Guild icon
Teeworlds
discord.gg/teeworlds / development
For discussions around the development of the official Teeworlds
Between 2020-04-11 00:00:00Z and 2020-04-12 00:00:00Z
Avatar
Well not clamping the fields that dont actually have a limitation would be the first half
07:07
To complete it when we do clamp fields due to a limitatiom in the map format or the editor itself, we refer to a constant defined in the respective one
Avatar
a04a33e Add command usage - Learath2 c366b3f Handle new type argument strings in commands - Learath2 60fbcf9 Truncate the helptext to fit - Learath2 2165d99 Get rid of the evil goto - Learath2 2fc4a0d Usage for commands, make whisper consistent - Learath2
12:09
Basically a feature that snaps automatically a quad to another quad's corner in the same layer. As it is now I find it difficult sometimes to align the quads properly without having some visible 1px space between them or some edges going off. When the cornes of both quads are close enough the quad you're moving should automatically snap to the closest corner
Avatar
As mentioned in #1911 this tileset had an assymetrical issue which when repeated in a map resulted in displaying an unnatural pattern. This should improve the visual appeal of it. Optimized and tested in game. Before: !jungle_background_fix_before After: ![jungle_background_fix_after](https://user-images.githubusercontent.com/2030615/79043951-c4196980-7c02-11ea-8683-049296b761ac.jp...
Avatar
Signal when the map has been successfully saved. Something like this: !image Have it pop up and hide after a moment. Could also be used for error reporting? If the map fails to save for example. This is useful now since we don't have a save icon to save the map and indicate that it is saved by disabling the icon, but it could be redundant later.
Avatar
Part 1 Extract the important part from the DoFileSelect() function. !image Basically remove the preview and bottom part. Have it return true when selecting a new file. Then update the RenderPopupMapLoad() and RenderPopupMapSave() accordingly (move the code from the DoFileSelect() function). Part 2 Make an "Add Image" popup in the Asset Manager, to load and add an im...
Avatar
Can anyone confirm that the none background on master is broken?
Avatar
i think it always needed cl_show_menu_map 0
Avatar
on 0.7.4 it works just by clicking on the thing, for me
16:37
not on master
Avatar
@LordSk 🦋 Why do you want to abstract away a weird partition of the file selection dialog?
Avatar
It is more flexible this way, it only does the file selection. Then we can do whatever we want with it. It is either this or build a "general" thing. Like how do you tell it to preview maps instead of images?
Avatar
Well at the end of the day all previews will be images, (preferably cached in a hashmap). So what's displayed will always be an image
Avatar
Will they though? We could render a tiny portion of the map (say the flag area) as the preview.
Avatar
I'm not even sure the map rendering code is efficient enough to be doing this
16:49
I was thinking an asychronous job that builds a preview image and that job can decide how to build a preview for a given file
Avatar
Could be the way. But until we know, I think splitting up the file explorer a bit and only keep the part that likely won't change is a good solution.
Avatar
Can one of you check no-theme master?
Avatar
sure
16:52
Avatar
eh whatever, I need something to code to distract myself, I'll just do whatever you want
Avatar
@LordSk 🦋 thanks, I guess the Robyte fast-loading thing broke it
Avatar
How do you want to resize the common part when you want to insert a preview there, or will there always be space left for a preview?
Avatar
probably pass a rect to the DoFileSelect and work from there?
Avatar
The region you highlighted is not a rect
Avatar
Yeah that's true hmm
Avatar
Move the search to the left side and eat up a bit of the path string rect thing?
16:57
It does not have to be pretty for now
Avatar
That, or we could pass a CUIRect pointer DoFileSelect can assign to, if the pointer is null don't leave space for a preview, if the pointer is not null assign the preview rect to it?
16:58
Tbh I don't quite get your design goals with this new editor: The code doesn't have to be pretty, the ui doesn't have to be pretty. We already have an editor lacking in both
Avatar
I did not say the code does not have to be pretty. The UI is temporary.
Avatar
code that is more general is not always prettier 😛 using as little abstractions as possible should be a goal by itself
Avatar
I guess
Avatar
@David - SFM got any updates on https://github.com/teeworlds/teeworlds/issues/2388 ?
Add sounds for events (start, finish, record, checkpoint?) in race.
Avatar
@Learath2 do the issue however you like if you have a better solution
Avatar
If you prefer we could revert the change to ed2_map and instead strip mapres from the beginning before calling EditAddImage. The background checkers for the preview are also scaled for assets that aren't square, we could fix that if you want. Closes #2550
Avatar
@Robyt3 what was the intention with moving the clear out of RenderBackground anyway? Closes #2552
Avatar
oh, nice find @Learath2
Avatar
@Dune what kinda worries me is that it was working for robyt
Avatar
maybe he had gfx_clear 1?
23:13
that would explain why it looks yellow like this
23:13
I'm surprised robyte wouldn't have noticed a change in color of his menus....
Avatar
ah, that would indeed explain it
23:24
A retro multiplayer shooter. Contribute to teeworlds/teeworlds development by creating an account on GitHub.
23:26
Loading the map files is undefined behaviour, the only reason it's not been an issue is that sane modern compilers just don't put random padding between ints. But they are completely free to, thus you are not allowed to assume the object representation of a struct
Avatar
That's an issue in editor v1 and v2, and not ingame, right @Learath2
Avatar
Well we save the quad layers without serialization. Thus whenever we load them we assume the object representation of struct CQuad in the map file matches the running clients representation. So its not isolated to the editors
Avatar
Ah. Well at least it does not seem exploitable
Avatar
Or I am not intepreting the standard correctly, but even if the compiler werent allowed to pad it, the ints themselves are not guaranteed to be compatible
23:34
But yes, not exploitable at all
Exported 55 message(s)