early jeash test

classic Classic list List threaded Threaded
16 messages Options
Reply | Threaded
Open this post in threaded view
|

early jeash test

Tony Polinelli
Hi guys, 

Thought that i'd share a little test using jeash. 


It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time 

Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)

If people want to test the code more pm me. 

cheers
Tony




--
Tony Polinelli
http://touchmypixel.com

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Tony Polinelli
Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
mobile replacement!

Fortunately haxe obviously compiles nativly too tho ;)

On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:

> Hi guys,
> Thought that i'd share a little test using jeash.
> http://touchmypixel.com/test/catch/
>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
>
> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
>
> If people want to test the code more pm me.
> cheersTony
>
>
>
> --
> Tony Polinelli
> http://touchmypixel.com
>
>

--
Tony Polinelli
http://touchmypixel.com

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

alexander.konotop
In reply to this post by Tony Polinelli
В Sat, 30 Apr 2011 16:20:43 +1000
Tony Polinelli <[hidden email]> пишет:

> Hi guys,
>
> Thought that i'd share a little test using jeash.
>
> http://touchmypixel.com/test/catch/
>
> <http://touchmypixel.com/test/catch/>It runs really well in chrome. It
> starts off well - but as soon as collisions occur, the framerate
> drops, and then never recovers. I am thinking that it is probably
> gc'ing not working correctly? I'll have to poke around a little when
> i get time
>
> Definately shows great promise for the jeash project, as i didnt have
> to put in too much work to get this little test to work. It shows
> that there are still a few bugs, but on a whole, *nearly* works as
> expected! great work Niel (i think)
>
> If people want to test the code more pm me.
>
> cheers
> Tony
>
>
>
>

Wow! Adobe RIP? :-)

P.S. It's VERY smooth in FF4. Opera = Chromium (fast enought but FF4 is
twice faster), in Midori (another webkit browser) FPS is jumping
constantly (actually it's normal for JS in Midori). So I don't think
that it's a gc/haxe issue. I think it's just a browser particularity.

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Tarwin Stroh-Spijer
I think the jumping framerate is normal inmost browsers, at least to a
point. Browsers seem to give JS a pretty low priorty, and so if
anything else happens it simply slows things down. Alas I don't think
there's any good fixes for this ATM but here's hoping (yay- HTML
games, totally).

The maybe GC issue is something different though. It could be worth
checking the inner workings of some of the multitude of current canvas
game engines to see how they handle this.

Does anyone know a good way of checking JS speed in real apps versus
simple loop checks like Sunspider?

On Saturday, April 30, 2011, Alexander Konotop
<[hidden email]> wrote:

> В Sat, 30 Apr 2011 16:20:43 +1000
> Tony Polinelli <[hidden email]> пишет:
>
>> Hi guys,
>>
>> Thought that i'd share a little test using jeash.
>>
>> http://touchmypixel.com/test/catch/
>>
>> <http://touchmypixel.com/test/catch/>It runs really well in chrome. It
>> starts off well - but as soon as collisions occur, the framerate
>> drops, and then never recovers. I am thinking that it is probably
>> gc'ing not working correctly? I'll have to poke around a little when
>> i get time
>>
>> Definately shows great promise for the jeash project, as i didnt have
>> to put in too much work to get this little test to work. It shows
>> that there are still a few bugs, but on a whole, *nearly* works as
>> expected! great work Niel (i think)
>>
>> If people want to test the code more pm me.
>>
>> cheers
>> Tony
>>
>>
>>
>>
>
> Wow! Adobe RIP? :-)
>
> P.S. It's VERY smooth in FF4. Opera = Chromium (fast enought but FF4 is
> twice faster), in Midori (another webkit browser) FPS is jumping
> constantly (actually it's normal for JS in Midori). So I don't think
> that it's a gc/haxe issue. I think it's just a browser particularity.
>
> --
> haXe - an open source web programming language
> http://haxe.org
>

--


Tarwin Stroh-Spijer
_______________________

Touch My Pixel
http://www.touchmypixel.com/
phone: +61 3 8060 5321
_______________________

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Niel Drummond-3
In reply to this post by Tony Polinelli
Thanks for the heads up, nice to see some game demos appearing in the
haxe-js world - I have been investigating how to run physics in a
separate thread with box2d when I get a free moment, which is the
bottleneck here and would give a significant performance boost.

I guess the drop in framerate comes when more than a few objects are
added to the physics engine - maybe objects are not being properly
dereferenced when they are removed? Should be worth investigating..

I should give credit where it is due, some contributors supplied nice
patches recently: brett periscopic (author of the gtween port), also
tecteun. Zjnue also continues to provide a wealth of qualitative test
cases / bug reports, and has successfully used the library in a couple
of commercial projects (one at SXSW I think).

As you say, it's *almost* there.. still have a couple of bugs to fix :-)

@tarwin: chrome has a very extensive profiling suite, though it's no
good for microbenchmarks.

- Niel

On 04/30/11 09:06, Tony Polinelli wrote:

> Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> mobile replacement!
>
> Fortunately haxe obviously compiles nativly too tho ;)
>
> On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
>> Hi guys,
>> Thought that i'd share a little test using jeash.
>> http://touchmypixel.com/test/catch/
>>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
>>
>> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
>>
>> If people want to test the code more pm me.
>> cheersTony
>>
>>
>>
>> --
>> Tony Polinelli
>> http://touchmypixel.com
>>
>>


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

hosey hosey

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Tony Polinelli
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)





On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--
Tony Polinelli
http://touchmypixel.com

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Cauê W.
2011/4/30 Tony Polinelli <[hidden email]>
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

That really makes sense!
I was thinking about that for java/c# targets. haXe could have its own internal file format to describe its AST as externs in a fast and concise way. This way we could build modules that interact with each other. (without having to define the externs by hand), and while still keep the way we do that as cross-target as possible (e.g. not having to parse .js / .swf / .dll / .jar files)

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Niel Drummond-3
On 05/01/11 03:21, Cauê Waneck wrote:
2011/4/30 Tony Polinelli <[hidden email]>
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

What you mention was implemented by nathan as a fork of the haxe-2.06 compiler, but never integrated into haxe mainline. https://github.com/webr3/haxe

However, in my experience deployment is best kept in one file, because browsers limit concurrent javascript file downloads, and also block until they execute each javascript file. Code size can be reduced to about 10-15% of the original size, through

- employing deflate on the http server
- --dead-code-elimination (box2d may need some extra work here, jeash seems to be fine)
- using a minimizer (e.g. YUI, jsmin is also in haxelib)

- Niel


haxe -jslib jeash -jslib box2d

That really makes sense!
I was thinking about that for java/c# targets. haXe could have its own internal file format to describe its AST as externs in a fast and concise way. This way we could build modules that interact with each other. (without having to define the externs by hand), and while still keep the way we do that as cross-target as possible (e.g. not having to parse .js / .swf / .dll / .jar files)


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

hosey hosey
In reply to this post by Tony Polinelli
Has there been any best practices written on Jeash and SWF programming.  I use the display objects like candy (anything that has interaction is a new Display Object).   I am assuming we try to limit that as much as possible for Jeash given it makes a full sized canvas for each display object? Basically we should track the mouse position over the main canvas and use a "mouse layer" canvas for interactions.   I am making a Visio-esque application.

At this point I am looking at using CSS to control the layout and haxe/jquery to write manipulate it or using Jeash and pulling everything into the canvas.  This will be the fall back when flash is not available.

Also I am looking at CPP output for Android/Iphone/Playbook as end goals as well....


On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]> wrote:
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)






On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Tarwin Stroh-Spijer
Interesting. I'm trying to come up with some best practices at the moment, as I'm giving a talk at Adobe tomorrow night. I think Jeash (haxe being able to do canvas stuff) will be something really interesting to people and I'd like to give them the best info.

Any examples would be great !


Tarwin Stroh-Spijer
_______________________

Touch My Pixel
http://www.touchmypixel.com/
phone: +61 3 8060 5321
_______________________


On Tue, May 3, 2011 at 11:33 AM, hosey hosey <[hidden email]> wrote:
Has there been any best practices written on Jeash and SWF programming.  I use the display objects like candy (anything that has interaction is a new Display Object).   I am assuming we try to limit that as much as possible for Jeash given it makes a full sized canvas for each display object? Basically we should track the mouse position over the main canvas and use a "mouse layer" canvas for interactions.   I am making a Visio-esque application.

At this point I am looking at using CSS to control the layout and haxe/jquery to write manipulate it or using Jeash and pulling everything into the canvas.  This will be the fall back when flash is not available.

Also I am looking at CPP output for Android/Iphone/Playbook as end goals as well....


On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]> wrote:
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)






On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Tony Polinelli
I still think that it is too early for best practices and as the saying goes - dont prematurely optimize. Jeash is in a very early stage, and is aimed at replicating the flash API. I think that this is great for crossplatform functionality, and great to slow the power of haxe js. If you are wanting to dev for older js engines, you might want to optimize a lot, but then you mightn't want to use jeash. I think the focus should be on jeash being crossplatform, and allowing the developer to dev, like they were in flash - and then wait for the browsers to catch up - FF4 seems kinda quick - very promising in this respect. 




On Tue, May 3, 2011 at 11:49 AM, Tarwin Stroh-Spijer <[hidden email]> wrote:
Interesting. I'm trying to come up with some best practices at the moment, as I'm giving a talk at Adobe tomorrow night. I think Jeash (haxe being able to do canvas stuff) will be something really interesting to people and I'd like to give them the best info.

Any examples would be great !



Tarwin Stroh-Spijer
_______________________

Touch My Pixel
http://www.touchmypixel.com/
phone: <a href="tel:%2B61%203%208060%205321" value="+61380605321" target="_blank">+61 3 8060 5321
_______________________


On Tue, May 3, 2011 at 11:33 AM, hosey hosey <[hidden email]> wrote:
Has there been any best practices written on Jeash and SWF programming.  I use the display objects like candy (anything that has interaction is a new Display Object).   I am assuming we try to limit that as much as possible for Jeash given it makes a full sized canvas for each display object? Basically we should track the mouse position over the main canvas and use a "mouse layer" canvas for interactions.   I am making a Visio-esque application.

At this point I am looking at using CSS to control the layout and haxe/jquery to write manipulate it or using Jeash and pulling everything into the canvas.  This will be the fall back when flash is not available.

Also I am looking at CPP output for Android/Iphone/Playbook as end goals as well....


On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]> wrote:
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)






On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--
Tony Polinelli
http://touchmypixel.com

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

jlm@justinfront.net
The issue that keeps coming back to me on Jeash, is can we justify Canvas as a minimum for instance...


you can do a lot in HaXe javascript without Canvas.  I think experimenting with Jeash will inform our flash coding, and this is why I suggested we should consider improving flash so that the targets are closer and also faster with a single workflow but I think this is the future evolving that we have access to now but the workflows are still evolving, so very exciting times, but maybe on the current web this is still more for tomorrow than today.


On 3 May 2011, at 04:13, Tony Polinelli wrote:

I still think that it is too early for best practices and as the saying goes - dont prematurely optimize. Jeash is in a very early stage, and is aimed at replicating the flash API. I think that this is great for crossplatform functionality, and great to slow the power of haxe js. If you are wanting to dev for older js engines, you might want to optimize a lot, but then you mightn't want to use jeash. I think the focus should be on jeash being crossplatform, and allowing the developer to dev, like they were in flash - and then wait for the browsers to catch up - FF4 seems kinda quick - very promising in this respect. 




On Tue, May 3, 2011 at 11:49 AM, Tarwin Stroh-Spijer <[hidden email]> wrote:
Interesting. I'm trying to come up with some best practices at the moment, as I'm giving a talk at Adobe tomorrow night. I think Jeash (haxe being able to do canvas stuff) will be something really interesting to people and I'd like to give them the best info.

Any examples would be great !



Tarwin Stroh-Spijer
_______________________

Touch My Pixel
http://www.touchmypixel.com/
phone: <a href="tel:%2B61%203%208060%205321" value="+61380605321" target="_blank">+61 3 8060 5321
_______________________


On Tue, May 3, 2011 at 11:33 AM, hosey hosey <[hidden email]> wrote:
Has there been any best practices written on Jeash and SWF programming.  I use the display objects like candy (anything that has interaction is a new Display Object).   I am assuming we try to limit that as much as possible for Jeash given it makes a full sized canvas for each display object? Basically we should track the mouse position over the main canvas and use a "mouse layer" canvas for interactions.   I am making a Visio-esque application.

At this point I am looking at using CSS to control the layout and haxe/jquery to write manipulate it or using Jeash and pulling everything into the canvas.  This will be the fall back when flash is not available.

Also I am looking at CPP output for Android/Iphone/Playbook as end goals as well....


On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]> wrote:
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)






On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--
Tony Polinelli
http://touchmypixel.com
--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Niel Drummond-3
On 05/03/11 06:43, [hidden email] wrote:
The issue that keeps coming back to me on Jeash, is can we justify Canvas as a minimum for instance...


you can do a lot in HaXe javascript without Canvas.  I think experimenting with Jeash will inform our flash coding, and this is why I suggested we should consider improving flash so that the targets are closer and also faster with a single workflow but I think this is the future evolving that we have access to now but the workflows are still evolving, so very exciting times, but maybe on the current web this is still more for tomorrow than today.


Yes, it's true that Apple's CoreGraphics library is hindering the performance of canvas pixel manipulation compared to Google's Skia on Windows and Linux, though it's potentially just a matter of time that this is fixed.

Afterall, there isn't anything stopping you from blitting images per frame in jeash, caching the results, and swapping the displayobject just like in the game you posted.

I think it's going to be quite difficult convincing haxers and non-haxers alike that a non-flash API is the way forwards for a language with many associations to Actionscript, even if it sometimes makes sense to do so (sorry to be blunt), though I am happy to be proved wrong about this.

- Niel


On 3 May 2011, at 04:13, Tony Polinelli wrote:

I still think that it is too early for best practices and as the saying goes - dont prematurely optimize. Jeash is in a very early stage, and is aimed at replicating the flash API. I think that this is great for crossplatform functionality, and great to slow the power of haxe js. If you are wanting to dev for older js engines, you might want to optimize a lot, but then you mightn't want to use jeash. I think the focus should be on jeash being crossplatform, and allowing the developer to dev, like they were in flash - and then wait for the browsers to catch up - FF4 seems kinda quick - very promising in this respect. 




On Tue, May 3, 2011 at 11:49 AM, Tarwin Stroh-Spijer <[hidden email]> wrote:
Interesting. I'm trying to come up with some best practices at the moment, as I'm giving a talk at Adobe tomorrow night. I think Jeash (haxe being able to do canvas stuff) will be something really interesting to people and I'd like to give them the best info.

Any examples would be great !



Tarwin Stroh-Spijer
_______________________

Touch My Pixel
http://www.touchmypixel.com/
phone: <a moz-do-not-send="true" href="tel:%2B61%203%208060%205321" value="+61380605321" target="_blank">+61 3 8060 5321
_______________________


On Tue, May 3, 2011 at 11:33 AM, hosey hosey <[hidden email]> wrote:
Has there been any best practices written on Jeash and SWF programming.  I use the display objects like candy (anything that has interaction is a new Display Object).   I am assuming we try to limit that as much as possible for Jeash given it makes a full sized canvas for each display object? Basically we should track the mouse position over the main canvas and use a "mouse layer" canvas for interactions.   I am making a Visio-esque application.

At this point I am looking at using CSS to control the layout and haxe/jquery to write manipulate it or using Jeash and pulling everything into the canvas.  This will be the fall back when flash is not available.

Also I am looking at CPP output for Android/Iphone/Playbook as end goals as well....


On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]> wrote:
17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is definately enough to make a fullblown game in! flash game sans-flash!

pity they code is ~700kb in JS. (probably ~50 swf)  the js target really needs to be able to complie to modules, so you could have jeash.js and box2d.js cached by the browser. I was thinking this could be done by:

haxe -jslib jeash -jslib box2d

it would use those projects - for autocompletion etc, but not compile the code in. You would have a jeash.js already compiled out by itself. (kinda like an exclude package feature - there for dev, not for compile)






On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]> wrote:

17fps on notion ink adam

On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]> wrote:
>
> Thanks for the heads up, nice to see some game demos appearing in the
> haxe-js world - I have been investigating how to run physics in a
> separate thread with box2d when I get a free moment, which is the
> bottleneck here and would give a significant performance boost.
>
> I guess the drop in framerate comes when more than a few objects are
> added to the physics engine - maybe objects are not being properly
> dereferenced when they are removed? Should be worth investigating..
>
> I should give credit where it is due, some contributors supplied nice
> patches recently: brett periscopic (author of the gtween port), also
> tecteun. Zjnue also continues to provide a wealth of qualitative test
> cases / bug reports, and has successfully used the library in a couple
> of commercial projects (one at SXSW I think).
>
> As you say, it's *almost* there.. still have a couple of bugs to fix :-)
>
> @tarwin: chrome has a very extensive profiling suite, though it's no
> good for microbenchmarks.
>
> - Niel
>
> On 04/30/11 09:06, Tony Polinelli wrote:
> > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
> > mobile replacement!
> >
> > Fortunately haxe obviously compiles nativly too tho ;)
> >
> > On Saturday, April 30, 2011, Tony Polinelli <[hidden email]> wrote:
> >> Hi guys,
> >> Thought that i'd share a little test using jeash.
> >> http://touchmypixel.com/test/catch/
> >>  <http://touchmypixel.com/test/catch/>It runs really well in chrome. It starts off well - but as soon as collisions occur, the framerate drops, and then never recovers. I am thinking that it is probably gc'ing not working correctly? I'll have to poke around a little when i get time
> >>
> >> Definately shows great promise for the jeash project, as i didnt have to put in too much work to get this little test to work. It shows that there are still a few bugs, but on a whole, *nearly* works as expected! great work Niel (i think)
> >>
> >> If people want to test the code more pm me.
> >> cheersTony
> >>
> >>
> >>
> >> --
> >> Tony Polinelli
> >> http://touchmypixel.com
> >>
> >>
>
>
> --
> haXe - an open source web programming language
> http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--
Tony Polinelli
http://touchmypixel.com

--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org


--
haXe - an open source web programming language
http://haxe.org



--
Tony Polinelli
http://touchmypixel.com
--
haXe - an open source web programming language
http://haxe.org



--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Dion Whitehead Amago
The haxe Hydrax game engine has rendering pipelines to both canvas and
css, because yes, on iOs devices, css transforms are much, much
faster.

http://dionamago.net/?p=410&cpage=1#comment-390

I'll be posting an update soon that allows you to mix canvas and css
layers, and identical display of SVGs across flash, canvas, and css.

Dion

On Tue, May 3, 2011 at 1:37 AM, Niel Drummond
<[hidden email]> wrote:

> On 05/03/11 06:43, [hidden email] wrote:
>
> The issue that keeps coming back to me on Jeash, is can we justify Canvas as
> a minimum for instance...
> http://sebleedelisle.com/2011/04/html5javascript-platform-game-optimised-for-ipad/
> you can do a lot in HaXe javascript without Canvas.  I think experimenting
> with Jeash will inform our flash coding, and this is why I suggested we
> should consider improving flash so that the targets are closer and also
> faster with a single workflow but I think this is the future evolving that
> we have access to now but the workflows are still evolving, so very exciting
> times, but maybe on the current web this is still more for tomorrow than
> today.
>
> Yes, it's true that Apple's CoreGraphics library is hindering the
> performance of canvas pixel manipulation compared to Google's Skia on
> Windows and Linux, though it's potentially just a matter of time that this
> is fixed.
>
> Afterall, there isn't anything stopping you from blitting images per frame
> in jeash, caching the results, and swapping the displayobject just like in
> the game you posted.
>
> I think it's going to be quite difficult convincing haxers and non-haxers
> alike that a non-flash API is the way forwards for a language with many
> associations to Actionscript, even if it sometimes makes sense to do so
> (sorry to be blunt), though I am happy to be proved wrong about this.
>
> - Niel
>
>
> On 3 May 2011, at 04:13, Tony Polinelli wrote:
>
> I still think that it is too early for best practices and as the saying goes
> - dont prematurely optimize. Jeash is in a very early stage, and is aimed at
> replicating the flash API. I think that this is great for crossplatform
> functionality, and great to slow the power of haxe js. If you are wanting to
> dev for older js engines, you might want to optimize a lot, but then you
> mightn't want to use jeash. I think the focus should be on jeash being
> crossplatform, and allowing the developer to dev, like they were in flash -
> and then wait for the browsers to catch up - FF4 seems kinda quick - very
> promising in this respect.
>
>
>
> On Tue, May 3, 2011 at 11:49 AM, Tarwin Stroh-Spijer
> <[hidden email]> wrote:
>>
>> Interesting. I'm trying to come up with some best practices at the moment,
>> as I'm giving a talk at Adobe tomorrow night. I think Jeash (haxe being able
>> to do canvas stuff) will be something really interesting to people and I'd
>> like to give them the best info.
>> Any examples would be great !
>>
>>
>> Tarwin Stroh-Spijer
>> _______________________
>>
>> Touch My Pixel
>> http://www.touchmypixel.com/
>> phone: +61 3 8060 5321
>> _______________________
>>
>>
>> On Tue, May 3, 2011 at 11:33 AM, hosey hosey <[hidden email]> wrote:
>>>
>>> Has there been any best practices written on Jeash and SWF programming.
>>>  I use the display objects like candy (anything that has interaction is a
>>> new Display Object).   I am assuming we try to limit that as much as
>>> possible for Jeash given it makes a full sized canvas for each display
>>> object? Basically we should track the mouse position over the main canvas
>>> and use a "mouse layer" canvas for interactions.   I am making a Visio-esque
>>> application.
>>> At this point I am looking at using CSS to control the layout and
>>> haxe/jquery to write manipulate it or using Jeash and pulling everything
>>> into the canvas.  This will be the fall back when flash is not available.
>>> Also I am looking at CPP output for Android/Iphone/Playbook as end goals
>>> as well....
>>>
>>> On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli <[hidden email]>
>>> wrote:
>>>>
>>>> 17fps on a tablet is HUGE - also, i just d/l ff4 = amazing ~50, that is
>>>> definately enough to make a fullblown game in! flash game sans-flash!
>>>>
>>>> pity they code is ~700kb in JS. (probably ~50 swf)  the js target really
>>>> needs to be able to complie to modules, so you could have jeash.js and
>>>> box2d.js cached by the browser. I was thinking this could be done by:
>>>>
>>>> haxe -jslib jeash -jslib box2d
>>>>
>>>> it would use those projects - for autocompletion etc, but not compile
>>>> the code in. You would have a jeash.js already compiled out by itself.
>>>> (kinda like an exclude package feature - there for dev, not for compile)
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Sun, May 1, 2011 at 10:24 AM, hosey hosey <[hidden email]>
>>>> wrote:
>>>>>
>>>>> 17fps on notion ink adam
>>>>>
>>>>> On Apr 30, 2011 3:20 AM, "Niel Drummond" <[hidden email]>
>>>>> wrote:
>>>>> >
>>>>> > Thanks for the heads up, nice to see some game demos appearing in the
>>>>> > haxe-js world - I have been investigating how to run physics in a
>>>>> > separate thread with box2d when I get a free moment, which is the
>>>>> > bottleneck here and would give a significant performance boost.
>>>>> >
>>>>> > I guess the drop in framerate comes when more than a few objects are
>>>>> > added to the physics engine - maybe objects are not being properly
>>>>> > dereferenced when they are removed? Should be worth investigating..
>>>>> >
>>>>> > I should give credit where it is due, some contributors supplied nice
>>>>> > patches recently: brett periscopic (author of the gtween port), also
>>>>> > tecteun. Zjnue also continues to provide a wealth of qualitative test
>>>>> > cases / bug reports, and has successfully used the library in a
>>>>> > couple
>>>>> > of commercial projects (one at SXSW I think).
>>>>> >
>>>>> > As you say, it's *almost* there.. still have a couple of bugs to fix
>>>>> > :-)
>>>>> >
>>>>> > @tarwin: chrome has a very extensive profiling suite, though it's no
>>>>> > good for microbenchmarks.
>>>>> >
>>>>> > - Niel
>>>>> >
>>>>> > On 04/30/11 09:06, Tony Polinelli wrote:
>>>>> > > Unfortunately it peeks at 2fps on iphone4- not exactly a flashgame
>>>>> > > mobile replacement!
>>>>> > >
>>>>> > > Fortunately haxe obviously compiles nativly too tho ;)
>>>>> > >
>>>>> > > On Saturday, April 30, 2011, Tony Polinelli
>>>>> > > <[hidden email]> wrote:
>>>>> > >> Hi guys,
>>>>> > >> Thought that i'd share a little test using jeash.
>>>>> > >> http://touchmypixel.com/test/catch/
>>>>> > >>  <http://touchmypixel.com/test/catch/>It runs really well in
>>>>> > >> chrome. It starts off well - but as soon as collisions occur, the framerate
>>>>> > >> drops, and then never recovers. I am thinking that it is probably gc'ing not
>>>>> > >> working correctly? I'll have to poke around a little when i get time
>>>>> > >>
>>>>> > >> Definately shows great promise for the jeash project, as i didnt
>>>>> > >> have to put in too much work to get this little test to work. It shows that
>>>>> > >> there are still a few bugs, but on a whole, *nearly* works as expected!
>>>>> > >> great work Niel (i think)
>>>>> > >>
>>>>> > >> If people want to test the code more pm me.
>>>>> > >> cheersTony
>>>>> > >>
>>>>> > >>
>>>>> > >>
>>>>> > >> --
>>>>> > >> Tony Polinelli
>>>>> > >> http://touchmypixel.com
>>>>> > >>
>>>>> > >>
>>>>> >
>>>>> >
>>>>> > --
>>>>> > haXe - an open source web programming language
>>>>> > http://haxe.org
>>>>>
>>>>> --
>>>>> haXe - an open source web programming language
>>>>> http://haxe.org
>>>>
>>>>
>>>>
>>>> --
>>>> Tony Polinelli
>>>> http://touchmypixel.com
>>>>
>>>> --
>>>> haXe - an open source web programming language
>>>> http://haxe.org
>>>
>>>
>>> --
>>> haXe - an open source web programming language
>>> http://haxe.org
>>
>>
>> --
>> haXe - an open source web programming language
>> http://haxe.org
>
>
>
> --
> Tony Polinelli
> http://touchmypixel.com
> --
> haXe - an open source web programming language
> http://haxe.org
>
>
> --
> haXe - an open source web programming language
> http://haxe.org
>

--
haXe - an open source web programming language
http://haxe.org
Reply | Threaded
Open this post in threaded view
|

Re: early jeash test

Gamehaxe
In reply to this post by Niel Drummond-3
Hi,
I have been slowly developing the "gm2d" project  
(http://code.google.com/p/gm2d/)
which is built on top of the flash API, but can also side-step it
and go direct to NME.  It has abstractions/optimisations for
"A bunch of sprites(tiles) moving around in a viewport".

The idea could then be to draw people in with the familiar flash API, then
sell them the optimised API (which is a "copyPixels" engine in flash).
Since this mixes with the flash API, it is not an all-or-nothing
"mega framework", but a logical optimisation. Since is it cross-platform,
it should maintain the advantages of haxe-eash philosophy.

Hugh

> On 05/03/11 06:43, [hidden email] wrote:
>> The issue that keeps coming back to me on Jeash, is can we justify
>> Canvas as a minimum for instance...
>>
>> http://sebleedelisle.com/2011/04/html5javascript-platform-game-optimised-for-ipad/
>>
>> you can do a lot in HaXe javascript without Canvas.  I think
>> experimenting with Jeash will inform our flash coding, and this is why
>> I suggested we should consider improving flash so that the targets are
>> closer and also faster with a single workflow but I think this is the
>> future evolving that we have access to now but the workflows are still
>> evolving, so very exciting times, but maybe on the current web this is
>> still more for tomorrow than today.
>>
>
> Yes, it's true that Apple's CoreGraphics library is hindering the
> performance of canvas pixel manipulation compared to Google's Skia on
> Windows and Linux, though it's potentially just a matter of time that
> this is fixed.
>
> Afterall, there isn't anything stopping you from blitting images per
> frame in jeash, caching the results, and swapping the displayobject just
> like in the game you posted.
>
> I think it's going to be quite difficult convincing haxers and
> non-haxers alike that a non-flash API is the way forwards for a language
> with many associations to Actionscript, even if it sometimes makes sense
> to do so (sorry to be blunt), though I am happy to be proved wrong about
> this.
>
> - Niel
>
>>
>> On 3 May 2011, at 04:13, Tony Polinelli wrote:
>>
>>> I still think that it is too early for best practices and as the
>>> saying goes - dont prematurely optimize. Jeash is in a very early
>>> stage, and is aimed at replicating the flash API. I think that this
>>> is great for crossplatform functionality, and great to slow the power
>>> of haxe js. If you are wanting to dev for older js engines, you might
>>> want to optimize a lot, but then you mightn't want to use jeash. I
>>> think the focus should be on jeash being crossplatform, and allowing
>>> the developer to dev, like they were in flash - and then wait for the
>>> browsers to catch up - FF4 seems kinda quick - very promising in this
>>> respect.
>>>
>>>
>>>
>>>
>>> On Tue, May 3, 2011 at 11:49 AM, Tarwin Stroh-Spijer
>>> <[hidden email] <mailto:[hidden email]>> wrote:
>>>
>>>     Interesting. I'm trying to come up with some best practices at
>>>     the moment, as I'm giving a talk at Adobe tomorrow night. I think
>>>     Jeash (haxe being able to do canvas stuff) will be something
>>>     really interesting to people and I'd like to give them the best
>>>     info.
>>>
>>>     Any examples would be great !
>>>
>>>
>>>
>>>     Tarwin Stroh-Spijer
>>>     _______________________
>>>
>>>     Touch My Pixel
>>>     http://www.touchmypixel.com/
>>>     phone: +61 3 8060 5321 <tel:%2B61%203%208060%205321>
>>>     _______________________
>>>
>>>
>>>     On Tue, May 3, 2011 at 11:33 AM, hosey hosey
>>>     <[hidden email] <mailto:[hidden email]>> wrote:
>>>
>>>         Has there been any best practices written on Jeash and SWF
>>>         programming.  I use the display objects like candy (anything
>>>         that has interaction is a new Display Object).   I am
>>>         assuming we try to limit that as much as possible for Jeash
>>>         given it makes a full sized canvas for each display object?
>>>         Basically we should track the mouse position over the main
>>>         canvas and use a "mouse layer" canvas for interactions.   I
>>>         am making a Visio-esque application.
>>>
>>>         At this point I am looking at using CSS to control the layout
>>>         and haxe/jquery to write manipulate it or using Jeash and
>>>         pulling everything into the canvas.  This will be the fall
>>>         back when flash is not available.
>>>
>>>         Also I am looking at CPP output for Android/Iphone/Playbook
>>>         as end goals as well....
>>>
>>>
>>>         On Sat, Apr 30, 2011 at 6:11 PM, Tony Polinelli
>>>         <[hidden email] <mailto:[hidden email]>> wrote:
>>>
>>>             17fps on a tablet is HUGE - also, i just d/l ff4 =
>>>             amazing ~50, that is definately enough to make a
>>>             fullblown game in! flash game sans-flash!
>>>
>>>             pity they code is ~700kb in JS. (probably ~50 swf)  the
>>>             js target really needs to be able to complie to modules,
>>>             so you could have jeash.js and box2d.js cached by the
>>>             browser. I was thinking this could be done by:
>>>
>>>             haxe -jslib jeash -jslib box2d
>>>
>>>             it would use those projects - for autocompletion etc, but
>>>             not compile the code in. You would have a jeash.js
>>>             already compiled out by itself. (kinda like an exclude
>>>             package feature - there for dev, not for compile)
>>>
>>>
>>>
>>>
>>>
>>>
>>>             On Sun, May 1, 2011 at 10:24 AM, hosey hosey
>>>             <[hidden email] <mailto:[hidden email]>> wrote:
>>>
>>>                 17fps on notion ink adam
>>>
>>>                 On Apr 30, 2011 3:20 AM, "Niel Drummond"
>>>                 <[hidden email]
>>>                 <mailto:[hidden email]>> wrote:
>>>                 >
>>>                 > Thanks for the heads up, nice to see some game
>>>                 demos appearing in the
>>>                 > haxe-js world - I have been investigating how to
>>>                 run physics in a
>>>                 > separate thread with box2d when I get a free
>>>                 moment, which is the
>>>                 > bottleneck here and would give a significant
>>>                 performance boost.
>>>                 >
>>>                 > I guess the drop in framerate comes when more than
>>>                 a few objects are
>>>                 > added to the physics engine - maybe objects are not
>>>                 being properly
>>>                 > dereferenced when they are removed? Should be worth
>>>                 investigating..
>>>                 >
>>>                 > I should give credit where it is due, some
>>>                 contributors supplied nice
>>>                 > patches recently: brett periscopic (author of the
>>>                 gtween port), also
>>>                 > tecteun. Zjnue also continues to provide a wealth
>>>                 of qualitative test
>>>                 > cases / bug reports, and has successfully used the
>>>                 library in a couple
>>>                 > of commercial projects (one at SXSW I think).
>>>                 >
>>>                 > As you say, it's *almost* there.. still have a
>>>                 couple of bugs to fix :-)
>>>                 >
>>>                 > @tarwin: chrome has a very extensive profiling
>>>                 suite, though it's no
>>>                 > good for microbenchmarks.
>>>                 >
>>>                 > - Niel
>>>                 >
>>>                 > On 04/30/11 09:06, Tony Polinelli wrote:
>>>                 > > Unfortunately it peeks at 2fps on iphone4- not
>>>                 exactly a flashgame
>>>                 > > mobile replacement!
>>>                 > >
>>>                 > > Fortunately haxe obviously compiles nativly too
>>>                 tho ;)
>>>                 > >
>>>                 > > On Saturday, April 30, 2011, Tony Polinelli
>>>                 <[hidden email]
>>>                 <mailto:[hidden email]>> wrote:
>>>                 > >> Hi guys,
>>>                 > >> Thought that i'd share a little test using jeash.
>>>                 > >> http://touchmypixel.com/test/catch/
>>>                 > >>  <http://touchmypixel.com/test/catch/>It runs
>>>                 really well in chrome. It starts off well - but as
>>>                 soon as collisions occur, the framerate drops, and
>>>                 then never recovers. I am thinking that it is
>>>                 probably gc'ing not working correctly? I'll have to
>>>                 poke around a little when i get time
>>>                 > >>
>>>                 > >> Definately shows great promise for the jeash
>>>                 project, as i didnt have to put in too much work to
>>>                 get this little test to work. It shows that there are
>>>                 still a few bugs, but on a whole, *nearly* works as
>>>                 expected! great work Niel (i think)
>>>                 > >>
>>>                 > >> If people want to test the code more pm me.
>>>                 > >> cheersTony
>>>                 > >>
>>>                 > >>
>>>                 > >>
>>>                 > >> --
>>>                 > >> Tony Polinelli
>>>                 > >> http://touchmypixel.com
>>>                 > >>
>>>                 > >>
>>>                 >
>>>                 >
>>>                 > --
>>>                 > haXe - an open source web programming language
>>>                 > http://haxe.org
>>>
>>>
>>>                 --
>>>                 haXe - an open source web programming language
>>>                 http://haxe.org
>>>
>>>
>>>
>>>
>>>             --
>>>             Tony Polinelli
>>>             http://touchmypixel.com
>>>
>>>             --
>>>             haXe - an open source web programming language
>>>             http://haxe.org
>>>
>>>
>>>
>>>         --
>>>         haXe - an open source web programming language
>>>         http://haxe.org
>>>
>>>
>>>
>>>     --
>>>     haXe - an open source web programming language
>>>     http://haxe.org
>>>
>>>
>>>
>>>
>>> --
>>> Tony Polinelli
>>> http://touchmypixel.com
>>> --
>>> haXe - an open source web programming language
>>> http://haxe.org
>>

--
haXe - an open source web programming language
http://haxe.org