#0.14 Bevy Jam
1 messages Β· Page 6 of 1
well, it does but it doesn't
it can access the cache from main branch releases but not from non-main branch releases
i'll give it a bit more thought
i think the tradeoff in the PR is probably worth it. may want to allow cache-group to override the workflow file hash as well
so that cache-group can work between different workflow files -- but that also means different versions of the same workflow file
actually before this PR you could have another very strange thing happen
suppose we had a job in ci.yaml named build for some reason
then it would use the same cache key as build in release.yaml, because the tuple is the same: (runner.os, job name, cargo --version, Cargo.toml, Cargo.lock)
so if the two build jobs compile with different profiles / features, then one of the jobs will always get a "cache hit" but forced to recompile
there's a problem with cache + matrix as well
job name is the same for each matrix job ("build")
since web and linux releases both run on ubuntu-latest... they use the same cache key
i saw in one of our releases, the linux build was super fast because got a correct cache hit but the web build was slow because it got a false positive cache hit and had to recompile
that's probably on our end to fix though, with cache-group: ${{ matrix.platform }}
or maybe upstream could do ${{ github.job }}-${{ strategy.job-index }} EDIT: updated pr
gonna leave bevy-template CI as it currently is atp; we know what the issues are, and it can be improved after the jam
Thanks for upstreaming the CI wizardry π
Thanks, I will review this tomorrow (if not feel free to ping me)
(replied to the wrong link, meant the cargo-cache PR)
It might be better to have the workflow file hash at the end so that we have a higher chance of being able to reuse the cavbe but I'm not sure yet
But definitely a good addition, that might explain some recompile issues I had in the past.
I wish cargo would provide better verbose output explaining why things are recompiled...
Moving to voting in #jam-theme-voting now
yeah you kinda have to look at the cache key and compare it to the cache keys in other jobs / workflow runs to figure out what's going on
planning to get the README finished today so we can announce that the template is ready by tomorrow
thx!
I was planning on making the design decisions and third party tools separate docs. What do you think?
sounds good to me, i think the CI should be a separate doc too lol
huge linear README is just not readable
and i don't think we should leave out details to try to keep the README small
put the separate docs in a docs/ folder or something? idk what the convention is
Stick them close to whatever they're documenting ideally
ah okay. so third-party tools probably in the root by Cargo.toml and CI in .github/workflows?
assuming it's valid to put a .md in the workflows folder
probably fine if they're all in the root ig
I'd put them in a docs directory in this case
idk why hackmd is not letting me share this note i made
Also, thoughts on renaming the template to bevy-jam-template?
I'm asking because bevy-template is a really really really generic name
Which I think would be fine if it was in the Bevy org
lol my jam template is named bevy_jam_template but i was thinking of renaming it because it was too similar to bevy-template
But as-is, it makes it a bit awkward to talk about
are we specifically targetting jam games though?
i thought it was meant to be a general template, jam-friendly
Kinda?
That's true ofc.
But the fact that it targets the Bevy jam specifically brings with it some consequences
like beginner friendliness
and easy itch uploads
semi-official-bevy-template π
starter-template or something?
That's at least better than bevy-template
@prisma delta do you have some wise recommendations?
if we're renaming, can we switch to underscores π©
sure!
itch page can still be hyphens cause it's more natural in a url
I kinda like bevy_starter_template, but I also like bevy_jam_template 
bevy_quickstart (like https://bevyengine.org/learn/quick-start/introduction/)
uuuh I like that one as well
bevy_jam_5_starter
Not durable :p
And then rename it every jam? π
bevy_jam_5+_starter /j
heh, well I look at this template and I see it extending bevy api with extension traits and stuff and it feels like a departure tbh
i think bevy users should be using extension traits
but I'll bow out, it is not my place to be giving opinions, it is y'alls work
sugary_template, you need some sugar to start making a jam
You really really have to do that ATM to get a nice API
I kinda want to have a more "boring" name because it is supposed to feel like how an official Bevy template would look like if it existed today
I do think this captures the vibes pretty well π
I like this one most so far
It is a pretty cute name @bleak grail :>
The more I think about it, the more I like it
Or something like bevy_sugar
Aaaaaa
fwiw, I agree. I think we're gonna need a great mini-tutorial about how to use the template. It's just a lot. This will be something that people who are completely new to bevy try to use, and I just don't see it going super well for them.
but I've also taken the position that there's no reason this thing shouldn't exist and I don't wanna be a party pooper π
i'd only think it sounds like syntax sugar
We are cursed with so many great names π
I think we already cut down a lot of that
But I'd love to hear some beginners' thoughts on it
yeah I agree and I'm planning on making a video on it to promote and explain what's going on if its ready far enough in advance of the jam start. I also don't want to party poop.
iirc extension traits are only used for ui widgets rn
There's definitely room for cutting more stuff that we simply don't see since we are blind to issues beginners might face
That's pretty sweet of you β€οΈ
Yeah, ideally that would be in Bevy proper
ah we have a Spawn extension trait defined on Commands and ChildBuilder ^\_(``/)_/^
But as a bit of a context, this way of doing UI was recommended to us by #ui-dev as the current best practice when not using an existing UI library
So I'd be very very skeptical of changing that part
I "get" that trait too / it makes sense. It's just a bit advanced.
Yeah that one is more contentious π
That's fair π
it's only used internally in ui/widgets.rs but yes
Do you think it would be more approachable with documentation in a dedicated "design philosophy" document or with inline documentation?
Also, I hope I'm not coming across as overly protective of the project or anything. Criticisms are still very much welcome π
we already have our own templates we're presumably more protective of hehe
hmm, I feel like the most important thing is a good top-down overview. I feel like there's no good place to "just start writing code." You have to first grok the whole project layout and figure out what you're doing and think of a good name for your little module and blah blah.
yeah that's a consequence of splitting into modules by feature
agree we should explain the project layout
I don't think I have access to the readme hackmd so I don't know what's already been done
should have access if you're signed in
oh maybe I just wasn't signed in, lol
Ideally, that would be the game module. Everything there is "just for show", i.e. the ducky moving around. You can remove all of that and just write stuff
I think to start with it wouldn't be wrong to just write all of your game code directly into /game/mod.rs and then split it as you see fit
Maybe we should also mention that π
i think the screen/ directory is pretty easy to slot into as well
delete a file if you don't want that screen, add a file if you want a new screen
but ofc that needs to be explained
@dawn crag feel free to edit the readme as you see fit or add comments if you find something is written in a weird way
btw @slender belfry, no license files => all rights reserved, not public domain π
Yes. Does the readme imply otherwise?
@slender belfry does trunk serve create a dev build or a release build?
dev
you can pass --release to get a release build
well, --release --no-default-features
BTW we could add trunk serve --no-spa as a VS Code task in tasks.json
that's good idea π
so im fine with whatever name is chosen for the template, doesn't really matter to me, but i'll lay out my reasoning against sugary_template:
- may be confused with "syntax sugar", aka this template provides a lot of api ergonomics, which is not the intent
- user has to make the connection to "game jams" to get the joke
- the name implicitly ties the template to game jams, which isn't its sole focus
it is cute though and i think a cute name is good, if it fits better
You just convinced me to change my vote
Yeah, I didn't for that one for reason 1
I would've suggested Bevy Pectin myself π
Feel like there might be some potential confusion with https://bevyengine.org/learn/quick-start/getting-started/#quick-start being a thing that exists
that was intentional but maybe it's not a good thing? ;p
a couple refactor proposals:
- merge
camera.rsintolib.rs(imo it should be separate, but with window / asset / etc. all being merged intolib.rs, it doesn't make sense rn)
2. renamesrc/ui_tools->src/utiland movesrc/ui_tools/widgets.rstosrc/ui/widgets.rs(user will add their own utils and their own ui. i think this makes more sense and is easier to explain)
having to explain src/ui_tools in the README feels really weird
I agree with both
Wait, I misread
so there will be /util and /ui?
as proposed yes
That feels a bit weird, considering everything in there is ui themed
true. could just rename src/ui_tools -> src/ui instead
sounds good π
is there a particular reason we are urging people not to use bevy_audio?
missing features iirc
note that we had a ui folder initially and we specifically changed it to ui_tools so it's not confused with screens
well we had it as src/util/ui at the time
the issue was apparent complexity i thought?
that was the main drive, but we also did naming revamp and that was a part of the reason
not gonna fight it, just letting you know
potentially confused with screen, maybe
we should include reasoning alongside the recommendations anyways
Best to ask that in #1236113088793677888
or #audio-dev
i mean, it's basically confirmed to be replaced
I sort of understand the reasons that bevy is moving towards kira, (rodio just being poorly designed, apparently). But afaik there aren't any showstopping things in bevy_audio right now and I don't know if telling people that an audio engine swap is a "must have" is right.
just a matter or release or few, there was serious discussion about making a custom sound lib for Bevy
i remember running into a limitation for something super basic in bevy jam 3
so i had to switch
bevy is moving towards an in-house audio engine afaik, so not kira anymore
The current implementation is not reliably real-time (as far as I know) and also makes it really difficult to do stuff as basic as loop tracks.
For a jam I'd probably consider using it though
yeah i think my issue was something like this
i like.. couldn't change the track to a victory track. or something
was just impossible
But yes, please give reasoning for recommendations like that
Yeah, I think for jams usability is key
They shouldnβt be told not to use it, but we could warn them they are going to have a hard time and apologize in advance.
Missing features probably come second
do tracks not... loop? like I recall having looping music in all of my jam games, lol.
Okay nvm, that sounds bad even when only used in a jam
@bitter sparrow I'm still feeling under the weather, but I think I should be able to do a pass on the readme tomorrow. Do you mind waiting with announcing everything until then?
it seems we need to use
bevy_kira_audioto support stopping a looping sound (the main track) to replace it with the victory track
found my old quote explaining my reasoning
yeah for sure, readme needs to go through code review anyways
But yeah, if there's a good reason for users to do that, please explain that reasoning. Ideally, make sure that reasoning is actually true circa Bevy 0.14.
Otherwise, I just really don't like telling new users that bevy_audio "is not usable." (that is the leap I am making from "bevy_kira_audio is a must-have plugin.")
yeah that's fair, the quote was true as of bevy 0.10
it seems we need to use bevy_kira_audio to support stopping a looping sound (the main track) to replace it with the victory track
This seems sort of surprising, but also I would be very surprised if there wasn't a very simple workaround if it's still a thing. It definitely sounds pre-"ecs-based audio"
We could leave it in the "nice to have" bucket and then re-evaluate after the jam
It would be pretty nice if bevy_ui turned out to be good enough for jams π
eh i used pure bevy_ui for bevy jam simulator
not that that was a good idea
bad multi-line text rendering and no scrollbar widget were the most annoying things
hopefully text is good in 0.15
was there a doc for the direction of the template?
we never ended up making one (yet)
it feels really awkward to be seeing things like "bevy_audio and bevy_ui aren't good enough to use", so was wondering what the charter was for the template
I do wanna write about that when my energy returns
i mean it's well-known that these are sore spots in bevy today no?
could be better and "unusable" are very different things
We're sticking to pure Bevy with no 3rd party plugins as that was a hole in the template ecosystem
if you're making a jam game you don't want to waste days working around upstream issues
i don't think anyone is saying unusable
So that's where that sentiment is coming from
We're phrasing it as "pretty much a must-have for jams". Not saying that the core Bevy libraries are unusable (bevy UI certainly is pretty capable, just missing widgets)
yeah I suppose my read of "its a must-have" is that the built-in functionality is not enough, even for game jams.
can be careful about phrasing when that section is written out
Yeah, strike some sort of balance where we don't present Bevy as being incapable on its own, but also glorify the third party ecosystem.
I definitely see where that read comes from
something like "these 3rd-party libraries are highly likely to save you time during a game jam"
Also feel free to just write that directly into the hackmd
comments welcome π
I don't have access because its private and not on github π
You should be able to access it when you're logged into hackmd
Inside that document, there are links for the hackmds of other docs
I don't really want to make a hackmd account, so I'll leave it to y'all and stop commenting.
understandable, i'll prob @ you when i make the README PR
Feel free to comment on Discord, I can copy your sentiments into the hackmd π
I very much value the comments we are getting β€οΈ
IMO bevy_ui is usable for 90% of jam games
Unless you're doing something UI heavy, buttons are generally enough
even for buttons you kinda want bevy_mod_picking :)
but yes it's simple enough to implement buttons without it ofc
iirc the main issue with Interaction is like having the button remain pressed on click & drag, or having a button take effect upon release (but only if the mouse is still above the button), stuff like that
Yeah, but again, for jam games, that's generally fine
if you just want "on mouse down" it's good
@prisma delta could you pin this? I want to leave it up until tomorrow
75% of the top 3 entries from previous bevy jams used bevy_ui exclusively. The rest used bevy_egui, a combination of bevy_ui and bevy_egui, and one said f-it and did their UI with sprites.
the sprite ui π
Thanks for checking π
also notably sickle_ui didn't exist for previous jams. (and as a counterpoint, bevy_ui has improved over time)
Curious to see how many will use sickle_ui this jam
I think we should copy-paste this information into the hackmd
imo bevy changes so much every 2 releases that it'd be too outdated to be relevant
but maybe
i'm gonna make separate PRs for each readme doc
so there will be a few broken links until they're all uploaded (will link to TODO at first)
@unreal narwhal (3rd-party recommendations not included in this PR yet, just the main README)
there may be a few other things we still want to add to the main README, but i think it's in a good enough place for an initial PR to get things moving
think adding --open to the trunk serve command (in the README and the tasks.json) is worth, equivalent to how cargo run will open a window once it's done building
The Trunk.toml already sets --open
In theory. It never opens the browser for me, but that might be a Windows bug
hm ok. did not work for me on linux :)
Maybe itβs parts of the next version of trunk, like no-spa
In that case, sure, letβs add the flag to the CLI invocation
ok on that note, created pr https://github.com/TheBevyFlock/bevy-template/pull/141
i'll update the readme pr as well
ok so uh
in our Trunk.toml
[server] is supposed to be [serve]...
i changed it to [serve] and open = true took effect
π
idk how to test if no_spa = true is working?
the game runs fine
updated other PRs accordingly
game runs fine with no_spa = false as well so idk
okay i checked web console
confirmed that no_spa = true in Trunk.toml works when [serve] is used instead of [server]
If changes what happens when a resource is missing. Bevy expects missing assets to return 404, but trunk by default just redirects to index.html
You can test it by navigating to localhost/foo
If it returns a 404, thatβs good. If it redirects to localhost, thatβs bad
Well that explains things π
i set up Bevy Jam 5 and Bevy Jam 6 milestones to organize issues better
can't recommend blenvy / blender bevy components workflow, because the latter is still 0.13 and the former is not released yet
ill omit the "other tools" section
Yeah, optional. I can definitely see arguments both ways.
Question motivated by my brain, for whatever reason, telling my fingers to type https://thebevyflock.github.io/bevy-template into my address bar and expecting it to work.
IIRC, it was done previously in a PR to bevy_github_ci_template, but that PR also massively reorganized the CI in sort of a controversial way and stalled out.
Seems unimportant for Bevy Jam, but maybe slightly useful for jamming in general.
That's a good idea. I think it's nice to give users the option to self-host. It shouldn't be too hard to implement either. Adding an issue so we don't forget π
Ehh. I feel like deploying to GH Pages is a bit much, and doesn't really bring much value, while also adding more areas where someone who is new could get confused.
(Also feels kinda sketchy to me tbh, but that's more of a me thing.)
I also don't think it brings that much value, but it's trivial enough. I don't think there is much of a vector for confusion since everything there can be automated away.
With itch, you have to add your API keys and create the project. With GitHub pages, it's enough to just have a workflow.
You mean you find games that are self-hosted sketchier than games hosted on itch.io?
Also useful to have beta builds deployed that way
Also I just generally dislike adding in more things that tie this template specifically to github.
actually @lunar sail i just had a thought. it may be better to leave the workflow file hash in cache-group, but add the entire cache-group to fallbacks
so there can be cross-job (or cross-cache group) fallback caching
different jobs might run the same or similar commands
would be even better to make the fallback more granular, preferring the same workflow file + job name but different matrix index first, then the same workflow file but different job name + matrix index, then different workflow file
but that would be more complicated because the cache-group can be overwritten
Hmmm right
While trying to find the cargo-cache repo (thanks terrible github search) I also found another action doing something similar to ours and they also provided more than just one user configurable part
Basically at different positions of the hash key
But I'm also thinking I might just merge what we have now as its definitely am improvement and then we can refine the behavior later on
Might have to do a major version bump though
makes sense
Fixes #53.
Related to #14.
Will link to this from README.md once either PR gets merged.
@unreal narwhal tool recommendations this time
why is it a set of specific crates instead of a link to https://bevyengine.org/assets
it includes that link
but the reason is to highlight a subset that are particularly useful for pretty much any game
so there's a quickstart template repo that doesn't include crates that are particularly useful for pretty much any game?
yep...
there's not much actual game code in the template
so it's not too bad, but as soon as you want like, physics
or debugging tools
then these crates would become useful
I think a guiding document would've been useful for the template effort. its not clear to me what the goals are, so I can't really give any useful feedback on the list.
if its an official template, then these are official recommendations.
it's meant to be a semi-official template
this is the closest we have to a guiding principles writeup #1258521739395203174 message :p
if #1 is "no third party dependencies" then scrap the list entirely imo
point people to /assets
hmm purely from a pragmatic standpoint, i think the list adds a lot of value (in addition to linking to bevy assets)
I think the list plays favorites in a way that isn't clear, and I disagree with the top-level statement of "the following libraries will likely be useful for most games"
in terms of "being neutral" it is a deviation though
Agreed. The bevy/assets page has a lot of libraries. From looking at it, it's not clear to a beginner that something like LWIM is on a whole other level than other crates
A recommendation list is inherently playing favorites. I mean, that's its job, isn't it?
How would you rephrase it?
Maybe "We find the following libraries are particularly useful for many kinds of projects"?
"some crates the template authors have vetted and think you might find useful"
Or "The following libraries have, in our opinion, an exceptional level of usefulness / documentation / maturity / widespread usage"?
make it more clear that it's subjective?
That's good
yeah again my comment in response to being asked for feedback is "I don't think I have useful feedback here", because I don't know that the list should exist. Especially given the variety of things like ui libraries. Why sickle and not quill or lunex or any of the others, for example
yeah the UI libraries is one thing i was unsure on
sickle_ui is mainly because it adds immediately useful features (widgets) and builds on bevy UI
but there are a lot of great UI libraries
also worth noting concretely that bevy_hanabi doesn't work on wasm afaik
so if you use it in the jam, its unlikely that you get votes as a result
actually i believe it might now, i checked the issue yesterday
let me double check
i havent read the full issue context to know what exactly this comment means tho
I gravitate towards recommending sickle_ui because at this moment it seems like it is the most representative of how bevy_ui will evolve. I recall @prisma delta writing something about "We will not upstream bevy_ui, but with a very high likelihood port large parts of it in some form or another into bevy_ui"
But I think it's fair to also mention the other UI crates
yes, but the assets page already does that, doesn't it?
Yeah, that's why I'd rather keep this list short if possible
i could remove sickle_ui and bevy_hanabi π
i think all the rest have fewer solid alternatives
Its a bit of a toss-up between keeping the list short and making it more inclusive
well hanabi doesn't really have solid alternatives for particles afaik, but like vfx
There's some CPU particle crates
Haven't tested them tho
I'd feel bad not recommending any UI library
hm yeah.. could be a separate section for UI, to include a few more libraries :p
yeah again, my comment in response to being asked for feedback is "I don't think I have useful feedback here". The template is straddling a line between not having third party crates and turning around and immediately recommending crates that could require refactors (bevy_asset_loader and loading states, for example).
(and to be clear, I love bevy_asset_loader, so no knock on the crate)
yeah that's true. the reason for that is that bevy is in an awkward spot rn for an "unopinionated pure-bevy template"
and yet we're trying to make one anyways
to the extent that it's possible
I think that's alright. This is mostly Bevy's official response to such things, so it makes sense that we would adopt that as well.
if the list of (curated) recommendations could be hosted outside the template somewhere we could link to, that might be slightly better
The refactors should be trivial. If not, we should change the template to make refactors for third party crates trivial
Github Repos have Wikis we could put it in if we wanted.
yes and I think this is this opinion coming through in the implementation. The template authors feel that Bevy by itself is not enough for a game jam.
I wouldn't. At that point, you're just creating yet-another-awesome-bevy-list
fair enough
tbf.. i do feel that way
yeah I only said it because I've seen you say it π
"not enough" is a bit of a strong word. Can't speak for the others, but at least my personal opinion is that pure Bevy is certainly enough for a jam, it's just that some things require more boilerplate than what you'd expect from a game engine
i'd consider that "not enough" since jams come with a tight time constraint
personally
Hmm, fair point
Prime example is LWIM. I had a stab at adding input mapping to the template, but I realized that I was basically recreating a subset of LWIM. Instead, a user can just trivially add LWIM themselves now.
@bitter sparrow I'll later check your readme PRs in more detail. I'll probably accept them without much fuss and then do a follow-up PR myself with some changes.
@unreal narwhal do you think adding something like this to the recommendations would make this seem less arbitrary?
I think if there's a strong reason like that then definitely include it as a note
does make the rec even stronger
"likely to be upstreamed" could be noted for a few of those libraries
LWIM, picking, sickle (partially)
kira_audio was going to be potentially upstreamed but plans changed
avian might be upstreamed at some point in the future but that's more speculative. some type of ecs-first physics for sure
picking has its own working group, so definitely makes sense -- https://discord.com/channels/691052431525675048/1236111180624297984
Still multiple years worth of work left for that, according to @limber dragon
that makes sense, it's a tall task
depends on how busy university will make me :P
can't you enroll in physicengineology?
bevyology
Ornithology
Just don't tell them which kinds of "avian" you're studying π
@slender belfry I would put something like "Uses bevy_ui primitives and well-aligned with bevy_ui's planned future" for sickle_ui
first draft
In particular:
- `leafwing-input-manager` and `bevy_mod_picking` are very likely to be upstreamed to Bevy in the near future.
- `bevy_editor_pls` and `bevy_mod_debugdump` help fill the gap until Bevy has its own editor.
- `bevy_rapier` or `avian` help fill the gap until Bevy has its own physics engine.
- `sickle_ui` is well-aligned with `bevy_ui` and helps fill the gap until Bevy has a full collection of UI widgets.
after the full table
I'd probably recommend bevy-inspector-egui over bevy_editor_pls, just because it's a bit lighter weight
And I would probably recommend bevy_egui for making dev tools
interesting. they're both made by the same person also :)
Yeah, the latter wraps the former
i was considering recommending both, or just bevy_editor_pls
would both be reasonable?
Pick one IMO, but no strong feelings
actually the decision is easy. bevy_editor_pls does not support 0.14 yet, bevy-inspector-egui does
Anecdotally, I have started doing my tooling through bevy_console and found that process to be more productive than fiddling with UI. Of course that does not work for e.g. graphs, but it's nice to be able to just enter spawn enemy 10 π
i was considering adding a dev console to the recommendations
i think i didn't because of bevy version compatibility
How should we call the template repo?
5
21
2
bevy_quickstart
ope
bevy_quickstart it is π
btw bevy-console has a PR updating it to 0.14, but not merged yet
renamed the repo
bevy_dev_console does not
hype. now the renamening begins
imo we should add a link to the bevy quick-start guide in the README now
I mentioned hanabi in a hackmd comment -- it can't / won't ever support webgl2 and doesn't yet work properly with webgpu.
to ease the potential confusion a bit
ah. is this not relevant then? #1258521739395203174 message
if it doesn't work on web i will remove hanabi from the list
I also heard that a certain coding wizard in my area is working on a for-real-for-real console in Bevy π
-# @night saddle
ah cool, so looks like it should work in most browsers now. My bad about the webgpu part. Still, most people will be doing webgl2 builds / that's what the template is set up for.
hm ok. i either have to remove it or add a note then
renamed the itch.io page to bevy-quickstart
either is fine for me, but there are particle plugins that do work w/ webgl2
CPU-based?
yeah. https://github.com/mbrea-c/bevy_firework and I think there may even be 1-2 others.
You're such a busy wizard π
Wait, I know one that got featured in the Rust GameDev newsletter
okay. i think removing hanabi from the list will be the easiest option then
user can find it / others in bevy assets
I wouldn't. At that point, you're just creating yet-another-awesome-bevy-list
I mean, I guess that's an option. I'd pile my recommendations on if we did a ~~semi-~~totally-un-offical "crates the authors of this template love" list.
But I mostly wish we just had better tools for figuring out which crates listed in bevy-assets even work with the latest version of Bevy.
Or, we could do a totally neutral "this is a list of the most-used third party crates in Bevy Jam 4 that work with Bevy 0.14" perhaps.
That would be pretty in-spirit with the working group; I like that idea π
I'd be nice if it was sorted by something. Most recent supported Bevy version and then most GitHub stars would probably be a good start.
still 2 versions outdated, but for an "automated" approach that could be worse
also if you want to make a non-jam game, the most used libraries for jam games wouldn't necessarily match. but i'm sure in practice they'd all be useful so that's fine
Oh, found an issue: https://github.com/bevyengine/bevy-assets/issues/297
it's a bit tricky -- some link to github and some link to crates.io. some use gitlab, or something else entirely. some do dev in their main branch, following bevy's main branch, etc.
Ah, true, didn't consider that
and some aren't open source, or just don't link to their source
oh wait mb
misread context :)
If they're not open source, why would they be on that page? π
ah π
was thinking about jam games (to see which libraries they use)
a pr could probably make this more consistent by only linking to crates.io
or having two fields
one for crates.io link specifically, one for "where clicking this asset will take you"
which may or may not be the same link
ah, I forgot that workspaces are (or at least were) an issue too. lots of related discusson in https://github.com/bevyengine/bevy-website/pull/648
edit: seems resolved now, except for gitlab
maybe sort by recent downloads?
most likely supporting a more recent version of bevy => more recent downloads
although that also prioritizes more commonly used crates ofc. but that could be a feature?
@prisma delta jam page linking to the template should be updated, the template is now called bevy_quickstart at https://github.com/TheBevyFlock/bevy_quickstart
hm actually i see the jam page links to the CI template + bevy assets. ig that was changed?
Yep, we swapped that out.
That said, we should submit this template to Bevy Assets soon
And I've updated the initial message too π
yep agreed. there is an issue for this that was waiting on some more polish to get a nice screenshot / gif
unlikely to change in a meaningful way (before the jam) at this point so should be good to do that
cargo-cache upstream PR approved π
that'll fix the build matrix jobs for linux and web sharing the same cache, forcing one of them to always fully recompile
Awesome. @lunar sail do you want to merge and release this?
That would be great, thanks!
Can you tag it as both v2.0.0 and v2?
The action stuff is kinda weird in how they handle semver (its all fake)
Remind me tomorrow π
do you guys have a way to show a loading screen while stuff is still popping in?
while? if you're in a loading screen presumably you're not seeing stuff pop in right
there is a loading screen yes, i haven't personally reviewed that code though
I've done it now :)
@bitter sparrow @slender belfry We'll have to upgrade the action version to v2 to get these changes
should this particular template be able to deploy to github pages?
3
4
1
yes
will try to do a pass on the CI / CD doc and make a PR for that today
Approved. Will work on the readme stuff in like 2 hours until it is done. I wanna get the announcements out today π
hm, there's an issue with the cargo-cache change π
it's not fully broken, in fact it's still a slight improvement over the previous version as is so not too bad
the "workflow file hash" is an empty string, but the job name / matrix index are working properly
so the cache group looks like -test-0 instead of lkjsdfalsjkdf-test-0 or w/e
ok it's a silly mistake. good
I hope that does not require an immediate bump to v3. Idk how/if GitHub actions works with semver
i did a pass on the CI / CD doc
will make a PR for it (with broken images for now cause idk where to put the images π)
@unreal narwhal not sure if you wanted a ping for this part of the README but pinging just in case π
Something else unfortunately got me blocked. Iβll get to it as soon as I get up in like 8 hours. Sorry for the delay!
/docs/images?
maybe. this doesn't conflict with cargo doc right?
we'd also have to mention to delete docs/ when user clones the template, which is np
bevy repo does assets/branding and assets/docs, but i don't think we want that
i wonder if putting the "user files" in their own folder and having the top-level folder be the "template files" would make sense
where "user files" = assets, web, src, Cargo.toml, etc., and "template files" = README, LICENSE, doc images, etc.
not rn ofc :P
like how you can have multiple crates in a workspace
The versions are all fake, its just the name of the tags.
So you being able to use v1 instead of v1.0.0 is just the maintainers also assigning a v1 tag lol
so i suppose one could just move the v2 and v2.0.0 tags then
@prisma delta Are you able to release this?
I don't have access to my PC for the next three days.
You'd have to tag the latest commit v2.0.1 and move the v2 tag to the latest commit as well.
Thanks!
Brilliant π
Also I just realized a good way to test if this template works for Game Jams. At least I think.
https://github.com/Leafwing-Studios/cargo-cache/tags This should do it π
Cache your Cargo build files. Contribute to Leafwing-Studios/cargo-cache development by creating an account on GitHub.
confirmed that the fix works in bevy_quickstart release workflow
the windows build had a different workflow file hash than the linux, mac, and web builds o.O
files on windows being weird maybe? backslashes instead of slashes in file path?
if it's something like that it's probably fine, since we're splitting the cache key by ${{ runner.os }} already
Describe the bug hasFiles returns different hash in Windows than in Linux. This is problematic if you have a one job running on Windows and one job running on Linux and you need consistent hashes g...
known issue, "expected behavior" π ok then
the Cargo.toml and Cargo.lock file hashes are also different on windows vs other OSes ofc, and always have been
@bitter sparrow I'm ready for the readme stuff now πͺ
Approved all of your PRs. I have some suggestions, but I think it's best if I do them as a follow-up PR after merging them
Is it alright if I merge them all in?
yeah i was gonna do that
I β€οΈ automerge btw
real
@bitter sparrow I've got two small PRs ready for a review
@slender belfry small README pr: https://github.com/TheBevyFlock/bevy_quickstart/pull/152
afaic just needs readme images and it's good to announce
adding to bevy assets can happen after announce anyways
I'll do the following first:
- Do a pass on the readmes
- Write a little bit more on the structure
- Update
cargo-generate - Test that
cargo-generatestill works
ok sg. what's missing from structure, web/ and assets/?
trivial pr btw https://github.com/TheBevyFlock/bevy_quickstart/pull/153
I meant more in the sense of "design decisions"
whew, that was a tough merge
im gonna make a pr fixing the images / links in the docs
π
one issue, everything else seems fine
auto merge ci later
I'll make a pass through both workflows
links / images pr: https://github.com/TheBevyFlock/bevy_quickstart/pull/155
Alright, merge should be good now
checking
Approved, I assume you ran oxipng already
i know the gif compressor tool (gifski)
cargo binstall oxipng
ok ill update pr
TIL
ikr. it's crazy
gifs become actually a viable format
oxipng -o 6 -s docs/*.png good?
50% smaller is impressive
updated pr
oh my linux repo's package manager has oxipng. even better (so i can keep it updated)
You can also do -Z to squeezy a little bit more out, but it takes way longer
-o 6 is also -o max
If you add -Z also use --fast
IME is squeezes out some extra 3-4%, so not that much
Oof, thats good to know
there might be a way around it via a git setting described in the issue, i haven't explored
Although the OS is at the start of our cache key anyway IIRC so it shouldn't matter for the cache
yup in practice it's fine
I dont think cargo could reuse compilations from other OSs
probably not.. maybe it would help for like cargo doc? idk
you're unlikely to have more than 1 job that calls cargo doc anyways :p
The only contrived example I can come up with is using both a Linux machine and WSL and then wanting to cache that? But no one would do that.
BTW I also found this action which essentially does the same thing as ours: https://github.com/Swatinem/rust-cache
Offers some more customization, but it doesn't handle a missing Cargo.lock like we do
(Although I guess now cargo recommends not ignoring the lockfile even for libraries so that situation doesn't happen that often anymore)
Seems like that one does not check for changes in the workflow files by default
Lel 
Maybe some cross pollination between the two could help
They seem to put more care to reduce the size of the cache which we currently do not
I also like their save-if option, only saving caches on the main branch could be important for huge repos like bevy
it also caches the cargo registry which is nice
Btw @bitter sparrow I love the editing you did on the readme after the hackmd
Like the change in headers to "Write your game", "Run your game", "Release your game"
That's brilliant π
thank you thank you π
Also the way to compressed many of the instructions into single sentences
Like linking trunk serve directly to the trunk docs
that's pretty clever
π
yeah i wanted to make it look more approachable
while still including all the steps ofc
the collapsible <details> section suggested by @unreal narwhal is very cool, will have to use that more in my own readmes
inline html is an intentional feature of markdown :p
warning about it does make sense tho
I could add an mdlint file to root to allow it, but I think that would be fairly useless to a user, so I'm not doing that
btw i deleted the itch.io release packages and started a release workflow, to update the filenames
ah, thx!
im also double-checking caching behavior
the caches restored from the previous release's caches
because we changed Cargo.toml in between
(and they're using the same v0.1.0 tag so cache is shared)
so π
it's enabled by default, but there's a flag to disable it
confusingly, the windows cache actually did not restore, even though the cache keys seem fine to me
only the cargo.toml hash differs between the previous window release and the latest one
ignore me
i confused which release was latest and which was previous
trivial PR, fixing link that i missed: https://github.com/TheBevyFlock/bevy_quickstart/pull/156
also something i just noticed about itch.io
it's doing something with the filenames
on the github release they're all bevy_quickstart-os.zip format
downloaded from itch, the filenames are as shown in the screenshot
Already fixed in my working branch
Not merging so I don't get conflicts π
ah ok
i suspect that itch.io is converting _ to - on non-mac platforms
The heck
not sure how else to explain it
I'd like to only recommend Avian as a physics engine because I think people will have an easier time working with it. The main reasons for using Rapier are performance, maturity, and some extra advanced features like joint motors. I don't think these are important enough for jams to point beginners to it.
i can see that
it's not meant only for jams, and also performance can help with audio issues on web
(Speaking as someone who has banged their head against nphysics / rapier since 5 years)
no i agree
ive had my gripes with rapier
im kinda on the fence about only recommending avian
Yeah but you won't be running physics simulations that are complex enough to run into issues with Avian on Wasm builds anyways
they might not be complex but high-scale
Hmm, true
like last jam
Can I at least reorder them so that Avian is first?
no complaints there
@bitter sparrow https://github.com/TheBevyFlock/bevy_quickstart/pull/158
left some comments
i havent reviewed the last couple commits though
Agree with most of them. Fixing now.
Nah I appreciate it π
If I did not want this level of quality review, I would have submitted a bigger PR π
Addressed everything
i'll see about dependi rq
thx
i wonder if it's because i'm on vs code 1.89 instead of 1.91
i'll have to do a full system update to test that out though
Don't see anything about VSCode versions in https://github.com/filllabs/dependi/blob/main/package-lock.json
But idk how VSCode extensions work
Ah wait
"engines": {
"vscode": "^1.87.0"
}
So you should be seeing it π
Oh no, your nit bashed against my auto-merge π
ha. that's ok
interesting
@slender belfry https://github.com/TheBevyFlock/bevy_quickstart/pull/160
i like trying to make the near-oneliners fit on one line, but it's hard to tell until after changes are pushed :p
I like the split of "Pattern" / "Reasoning". I could make them just bold, but that's not really correct imo
i'm thinking maybe de-nest each pattern
How so?
so like ## Code structure
i'll be lenient on the PR adding the doc, as long as the high level makes sense
editing passes can optionally come after
also this doc is more of an optional read to use the template so it's not as critical
@bitter sparrow https://github.com/TheBevyFlock/bevy_quickstart/pull/162
I just remembered that that info had to also go somewhere
Note that I did not write about this being necessary on Windows since just last week someone found the issue in blake3 that caused recompiles and fixed it.
note + collapsible section is good
didn't know if that would render well
btw i'd be fine with announcing before the design doc is pushed
Is it alright to wait until that? I want to finish it in one go right now.
oh yeah for sure
sure i can go for that
Show off the splash, the menu and the ducky moving around
I want to use that GIF for the Bevy Assets page β¨
i should make a GIF of the CI / CD tbh /j
π
BTW I might try out the Jam template for TriJam if I get the time this weekend.
That'd be really really cool π
yea, will be great to see people try it out, maybe create issues / etc
I mean it's a 3 hour Game Jam hosted every weekend. Figured it would be a good test to see what else might be needed to get things up and running.
I'm in that intermediate area that I know enough to get in trouble, but there's still a lot I don't know.
You could also submit it to the assets page when you're ready. For the description, I was thinking of
A great way to get started on a new game β especially for a game jam!
Includes a basic project structure and automated deployments to itch.io.
Brought to you by the Bevy Jam working group.
So I figure I'm a good target to test the template out.
i'm not going to use this template for bevy jam 5 π
It's nice and short, so I'm not spending too much time, 3 hours would be enough that it would probably reveal any shortcomes pretty fast. Where-as a longer Game Jam would both take longer...and any issues could just be worked over. It's in some ways a Microcosm of Bevy Jam.
but only because i have my own template that uses 3rd-party libraries and has a few more opinionated features
but the core is otherwise very similar. in part because i updated it recently
I too have my own template, but:
- This needs some testing and a quick Game Jam to reveal some rough edges would be nice.
- I don't actually know that much about the way the template is structured so I would be coming in with just slightly more knowledge than someone who is new to the template. And I happen to be in a good position for that.
Also TriJam is very sensitive to rough issues in your pipeline and failure to plan properly. If I can get something semi-decent in 3 hours with Bevy, we can more refine things.
yeah that would be a great test case
But yeah. sorry for derailing the convo a bit
nah derail away
im mostly watching stuff compile
(FWIW I also know what third party crates I'll be using, more or less, in addition to what is provided, as I imagine many people using the template will reach for third party crates to fill in some things. I also only have 3 hours, so additional crates are kinda...necessary in some areas)
hopefully using no 3rd-party crates in the template means that users can easily slot in their own deps
We will find out.
raw footage
I thought it was going to be CI doing stuff lol
oh no i forgot to record 15 minutes of CI for the timelapse gif π©
Now you gotta do it again π© π© 
:wearydrops: is crazy
Is that a good thing or?
You mean the CSS Throbber? /jk
never heard it called that lol
A throbber, also known as a loading icon, is an animated graphical control element used to show that a computer program is performing an action in the background (such as downloading content, conducting intensive calculations or communicating with an external device). In contrast to a progress bar, a throbber does not indicate how much of the ac...
π
But yeah.
Idk why, but that term makes me feel uncomfortable π
Someone should probably sanity check that cargo-generate is still working
I will remove the link if you would like.
No no, it's alright π
Alright. Just wanted to offer, if you change your mind let me know.
I'll also spoiler the other usages of the word as well if you really want.
Nooooo it's not serious, no worries
I just never thought of an icon being able to "throb" instead of spin π
If you really want to do me a favor, you could check cargo generate π
Sure. Let me clean up some harddrive space really quick
Cargo generate seems to be fine. There's a small issue with the itchio stuff where it gets formatted like:
#Comment this out to disable.# ITCH_TARGET: your-itch-username/shaboingus# The organization or author that owns the rights to the game.
OWNER:
But that should be fixable.
the gif
Beyond that little issue, cargo-generate works!
Thanks for checking π
I just tried to fix it, but the tool keeps insisting on cutting away the newlines
heck!
It's the {%- something %-} specifically the dashes.
Oooooh
So if I do this
{% if itch_username != "" %}
It should work?
Yes, but it would leave a blank line in its place
The dash before removes whitespace before that, and the dash after does the same but after it. So if you remove the last dash I think the newline issue should be resolved.
Thanks, fixed it now
But I also get USE_GIT_LFS: falsejobs: due to the same kind of issue π
fixing that as well
will need to abbreviate this to fit, so stripping it down to the first sentence, and maybe one extra little thing
was thinking to somehow mention that it's pure bevy, since that's the main differentiator
fixed
Good idea
description = "A great way to get started on a new game β especially for a game jam! Start simple with a pure Bevy template."
I do miss the part about the project structure and itch deployments
i would agree but every template description says that at this point π
windows, mac, linux, web releases
well 4/6 of them
Fair enough
I guess my issue is that this doesn't communicate why this is more aligned with Bevy than "Random Bevy template #4928"
open to better wording, created pr: https://github.com/bevyengine/bevy-assets/pull/463
Start simple with a semi-official pure Bevy template.
?
a lil wordy
better
Does this fit?
"by the Bevy Jam working group"
At the end
ill check how that looks
We could also drop the entire "A great way to get started on a new game β especially for a game jam!"
"Start simple with a pure Bevy template brought to you by the Bevy Jam working group"
yeah we should drop this
that's better!
hooray
I'll also need to change the description of Foxtrot to point people towards bevy_quickstart
Don't want people to try using Foxtrot for jams π
I gave up on maintaining Wasm compatibility, and that is a must-have for jams
ahh, ok
I'm reordering widgets.rs to have pub stuff on top in this PR as that is what the design doc talks about
(And what a reader would care about)
hm, okay. personally i would also move trait Spawn into its own file, util/patch.rs, if this was my template
but we only have one patch
imo we should focus on getting w/e we need through that's blocking announcement
and more work can be done after the announcement / whenever ofc
but i'd quick approve this ofc
agreed, that's why I'm just reordering code in the same file
what is patch?
extension traits that "patch" missing functionality in upstream bevy
Ah, that makes sense
usually with a link to an issue
@hidden pond would you mind going through cargo generate again but this time do the whole shebang, including releasing onto itch.io?
Would be very embarassing if that didn't work π
If you don't have time, I (or @bitter sparrow if they want) can do it as well
I do not have an itch.io account, otherwise I would.
Mind creating one?
Sure, I'll have to do it tomorrow though as I'm about to head off to bed.
If that's alright.
No worries! Since we want to do the announcement soon, we'll just do it ourselves then π
Sleep well!
@bitter sparrow do you have time?
Sorry about not being able to help.
Hey, you showed me what those magic - did in the template
I would have no idea how google for my issue, haha
so you saved me a ton of time β€οΈ
maybe. i'm trying to figure out Dependi π‘
π
I can ship something to itch if y'all need someone to
My pleasure. I actually tried to figure out why there was - in the commands when I was working on the template initially and it came up blank. I only found it when looking up how to remove random blank lines from generated files lol.
Ooooh yes please!
current cargo generate instructions results in you being on a branch called cargo-generate
maybe a git checkout -b main in the readme?
I haven't looked at what branches the ci scripts are checking yet
Oooh right, that's an open issue on cargo-generate, I remember
Yep, I'll add that
They all expect main
git switch -c main is the slightly more modern variant FYI π
I have forgotten about switch since the day I read those release notes lmao
Good catch!
need to get that into my fingers
I also did, and then I got repeatedly shamed for it at my past workplace π
It worked though, haha
yeah thanks for the reminder
https://github.com/bevyengine/bevy-website/pull/1561 would give the template a small boost in the asset rankings π
Would you look at that, you've just received a totally impartial, unbiased approval
gonna make my crates brokenly depend on bevy 0.15 to take the top spot on the list
Oh wow @bitter sparrow look at https://bevyengine.org/assets/#templates
It looks like the ducky is walking on the black bar added by the design π
for me it's a bit higher
me updating the spawn code to place ducky lower and re-recording gif:
ok i see the pr is actually sorting "semver-compatible" followed by "non-semver-compatible" so i can't cheese the system like this
you must provide a stand-in html file when choosing html on itch
I believe anyway. I'm going to run the release script soon to see if it fails
the behavior i'm expecting:
- this error message exists before and after running the release workflow
- have to go into project settings and mark the web build as "run in web" or w/e, and then it'll work
I'll add that documentation to my current PR as well as soon as @unreal narwhal reports back
Nearly done with the design doc btw
this specifically
will let y'all know when it finishes π
i'm predicting 15 minutes for a fresh windows build heh
yea tests run on ubuntu-latest
Design doc is ready: https://github.com/TheBevyFlock/bevy_quickstart/pull/161
Pretty happy with how it turned out π
Let me know if I missed something important
will review
I will, in order to dogfood how easy it is to plug in third-party crates in practice
it feels easy to miss things in the github release section
Could you elaborate?
if you use cargo-generate you get to skip sections, and while I was doing the post-setup I had to find the "do a release" section in the main readme, then read through all of the release md skipping sections.
things like "the version must be a git tag'd commit that starts with v" are easy to miss
Oh, good point!
isn't cargo-generate the main flow? think i'm missing something
I was about to tag it without a v before I went to check because I've hit that problem in past lives
for the workflows.md file, it lists out all the variables that have already been set by cargo generate, for example
ah okay. that's true. i had the thought that maybe cargo generate shouldn't set any of those variables
if it's not going to set all of them and you have to look in that file anyways
If you name your itch.io project the same as your crate, you don't need to look in there
Yeah I see how that gives the vibe of "This is irrelevant for you, you already have these, go skip"
hm let me take a look at the release.yaml template rq
hm okay. technically the values we're populating it with may not be the values you want
like your itch.io username may not be what you want to set for OWNER
they're all reasonable defaults though
(that's only some silly macOS requirement anyways)
BINARY: {{crate_name}} is fine unless you use a src/bin/ folder, which the template does not by default, so that's fine
probably keeping the templating in there is reasonable. so docs should be clearer
yeah imo changing anything here is off the happy path
we could put this section last? ;P
yeah, any additional info can be deprioritized. details/summary or "at the end" or whatever
it also looks scary
its a lot of detail reading
On it
@bitter sparrow @unreal narwhal https://github.com/TheBevyFlock/bevy_quickstart/pull/163
Added a bit of wording as well to make sure people don't name their project bevy-jam-5 and their itch page bubble-popper-deluxe and then get confused why it does not work
Ideally, I'd like to add an option to cargo-generate
Enter your itch.io page name. Leave blank to use the project name
That would only show up when you have also added an itch.io user name
But I'm too tired to implement that rn π
maacos failure
yeah, in an ideal world there's plenty of time and energy lol
mm
so the source file of the mv doesn't exist ig
what's your crate name?
hm i wonder if the binary has hyphens instead of underscores
doing a local build now to find out
would be whacky
Thanks for helping out @unreal narwhal π
it's possible the conversion from - to _ only applies to the crate name so it can be used as a dependency as use bevy_flock::etc
happy to help
but not the binary name because there's technically no need
we set name = "{{project-name}} in Cargo.toml, so setting BINARY: {{project-name}} would be an easy fix if that's how it works
i think this should work as a fix, i'll make a PR
@unreal narwhal could you manually set the BINARY variable in the release workflow to the kebab-cased-version and re-run the release?
yep, doing it now
@hidden pond this might be something for you when you get up, if you feel like it: https://github.com/TheBevyFlock/bevy_quickstart/issues/164
macos build succeeded. just waiting on windows
heck yeah!
will I have to do this for every zip that's uploaded?
just the web one
oh ic what you mean
no, just one time ever
yes, every web zip that's uploaded
it uses itch.io "channels" so itch knows that the next release to the "web" channel is a new version of the same thing
Nice!
fantastic. so should be good once the final README pr is through explaining the missing steps
mac-native working too via download
This makes me happy π
would be nice to do signing and stuff for the next jam if it can be figured out
@unreal narwhal does this explain stuff better? https://github.com/TheBevyFlock/bevy_quickstart/pull/163
right now you have to do the right-click dance
what do you mean by that?
and signing
(i use linux)
ah okay
The classic macOS experience
tbf it is malware
macos throws supa-scurry warnings when you try to open non-signed dmgs these days
so you have to right-click to bypass it
Yeah it's sketchy how we ask for microphone permissions π
sketchier when no permissions are asked for
I imagine downloading random unsigned binaries is part of why wasm gets way more play in jams
No other way to get access to the microphone on macOS
@pyrious, @miniaczq, @chanku and I of the Bevy Jam working group have put together a nice little template for the jam: https://github.com/TheBevyFlock/bevy_quickstart
It includes a basic project structure and automated itch.io uploads. You can play it at https://the-bevy-flock.itch.io/bevy-quickstart
If you use it, let us know how it went! Feel free to open issues about anything that is unclear or missing.
Hello Bevy Jammers!
We at the #1258521739395203174 working group have finished preparing the Bevy Quickstart template for the upcoming jam. It's a simple pure-Bevy template so you can add your own 3rd-party dependencies on top.
i was typing π‘ /j
Huehuehue
when are y'all thinking of making the announcement?
Literally now
cool
@unreal narwhal give us your impartial opinion on which text to use
yeah timezone may not be great, but people should have time to check it out before the jam
I'll update the cargo-generate branch in the meantime
oh my msg is unfinished also π
Also: need to add the cool gif to the announcement
I would avoid value-judgements or words like "simple", "easy", etc
calling something simple and then a newer dev finding it not so simple will make that dev feel things that are not necessarily positive things
That's fair
Hello Bevy Jammers!
@hidden pond, @agile dirge, @slender belfry and @bitter sparrow of the #1258521739395203174 working group have finished preparing the Bevy Quickstart template for the upcoming jam. It includes automated itch.io uploads (https://the-bevy-flock.itch.io/bevy-quickstart) and a project structure that we think will start you off on the right foot; plus it only depends on Bevy, so you can add your own 3rd-party dependencies on top.
If you try it out, let us know how it goes! Feel free to open an issue for anything that's unclear or missing.
You're saying "try it out" twice
true
o whoops
And I think pinging the others for credit is nice π
But lgtm otherwise
@unreal narwhal would you replace "basic" with "guiding" or something like that?
the bevy assets description says "simple" still
mmm
it says "start simple with ..."
so maybe that's fine
not "this is a simple template"
ig still not great
"starting point" or "bevy-only" kind of wording
"basic" as in "basic project structure"
It includes a suggested project structure
That sounds like there is no structure, only the suggestion of one π
"An example project structure"?
it includes a project structure that we think will start you off on the right foot
that's great!
-Start simple with a pure-Bevy template made by the Bevy Jam working group!
+Start fast with a pure-Bevy template made by the Bevy Jam working group!
Is a "Bevy Jammer" something like a cell phone jammer for Bevy? 
yes
"an initial project structure" good too?
@unreal narwhal's suggestion sounds more human imo
πΈ
π
from the perspective of "avoiding easy/basic/etc" yeah
Also post it to #crates, templates do get posted there even though they're technically not a crate
did the dashes fix make it into the repo? I wasn't watching closely enough
yep π
sweet
Pretty much 20 seconds after you reported them, hehe
edited
er typo
Hello Bevy Jammers!
@chanku, @miniaczq, @janhohenheim and @pyrious of the #"0.14 Bevy Jam" working group have finished preparing the [Bevy Quickstart](https://github.com/TheBevyFlock/bevy_quickstart) template for the upcoming jam.
It includes automated itch.io uploads (https://the-bevy-flock.itch.io/bevy-quickstart) and a project structure that we think will start you off on the right foot; plus it only depends on Bevy, so you can add your own 3rd-party dependencies on top.
If you try it out, let us know how it goes! Feel free to open an issue for anything that's unclear or missing.
ah yea
