Comments on: Apparat is now Open Source

By: Sankar

Sankar — Tue, 25 Jan 2011 22:46:06 +0000

Hi Joa,

Is there any issues running apparat on style swfs when using the LZW option ?

When I run it without LZW, i.e.reducer -i $file -o $file i get a compression of 1.74 % and the swf loads up fine

with the -l true option applied on the reducer, the swc fails to load with the following errors

time 17:06:14.260 :: app.core.startup.LoadStylesheet :: Error loading compiled stylesheet: url=main.swf?version=0.1595955528318882, error=SWF is not a loadable module
setTitle(Object { name=”title”})history.js (line 482)
[Break on this error] backStack[backStack.length - 1].title = title;\n
history.js (line 482)
time 17:06:14.276 :: com.adobe.ac.bootstrap.AbstractWorkItem :: work item failed: name=Loading Application Styles
(?)() (line 54)
(?)() (line 12)
[Break on this error] \n
(line 54)
time 17:06:14.291 :: com.adobe.ac.bootstrap.AbstractWorkItem :: work item failed: name=app Bootstrap
(?)() (line 54)
(?)() (line 12)
[Break on this error] \n
(line 54)
time 17:06:14.291 :: com.adobe.ac.bootstrap.AbstractWorkItem :: work item failed: name=Initializing app Core
(?)() (line 54)
(?)() (line 12)
[Break on this error] \n
(line 54)
time 17:06:14.307 :: app.core.startup.LoadStylesheet :: Error loading compiled stylesheet: url=main.swf?version=0.1595955528318882, error=Unable to load style(SWF is not a loadable module): main.swf?version=0.1595955528318882.
(?)() (line 54)
(?)()

By: joa

joa — Sun, 16 Jan 2011 13:16:30 +0000

No, since the trigonometric methods and square root are just approximations.

Your physics engine will probably fail when using those methods because of the chaos theory. That means the small error will accumulate and your running simulation fails.

This means it is okay to update normals using FastMath.rsqrt to render a frame for instance, but anything that involves “x += FastMath.rsqrt(…)” will fail.

By: pallzoltan

pallzoltan — Thu, 13 Jan 2011 10:28:44 +0000

Joa, i was writing a racing game where I’ve used box2d and thought it would be a good idea to replace all Math functions with FastMatch and then put the file through your Apparat. It did help a bit, and I was glad about it, but then it ruined the physics engine…
So my question is if the FastMath is 100% compatible with the native AS3 Math.
Great stuff otherwise :)

By: Agile Chronicles #12: Technical Debt | Flex and Flash Developer – Jesse Warden dot Kizz-ohm

Fri, 30 Jul 2010 15:47:55 +0000

[...] C. We have certain limitations (no threads, a VM vs. all code being machine code, and a lot of missing optimizations). Brandon Hall once said if you have a performance problem in Flash, you’re probably doing [...]

By: Interview with Jesse Freeman – JXLTV Episode 9 | Flex and Flash Developer – Jesse Warden dot Kizz-ohm

Mon, 19 Jul 2010 15:20:13 +0000

[...] Apparat [...]

By: [Flash CS4] - SWFs komprimieren - Flashforum

[Flash CS4] - SWFs komprimieren - Flashforum — Fri, 22 Jan 2010 00:35:11 +0000

[...] [...]

By: Jonas

Jonas — Tue, 18 Aug 2009 16:04:57 +0000

I can confirm that it does not keep the order even in flat structures, and I can’t see any pattern either.

For my purposes, I guess I could identify them by checking pixel size and/or pixels with the list of files I’m sending to my tool, but it’s not a pretty solution.

By: joa

joa — Tue, 18 Aug 2009 15:13:08 +0000

Sorry Jonas, I have no idea about the order. Especially not when you think about other libraries that are embedded as well and they contain embeds too.

By: Jonas

Jonas — Tue, 18 Aug 2009 15:10:44 +0000

I feared as much. Do you know, though, if the mxmlc preserves the order in which the [embed]‘s were found when generating the swf? That would let me achieve the same result (for the “Embedder” tool I’m writing).

By: joa

joa — Tue, 18 Aug 2009 15:04:01 +0000

The original filename is lost after exporting a SWF. You could write a GUI that shows each picture and a quality slider right next to it. In that case it is no problem to have individual levels. But I do not see this option for the command line now.

By: Jonas

Jonas — Tue, 18 Aug 2009 14:53:16 +0000

For Reducer, it seems there is no way to set individual quality factors for different bitmaps – and maybe it’s a tricky task. When parsing the file, is there a way to find the original bitmap file’s name, so that one can match each DefineBitsLossless2Tag with an argument sent to the Reducer tool?
Thanks!

By: hristo

hristo — Thu, 13 Aug 2009 08:55:42 +0000

Exciting exciting exciting! Let’s see where this goes.
You really are becoming the Luke Skywalker of better as3/abc.

By: joa

joa — Wed, 12 Aug 2009 12:23:24 +0000

quarkus: Sorry but WordPress strips all XML tags. For the Ant Task I was thinking about a Java solution.

By: quarkus

quarkus — Wed, 12 Aug 2009 10:51:48 +0000

Here is a simple ANt-Task to call Reducer:

Could be configured with defined properties.
Thanks for this great tool !

By: TurboDieselSportInjection at blog.joa-ebert.com – Blog of Joa Ebert

TurboDieselSportInjection at blog.joa-ebert.com – Blog of Joa Ebert — Tue, 11 Aug 2009 23:37:05 +0000

[...] Update: TDSI is now open source! [...]

By: Reducer at blog.joa-ebert.com – Blog of Joa Ebert

Reducer at blog.joa-ebert.com – Blog of Joa Ebert — Tue, 11 Aug 2009 23:35:59 +0000

[...] Update: Reducer is now open source! [...]