5317: Fix ConSwitchDoorOpen r=C0D3D3V a=trml
Just figured out that my pr introduced a bug in switches set from map configs (not opening the correct switches). Sorry about that!
Thanks to Faulty for finding it!
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: trml <trml@users.noreply.github.com>
5311: drop small NETOBJTYPE_SWITCHSTATE r=def- a=C0D3D3V
I'm too lazy right now to write code to test this, but I think this should work.
What do you think, is it sufficient? `@heinrich5991`
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: c0d3d3v <c0d3d3v@mag-keinen-spam.de>
4705: Refactor switch state, add prediction of switch state toggles r=def- a=trml
This is an attempt to fix#4702, and also ended up being a more general attempt at predicting switch on/off tiles as well as timed switchers (with delay). A few things:
- Refactored the switch state array from CCollision to CGameCore (second commit)
The idea here was that CCollision holds information about the map that is immutable and there only exists one copy of it, while the switch state is mutable and would have both a predicted and non-predicted state in the client. While strictly not necessary, having it in the gamecore seems to represent the mutable state better and would be useful to avoid having to put the snapshot switch state back into the gameworld on each new prediction as well as to more easily keep track of the two copies.
A concern here is whether this could in any way change server behavior (time of initialization and/or destruction of switch states), so this change would require quite a bit of testing and/or review. However, from what I could tell from the code, CCollision is only initialized once and at that time where the gamecore already exists.
- Added a new extended netobject containing the end tick of switches (third commit)
This would be needed for adding general prediction for timed switchers. The idea is to only send it for a limited number of switchers (those with the lowest time left before they toggle back), since there will most likely be a limited number of timed switchers active at any point, and I randomly selected 4 as a number here. Ideally, this netobject should be merged with the previous one for simpler parsing of the information, but that would sacrifice compatibility with older servers (and/or clients). (Another possibility here would have been to treat timed switchers as regular on/off tiles and make the addition of the netobject a separate pr)
- Used NetArray from datasrc/seven/datatypes.py to simplify both the new and old switch state netobjects (first commit)
(Haven't checked if there might be reasons not to use it? But it did seem to be backwards compatible with older ddnet clients when keeping the memory layout the same)
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: trml <trml@users.noreply.github.com>
5271: Time out for POST requests too (hopefully fixes#5198) r=heinrich5991 a=def-
Untested because the issue is sporadic. But I think it makes sense to have a timeout even if this is not the root cause.
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: def <dennis@felsin9.de>
5277: penalty tiles can be used to get a time of 00:00:01, this is fixed by letting the tees die, before that happens r=def- a=C0D3D3V
Also map gets reloaded if the server tick overflows. We could add a warning, that the server reloads soon, or save all players and load them after map reload. But I think that is such a rare event, we can just reload the map without these nice things.
fixes#5170
we should check the following maps for cheated times:
- just every fly
- puzzle partners
- time shop
## Checklist
- [x] Tested the change ingame See https://youtu.be/lQPU60XN8Nk
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: c0d3d3v <c0d3d3v@mag-keinen-spam.de>
5269: Introduce and use constexpr CCharacterCore::PhysicalSize r=def- a=Kaffeine
Add `Collision()` getter to server CEntity to make it easier to sync the server and prediction implementations (e.g. `IsGrounded()` imps are now identical).
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: Alexander Akulich <akulichalexander@gmail.com>
5265: Cleanup prediction entity handling a little r=def- a=trml
I looked over the remaining code for creating/removing entities in the prediction, and only ended up making some smaller changes:
- More consistent use of InsertEntity/RemoveEntity/Destroy
- Fixed one instance in gameclient.cpp where an entity was removed but not deleted (could potentially have been a memory leak)
- Removed some old code for retaining information about the characters after they left the snapshot (the client should only rely on the snapshot for prediction of players)
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [x] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: trml <trml@users.noreply.github.com>
5252: Use std::vector instead of array, remove base/tl/array.h, algorithm.h, allocator.h, range.h r=def- a=Robyt3
Replaces all usages of `array` with `std::vector`.
I adjusted variable names of variables I changed to use the `v` prefix. Not so in the editor however, as there are already many changes due to clang-tidy enforcing the use of for-each loops.
This allows us to remove all remaining `base/tl` headers except `threading.h`.
Clang-tidy now finds `clang-analyzer-cplusplus.NewDelete` (Use of memory after it is freed), which is also fixed here, though it appears to be a false-positive.
This last remaining usages of `goto` are also removed.
## Checklist
- [X] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: Robert Müller <robytemueller@gmail.com>
5259: reintroduce shotgun bug r=def- a=C0D3D3V
fixes#5258
bug got fixed by the math changes in 68bcd21eff
This fix introduces a hard coded velocity that the player would get like before. (It could be that this has slightly other behaviour, because before the produced NaN values where converted to ints in CCharacterCore::Quantize (so some other logic may be confronted with the NaN values before), but it should most likely not be noticeable because it is fast enough)
We could now make this bug optional, by adding it to our mapbugs, or adding a server setting for it.
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [x] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: c0d3d3v <c0d3d3v@mag-keinen-spam.de>
5255: fix windows server crash on client spawn r=def- a=C0D3D3V
The continuation of removing mem_zero over CCharacterCore.
I have tested the server in a Windows VM -> no crash
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [x] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: c0d3d3v <c0d3d3v@mag-keinen-spam.de>
5254: Fix ingame menu button behavior when holding mouse button r=def- a=Robyt3
Fixes the following behavior:
1. Player is ingame and the ingame menu is opened with Esc
2. A mouse button is held down on one of the ingame menu buttons (e.g. the spectate button)
3. The ingame menu is closed without moving the mouse away from the button
4. The mouse button is released while ingame
5. The menu is opened again
- Current behavior: The hovered button is immediately activated, as soon as the menu opens, due to the released mouse state being handled as a click.
- Fixed behavior: The button is not activated. Instead another call to `FinishCheck` is added so it's also called when the menu is not active, to clear the active UI item in that case.
As well as the following:
1. The menu is closed while the mouse cursor is hovering over a button.
2. The mouse button is pressed and held while the menu is closed.
3. The menu is opened again.
- Current behavior: The menu button is already held down and when the player releases the mouse button it will immediately be activated.
- Fixed behavior: The button is not activated immediately, by clearing the hot item and next hot item variables in `FinishCheck`.
The latter behavior can still occur when entering the editor instead of closing the ingame menu.
Closes#3560. Closes#5229.
The first two commits are identical to upstream https://github.com/teeworlds/teeworlds/pull/3158, where those issues also exist. The last commit adapts upstream behavior so buttons can only become hot when the mouse is not pressed.
## Checklist
- [X] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
5257: Refactoring: Move CMenus::UseMouseButtons to CUI::SetEnabled, extract CRenderTools::RenderCursor r=def- a=Robyt3
Adopt changes from upstream.
## Checklist
- [X] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: Robert Müller <robytemueller@gmail.com>
