihrAnt, @Freagarach
This was SVN commit r25312.
Emplace does not replace existing element, insert_or_assign does.
While at it:
- Clean up the 'outdated' logic and reuse it for actors
- When an actor fails to load, return a placeholder. This improves
hotloading of broken actors, and makes Cunit behaviour more predictable.
- Some minor cleanup
Reported by: Stan
Fixes#6157
Differential Revision: https://code.wildfiregames.com/D3882
This was SVN commit r25308.
To render models, we need to know the maximum bounds it takes over the
course of an animation. This depends only on the ModelDef and the
AnimationDef (and thus the SkeletonDef).
Currently, we recompute this data for each model, which is inefficient.
Caching it in ModelDef is faster, particularly avoiding lag spikes at
game start on some maps.
The animations are referred by a unique ID to avoid pointer-related
issues. I would have preferred weak_ptr, but that cannot be stably
hashed for now.
While at it, switch to unique_ptr/vectors.
Differential Revision: https://code.wildfiregames.com/D2967
This was SVN commit r25306.
Infinite loop will instead trigger JS exceptions, which will make error
reports much nicer.
Differential Revision: https://code.wildfiregames.com/D3851
This was SVN commit r25280.
(Follows 76acc4e146 where I deleted actor.rnc)
This deletes the .rnc validation files, upgrading the .rng files to
being the source of truth.
- The engine uses, via LibXML2, .rng files to validate XML schemas, to
the .rnc files are redundant.
- The RelaxNG "Compact" format is a DSL for the RelaxNG XML format. Its
syntax is unique, unlike .rng files which are XML.
- Some errors are likely - I'm not sure anyone has converted the compact
files in years.
- The tool to generate .rng from .rnc files is trang
(https://github.com/relaxng/jing-trang), which runs on the JVM and is
quite annoying to install compared to "your usual text editor".
- The JS components use the full .rng format in XML, so editing this
format is already familiar to most people that mod the game.
The .rnc files were added in 2c9939ac74 along the .rng files.
Refs #413, #245
Differential Revision: https://code.wildfiregames.com/D3824
This was SVN commit r25258.
Essentially reverts D1918 / d592bf9cb6.
Instead of copying path requests to workers, setup the result vector,
then setup an index, and compute 'in-place'.
To send messages, the result vectors are read in order. This makes the
order trivially constant no matter how many workers there are, and the
architecture overall makes it much easier to efficiently paralellise.
Tested by: Langbart, Stan
Differential Revision: https://code.wildfiregames.com/D3849
This was SVN commit r25256.
Components have some initialisation-order dependency. They could also
have destroy-order dependency, and from a RAII-like point of view, it
makes sense to do this.
Differential Revision: https://code.wildfiregames.com/D3843
This was SVN commit r25251.
This changes ParamNode to use UTF8 values internally (XMB files are UTF8
since cb9d0733ef).
This removes the need for a lot of conversions, speeding things up and
allows cleaning up the validator interface & a few other callsites.
ConstructJSVal could be a tad slower because of UTF8->16 conversions
within Spidermonkey; but the difference is unlikely to be noticeable in
practica.
Also:
- Changes `ToXML` to `ToXMLString` for clarity.
- Add a simple "op" test & show a particular behaviour of merge nodes
that I intend to change somewhat in D3830.
- Remove Component.h from simulation2 PCH - brought in too much.
Tested by: langbart
Differential Revision: https://code.wildfiregames.com/D3834
This was SVN commit r25228.
Two issues:
- The wstring_view was created on a temporary string - the data got
overwritten sometimes.
- there was some weird UTF16->UTF8 conversion going on.
Reported by: langbart
Differential Revision: https://code.wildfiregames.com/D3828
This was SVN commit r25226.
This finishes 7c04ea0211 and 1b67a079fb.
GUI Proxy objects are now easier to specialize, code duplication is
reduced, code is made safer.
In details:
- the proxy private is always the IGUIObject* pointer
- the private data is accessed through a safer wrapper
- CreateJSObject returns an opaque type to allow easier extension &
prevent errors.
- The implementation of CreateJSObject is moved near the GUI Proxy
template instantiation, and both are wrapped in a convenient macro (this
makes it so that if you use the macro, you can't forget to overload the
method, and vice-versa).
- The common IGUIObject JS interface no longer needs to be repeated.
- All specialisations are again put in the same file, for improved
compile-time & clarity, given there are so few lines now.
- While at it, implement toSource which makes it possible to `uneval`
components (does the same as toString though).
Differential Revision: https://code.wildfiregames.com/D3826
This was SVN commit r25225.
This allows empty actor files -> they map to a dummy object that renders
nothing.
The use case is automatically hiding objects at lower quality levels &
LOD settings.
Differential Revision: https://code.wildfiregames.com/D3819
This was SVN commit r25224.
- Allows compositing any two templates in TemplateLoader, and not just
filters: 'A|B|C' is now valid for any template A, B and C.
- Allows parents to be composited paths 'A|B|C'. In such a schema, if A
or B themselves specify a parent, the actual composition becomes
A|pA|B|pB|C and so on.
This allows, by leveraging the common schema of our entities, to reduce
duplication.
For convenience, templates in "special/filters/" and "mixins/" can be
included by their direct name. Others have to be completely specified.
See the two provided cases for examples:
- 'hoplite' becomes a mixin that can be used to apply the Phalanx
formation
- 'builder' becomes a mixin that can be give a template the ability to
build the standard structures, and gives the 'Builder' identity class.
This also allows deduplicating that list of tokens.
Update checkrefs & swap std::map for std::unordered_map in
TemplateLoader.
Differential Revision: https://code.wildfiregames.com/D3801
This was SVN commit r25223.
Fixes 76acc4e146 (and some 2007-2010 commit).
ObjectManager needs the Simulation to be initialised, thus needs to come
after it in the class.
Tested by: langbart
Differential Revision: https://code.wildfiregames.com/D3827
This was SVN commit r25222.
An actor file, as referenced by the VisualActor, can now define
different actors for different "quality level" setting.
In this initial version, the quality is handled directly by the object
manager.
Actor format impact:
- '<qualitylevels>' may be used as the root node, containing actor nodes
as children.
- such actor nodes can refer to a file, or to an inline actor, or
simply be inlined.
- such actor nodes may have a 'quality' attribute, specifying the
maximum quality level of this actor. By default, 255 (the maximum) is
implied.
- The actor format remains valid, but 'groups', 'variants', 'material',
'castshadow' and 'float' can be given a [minquality, maxquality[ range
via XML attributes. Outside of this range, the XML node is ignored
(making it possible to define, in a single actor file, several quality
levels).
Quality is a 0-255 value, with:
- Range 0-99 intended for lower level-of-detail actors (billboards,
etc.)
- Range 100-200 the 'normal' range for models. 100 is "low", 150
"medium", and 200 "high".
- Range 201-255 used for higher quality actors that might be used for
e.g. cinematics.
The range is wide to make it easier to add intermediate levels in the
future and it seemed easier given that an integer value of some kind was
required anyways.
Engine impacts:
- A new CActorDef class is introduced, wrapping an art/actors XML file
and its different quality levels. ObjectBase remains the definition of a
given 'actor', now at a given quality level.
- CActorDef imposes a maximal # of quality level for a particular actor
definition (5 currently).
- CUnit is made to refer to an Actor Definition explicitly, not a
particular ObjectBase.
- As a minor optimisation, variation keys are calculated on
pointer-to-sets-of-selections, instead of raw sets-of-selections, as
this reduces copying.
- some refactoring, including better const-correctness and hotloading
support via std::shared_ptr.
Differential Revision: https://code.wildfiregames.com/D3787
This was SVN commit r25210.
Swap an std::set for an std::vector, the set was useful only in an
earlier iteration of the diff. This fixes nopch.
Differential Revision: https://code.wildfiregames.com/D3790
This was SVN commit r25195.
This basically does two things:
- Allow turrets on moving entities (fixes serialisation in cmpPos).
- Allow an entity to be present on another when the other is created.
That makes it basically a boiled down version of a prior patch, omitting
the part where orders could be passed on.
This does allow e.g. ranged units on chariots that fire while the
chariot moves.
Original diff by: @sanderd17
Redone by: @Stan
Differential revision: D1958
Comments by: @Alexandermb, @Angen, @bb, @Langbart, @Nescio, @Stan,
@wraitii
Refs. #2577 by implementing the entity on platform case.
This was SVN commit r25192.
This is necessary following 2d53308e1b, which now handles key-up in
input to prevent hotkeys from firing then.
Hotkeys are a global system, which means opening e.g. the chat window or
the summary does't reset things. If you press "S" to scroll the in-game
camera down, then press 'return' to open the chat, the camera keeps
scrolling down. Following 2d53308e1b, it never released since it never
got the keyup.
To handle this nicely, we need to explicitly clear hotkeys when input
boxes get focus (respectively the console).
Differential Revision: https://code.wildfiregames.com/D3797
This was SVN commit r25186.
The motion manager introduced in bae258f9a1 makes 'chasing' (e.g. an
entity targeting another entity, which also applies to fleeing)
behaviour symmetrical: both units see the initial state instead of it
being ID-dependent.
This allows removing hacks introduced in 6a66fb8205 (and refined in
7b88b1a0f9).
Differential Revision: https://code.wildfiregames.com/D3785
This was SVN commit r25185.
This implements a form of crowd movement that I've generally called
'unit pushing' in the last few years.
Essentially, any two units will push each other away when they're too
close. This makes it possible to ignore unit-unit obstructions, and thus
makes movement much smoother in crowds.
This first iteration of this system only allows pushing between idle
units and between moving units (i.e. a moving unit does not affect an
idle one).
This is because the unitMotion logic to detect it is stuck & needs to
use the pathfinders starts breaking: units can fail to move because they
are pushed away from their intended movement, and the current logic
fails to handle this gracefully.
Thankfully, the most value of this patch in terms of player experience
is found in the improvements to group movements and shuttling.
Other impacts:
- As the short pathfinder is called less often, we can increase the
starting search range & reduce the # of max turns, both improving
collision recovery.
- The performance of idle units is slightly worsened, as they must be
checked for idle-idle collisions. If needed a 'sleeping' system, as used
in physics engine, could be implemented.
- In general, however, expect slight performance improvements, as fewer
short paths are computed.
- Gathering efficiency should increase slightly, since shuttling times
are likely reduced slightly.
- As a sanity change to improve some edge cases (units that say they're
moving, i.e. pushable, but don't actually move), the 'going straight'
logic is turned off if a short path has been computed. This requires a
few cascading changes to work correctly.
Technical notes:
- To reduce the cost of the n^2 comparisons that pushing requires, units
are only compared within a small square on a grid which is lazily
reconstructed each turn. The overhead seems rather small, and this is
much simpler than keeping an up-to-date grid.
- The design is intended to be parallelisable if needed someday.
- The pathfinder's CheckMovement ignores moving units in UnitMotion, as
that is now the spec. Idle units are not ignored, which is required for
the 'collision' detection to work correctly (see above).
Refs #3442 (not fixed - idle units are not pushed by moving units).
Fixes#5084 (the overlap can still happen, but units will push each
other away).
Differential Revision: https://code.wildfiregames.com/D1490
This was SVN commit r25182.
- Fix mistake in 4b46c09222 (similar to one in a4852c4c01) that allowed
hotkeys to fire while typing in the console.
- Prevent Alt+key to fire hotkeys in input boxes & console as that is
often text input.
- Split the hotkey triggering logic in two: a preparatory phase & an
actual firing phase.
This allows the GUI code to check which hotkeys are about to fire and
selectively allow them to go through. This makes it easier to change
hardcoded hotkeys, such as the console toggling one.
Differential Revision: https://code.wildfiregames.com/D3786
This was SVN commit r25180.
Mouse wheel hotkeys behave differently from other hotkeys (mouse
button/keys) in that they are transient: they're always pressed/released
in one go.
As such:
- They shouldn't be added to the stored pressedHotkeys
- They shouldn't be added to active scancodes.
This fixes these issues from 4b46c09222 & also skips un-necessary
processing in those cases.
Reported by: Imarok.
Differential Revision: https://code.wildfiregames.com/D3782
This was SVN commit r25176.
Change the OOS notification logic to remember the OOS-ness of each
client. Reset it on client leave.
The server will thus continue checking for OOS if the OOS-client leaves.
This is convenient to ignore observer OOS, or wait for an OOS player
without restarting the game.
Also add the turn number to the OOS dump, to fix#3348: particularly
following d4c2cf4430 the turn is likely to not be the same between
different clients.
Agree by: asterix
Differential Revision: https://code.wildfiregames.com/D3753
This was SVN commit r25170.
Follows d0a42f2f00.
Because only the most specific hotkeys can be active at any time,
releasing a key may require re-activating less specific hotkeys.
There were two issues with this behaviour:
- It was buggy, as it only checked one active key, when any still active
key can trigger hotkeys.
- "HotkeyPress" and "HotkeyDown" events where sent, as if the hotkey was
pressed, which was unexpected for most code/users (it is unusual to have
a "Press" event on key release).
This fixes these issues by "silently" re-triggering the hotkeys in such
a case. It also makes it easier for JS code to use "hotkeyPress" instead
of "hotkeyDown" for non-continuous behaviour.
Accepted By: nani
Fixes#6123
Refs #6064 (fixes the problem, but not the code weirdness)
Differential Revision: https://code.wildfiregames.com/D3766
This was SVN commit r25169.
d8ea401a95 replaced the ExploreTerritories logic, but I did not account
for the territory grid being N cells wide when the LOS is N+1.
Therefore, we could get out of bounds.
Reported by: Vladislavbelov
Differential Revision: https://code.wildfiregames.com/D3765
This was SVN commit r25161.
Observers no longer lag the game for players. There is still some time
to serialise the game when sending it to a joining observer, and
depending on the chosen 'max lag' the game may stop while observers
sufficiently catch up, but this impact too is reduced.
- Make the NetServerTurnManager ignore players marked as 'observers' for
the purpose of ending a turn, effectively making it possible for
observers to lag without it affecting the players in any way.
- Add a config option (network.observermaxlag) that specifies how many
turns behind the live game observers are allowed to be. Default to 10
turns, or 2 seconds, to keep them 'largely live'.
- The controller is not treated as an observer.
- Implement a simple UI to show this delay & allow the game to speed up
automatically to try and catch up. This can be deactivated via
network.autocatchup.
- Move network options to the renamed 'Network / Lobby' options page.
- Do not debug_warn/crash when receiving commands from the past -
instead warn and carry on, to avoid DOS and "coop play" issues.
Refs #5903, Refs #4210
Differential Revision: https://code.wildfiregames.com/D3737
This was SVN commit r25156.
NaN values could not be serialised safely because of the multiple
possible NaN numbers.
Since NaN values are usually the result of bugs or dangerous code, it
seems simpler to refuse to serialise them.
(D3205 was a safe-serialization alternative, should the need arise).
Fixes#1879
Differential Revision: https://code.wildfiregames.com/D3729
This was SVN commit r25151.
9d82ae15af introduced logic to extend the search range of the short
pathfinder in some situations. This extension was unbounded, resulting
occasionally in search domains several hundred meters wide, and path
computation that can take several seconds.
Reported by: Vico (on the forums).
Differential Revision: https://code.wildfiregames.com/D3760
This was SVN commit r25150.
Following bae258f9a1, all units moved, then all messages were sent. This
could result in units overlapping, which cannot be resolved without unit
pushing. Fixing this makes for a fairer comparison with the pushing
diff, so it seems worthwhile.
Differential Revision: https://code.wildfiregames.com/D3746
This was SVN commit r25126.
bae258f9a1 implemented CCmpUnitMotionManager, which handles motion for
CCmpUnitMotion (as distinct from ICmpUnitMotion, the interface).
The tight coupling between these two components was awkward at the
interface level, leaking underlying implementation details.
This diff makes CmpUnitMotionManager explicitly manage CmpUnitMotion,
instead of any implementation of ICmpUnitMotion, and moves files around
as requir
ed.
This:
- Makes it impossible to accidentally try to have the wrong
IID_UnitMotion managed by the UnitMotionManager.
- Allows devirtualising the calls from the manager to UnitMotion itself
(and inlining, as they are compiled in the same TU).
- Cleans up the manager interface - MotionState is now part of
CCmpUnitMotionManager.
- Cleans up ICmpUnitMotion interface - no need to provide a private
interface for the manager.
Differential Revision: https://code.wildfiregames.com/D3732
This was SVN commit r25125.
