Sburb (Formerly Jterniabound)

[kwinse]

Try to break stairs

Done and done.

The mute button takes a little bit to register, but the music change is instant (dunno if that's known).

Everything else seems to be working cleanly. I miss the sound effect on the chests though.

Browsers: Firefox 12.0, Chrome 19.0.1084.52 m

Firefox bug: At the top of the room, if you walk up to the wall then left to the stairs some characters don't roll onto the stairs (for example Karkat ge, Terezi doesn't)

[Mme Sparklecakes XVII]

Good catch on the stairs. Seems you found a nice geometric flaw in my impact detection. I only check the "corners" of the sprite's hit box, and that corner of the stairs can pierce pretty deep into the sprite before it actually notices. I guess I could sample some more points along the box (pre-edit: I did). At very least when this happens you don't ever actually get trapped. Just an extra walkable region. I added an extra corner to these acute angles so they're gone from |/ to |_/. Means you can't roll onto them as easily, but at least you don't get the ugly visual dissonance. Rolling along the stairs by walking down isn't a big deal.

Getting stuck in firefox: Fucking weird! I guess this is a subtle difference in the firefox and chrome algorithms for "point in polygon" checks. Made a small tweak to the polys so it doesn't happen (in this one instance).

I also slowed down the stair velocities so now you get more of a feeling that you're actually on stairs. (An authentic stair experience is quintessential) Weird side-effect: You can walk down stairs faster by running into the top/bottom now. Might fix that later.

The mute button: I think it's just because it's "on click". It doesn't go when you press, only when you've pressed and released. Is this what you're experiencing, or is it something more? (Also visual-audio timing has some cognitive dissonances involved).

Edit: oops! Left on a debug flag! fixed.

[kwinse]

There is a noticeable delay after you click+release before the volume change takes effect. I'd estimate 2 seconds for Firefox, 1 for Chrome, and instant in flash-terniabound.

[Mme Sparklecakes XVII]

There is a noticeable delay after you click+release before the volume change takes effect. I'd estimate 2 seconds for Firefox, 1 for Chrome, and instant in flash-terniabound.

Oh... I'm confused now. You seemed to be describing the opposite effect before.

This issue, I'm afraid I might not be able to resolve. The Audio tag is still pretty experimental as far as I can tell. It seems that the only way to currently change the volume of a running sound is to call "song.volume = blah", and since that's not a function call, the browsers media playback would likely only notice if it checked. I'm afraid this is not a resolvable problem on my end, as far as I can tell. Hopefully as HTML5 standards mature this will improve.

Although, I suppose in strict theory, I could completely destroy the current song and start a new one where it was with the changed volume. That would certainly have immediate sound change ramifications. Although I suspect then you'd just get a skip in the song.

EDIT:

this.asset.volume = globalVolume;
this.asset.pause();
this.asset.play();

Dirty hack is dirty, but it works. Seem good now?

[kwinse]

Oh, whoops, my bad. Glad to know it's HTML5 being all new though.

[Mme Sparklecakes XVII]

Yeah this shit can be really frustrating. Every time I look at this table I get really pissed off: http://www.w3schools.com/html5/html5_audio.asp

Also, I ninja-edited in that I provided a dirty fix to the problem. Seem good now?

[kwinse]

No difference in Firefox (actually may be slower to change the volume) (and yes I cleared cache), instant in Chrome but you can notice the pause and play.

[Mme Sparklecakes XVII]

In your opinion, which is the better option of the two?

[kwinse]

Probably the version I see in Chrome, instant but with a noticeable pause in the music. Really it's not a big deal either way.

[Mme Sparklecakes XVII]

Made SBURBML much smarter, and moved some stuff around so even less needs to be hardcoded (even the HUD button behaviour is defined in SBURBML now).

Optimized the rendering so it should be a bit peppier in Firefox.

I also wrote down pretty much everything you need to know to start deploying fully functioning walkabouts on the wiki here. Literally you can just hit the big "save" button in the test page, edit the mark up there, and hit load and it should load a valid state. Interested to see if anyone can figure out how to do it all from my wiki notes. Would also appreciate any criticisms of the notes.

[Mme Sparklecakes XVII]

After hitting myself over the head several times, I have now made it so EVERY tag can be templated. So if you have a lot of actions that look like

<Action command="talk" name="suck it!!!" noWait="true">@! You suck!!!</Action>

You can now do this:

<Action class="suckAction" name="suck it!!" command="talk" noWait="true">@! fuck you!!!</Action>
...
<Action class="suckAction"></Action>
<Action class="suckAction" silent="true"></Action>
<Action class="suckAction" name="suck it x2 combo!!!"></Action>

And any fields you didn't re-specify should be filed in!

[ashdenej]

This is pretty awesome. In fact, it's very awesome. How much of your future workload in walkarounds is reduced by this engine, by the way?

And how swappable-outable are the backgrounds, out of interest?

[Mme Sparklecakes XVII]

There was a MASSIVE overhead of just assembling and aligning, and wiring assets. That should be almost non-existent now.

How swappable are backgrounds?

Go to the demo, hit save, and find this section in the output:
<Sprite name="compLabBG">
<Animation sheet="compLabBG"/>
</Sprite>

replace with

<Sprite name="compLabBG">
<Animation sheet="stairRoomBG"/>
</Sprite>

and press load. Boom.

[OrangeAipom]

You need to FLIP Vriska's talksprite TURNWAYS!

You already knew that though.

Also, when she first talks, she swoops in from the wrong direction.

[Mme Sparklecakes XVII]

I actually really wasn't sure which way she's supposed to face. Flip'd.

She is supposed to come in from that side, though. That's always how it's been (in my head, at very least?). The first dialog has to be MEGA DRAMATIC, so everyone knows this shit is SUPER SERIOUS.

[Mme Sparklecakes XVII]

I'm having a bit of trouble with the bottleneck of rendering. Running function run time tests, Animation.draw() seems to be the biggest glutton, and its impacting performance a bit on firefox (and certainly making me question scalability). I've made it so now only onscreen pixels/objects are rendered, but there's obviously still a lot of redundant rendering going on (every on screen pixel of every object gets re-rendered always, even it hasn't changed or gets covered by another pixel).

Unfortunately I'm at a loss for how to properly optimize for these sorts of things. Is there anyone here with experience optimizing graphics?

[Mme Sparklecakes XVII]

Started work on Strife Mode support. Head on over to the demo page and press "strife test" to see it in action.

[ashdenej]

Started work on Strife Mode support. Head on over to the demo page and press "strife test" to see it in action.

When I swapped, first note of the JUST1C3 theme played and then it paused for three or four seconds before resuming. Perhaps put a pause on the music until it loads sufficiently?

Dave has nice glide, mm. He's not doing the "charge" attack, but that's prolly because you haven't put it in yet. Seems to be all good!

Although he is a little small??? And there's no message saying that there's nothing downwards of the lab, so that might confuse people if you made edgeless arenas I guess.

Aaand once the theme has finished, it doesn't loop. ...Until you mess with the volume controls.

The "whatup" feature works, I didn't realise one could for a long time.

[Mme Sparklecakes XVII]

The sounds are all preloaded. I think the delay might just be some silence at the start of the song file. I'm starting it at 0, normally it gets started a few seconds in. Gonna fix that in a later update.

Yep, no charge attack yet.

He is definitely too small! I forget to scale up all his assets and then kinda went... "fuck it". Will fix... eventually (it's literally scaling 6 images 2x).

Removed the message just to minimize the the SBURB file. I could add it back. It was really only there as a demonstration of "triggers" though. I didn't realize people really thought it was important!

The doesn't loop thing is worrying though. I will have to look into that (I think I have an idea of what it might be).

[Mme Sparklecakes XVII]

Sound issues fixed (may need to refresh cache).

For anyone who does refresh their cache, can anyone reproduce this:
Clear cache, refresh page, wait for first set of files to load, hit "strife test": hangs indefinitely on 13/14.

This only seem to be happening to me in chrome, not in firefox. No errors are being reported in either.

Also just noticed if you can hit "strife test" before the first set has loaded the asset manager FLIPS OUT. Gonna go ahead and call this one "non-standard use".

[Frost]

You probably know this already, but physics are a little off in Strife Mode. Walk directly north against the desk thing, and the camera "shivers" rather than staying in place.

[Mme Sparklecakes XVII]

I hadn't noticed this, actually. However it's not surprising, I'm doing a lot of approximations for the impact resolution. I'm going to be setting the camera to "pan" to current location rather than snapping in a later version. Hopefully this will resolve the issue.

[kwinse]

For anyone who does refresh their cache, can anyone reproduce this:
Clear cache, refresh page, wait for first set of files to load, hit "strife test": hangs indefinitely on 13/14.

First try, it hung on 18/14 in Firefox, but I don't know where in the loading it was when I clicked (kinda hard to time it when stuff loads quick). Second try, 27/14.

Yeah it's probably best to call this non-standard use.

EDIT: Or am I testing the bug you described later in the post? I probably got confused somewhere.

In any case, it seems to load fine in Chrome, excepting the first time loading where it hung just loading stuff, but I haven't been able to reproduce it.

[ashdenej]

For anyone who does refresh their cache, can anyone reproduce this:

I got 17/14

Thanks for the fixes!

[Mme Sparklecakes XVII]

Yeah I was more interested in testing when everything's properly loaded, and then hitting strife. So you see karkat and shit, and then try. It reproduces 100% of the time for me on Chrome.

I set up some debug calls and it seems it's an issue with how Chrome handles readyState and loadeddata. Must be something like the file hasn't exactly cached yet but is sort of cached. I'm gonna pretend it's not really a problem, since it's an issue with re-initializing assets. At worst I could just make the system a bit smarter and just not do that.

Edit: fixed the fraction glitch, didn't fix the other one.