Proposed irc triage for nme

classic Classic list List threaded Threaded
22 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Proposed irc triage for nme

Alex Liebert
There's been a huge jump in activity on this list about nme- many thanks to Hugh and Joshua and everyone who's been making the project so appealing recently!

I've made a few commits and tried to keep track of any features I add in the googlecode tracker.  There's a few in there I'm working on, a few I'm stumped on, and a few from other people that need attention.  Plus quite a few issues showing up in Haxelist but not googlecode.

Would anyone be interested in this plan:

1- rebuild all binaries in svn and invite the list to try things out, based on Josh g's new how to

2- call for any pending issues to be submitted to the googlecode issues list

3- schedule a time for an irc meeting that works for the contributors + any interested parties.  Agenda:

i- review each issue and triage- accept, defer or assign to a comitter

ii- open discussion of feature plans / code issues

iii- q and a session

iv- consider some dates and milestones- (loosely since this a free time project for everyone )

Alex

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

Re: Proposed irc triage for nme

Nicolas Cannasse
Le 15/07/2011 20:46, Alex Liebert a écrit :
> There's been a huge jump in activity on this list about nme- many thanks to Hugh and Joshua and everyone who's been making the project so appealing recently!
>
> I've made a few commits and tried to keep track of any features I add in the googlecode tracker.  There's a few in there I'm working on, a few I'm stumped on, and a few from other people that need attention.  Plus quite a few issues showing up in Haxelist but not googlecode.
>
> Would anyone be interested in this plan:

+1

Nicolas

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

Re: Proposed irc triage for nme

Joshua Harlan Lifton
In reply to this post by Alex Liebert
> Would anyone be interested in this plan:

I'm in.

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

Re: Proposed irc triage for nme

jlm@justinfront.net
In reply to this post by Alex Liebert
I gather Lion is in beta and due for final release very soon in the  
next week or ?, I think it would be worth syncing this with its launch  
maybe preparing something good, so we are sure it does not fail, I  
think this maybe worthwhile and good to see the haXe community as  
responsive, ie lets try to test and turn round within a day, well I  
know Nicolas does that with adobe :)  Just may not be good if we get a  
stable version out there and it failed on Lion I know the audience is  
small but they blog a lot.  I currently have leopard, should I upgrade  
to snowy and lion, or would it be worthwhile for me to test it as I  
am? But cross platform tests before a launch I am all in favour of.

Cheers ;j

On 15 Jul 2011, at 19:46, Alex Liebert wrote:

> There's been a huge jump in activity on this list about nme- many  
> thanks to Hugh and Joshua and everyone who's been making the project  
> so appealing recently!
>
> I've made a few commits and tried to keep track of any features I  
> add in the googlecode tracker.  There's a few in there I'm working  
> on, a few I'm stumped on, and a few from other people that need  
> attention.  Plus quite a few issues showing up in Haxelist but not  
> googlecode.
>
> Would anyone be interested in this plan:
>
> 1- rebuild all binaries in svn and invite the list to try things  
> out, based on Josh g's new how to
>
> 2- call for any pending issues to be submitted to the googlecode  
> issues list
>
> 3- schedule a time for an irc meeting that works for the  
> contributors + any interested parties.  Agenda:
>
> i- review each issue and triage- accept, defer or assign to a comitter
>
> ii- open discussion of feature plans / code issues
>
> iii- q and a session
>
> iv- consider some dates and milestones- (loosely since this a free  
> time project for everyone )
>
> Alex
>
> Sent from my Phone
> --
> 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: Proposed irc triage for nme

Alex Liebert
Sounds like a good concern; I'd suggest adding 'test os x lion
support' to http://code.google.com/p/nekonme/issues/list .

On Fri, Jul 15, 2011 at 2:25 PM, [hidden email]
<[hidden email]> wrote:

> I gather Lion is in beta and due for final release very soon in the next
> week or ?, I think it would be worth syncing this with its launch maybe
> preparing something good, so we are sure it does not fail, I think this
> maybe worthwhile and good to see the haXe community as responsive, ie lets
> try to test and turn round within a day, well I know Nicolas does that with
> adobe :)  Just may not be good if we get a stable version out there and it
> failed on Lion I know the audience is small but they blog a lot.  I
> currently have leopard, should I upgrade to snowy and lion, or would it be
> worthwhile for me to test it as I am? But cross platform tests before a
> launch I am all in favour of.
>
> Cheers ;j
>
> On 15 Jul 2011, at 19:46, Alex Liebert wrote:
>
>> There's been a huge jump in activity on this list about nme- many thanks
>> to Hugh and Joshua and everyone who's been making the project so appealing
>> recently!
>>
>> I've made a few commits and tried to keep track of any features I add in
>> the googlecode tracker.  There's a few in there I'm working on, a few I'm
>> stumped on, and a few from other people that need attention.  Plus quite a
>> few issues showing up in Haxelist but not googlecode.
>>
>> Would anyone be interested in this plan:
>>
>> 1- rebuild all binaries in svn and invite the list to try things out,
>> based on Josh g's new how to
>>
>> 2- call for any pending issues to be submitted to the googlecode issues
>> list
>>
>> 3- schedule a time for an irc meeting that works for the contributors +
>> any interested parties.  Agenda:
>>
>> i- review each issue and triage- accept, defer or assign to a comitter
>>
>> ii- open discussion of feature plans / code issues
>>
>> iii- q and a session
>>
>> iv- consider some dates and milestones- (loosely since this a free time
>> project for everyone )
>>
>> Alex
>>
>> Sent from my Phone
>> --
>> 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: Proposed irc triage for nme

Gamehaxe
In reply to this post by Alex Liebert
Hi,
Well, my idea on this is:

1. Release new (minor) version of HXCPP.  This will include webOS.
    -> This depends on me being able to test basic functionality on webOS.
       I have been moving towards this this week, since I now have a webOS  
device.
    -> Depends on having the install-tool working for webOS - this is  
happening.

Once HXCPP is released, I will release NME.
My main criteria for this is are:
1. That it has not gone backwards (all samples compile + work)
2. The install tool works for most use cases.
    -> This has been in flux this week.

The main thing that was holding me up was getting icons working on all  
platforms,
which I have now mostly done.
I have deferred the idea of getting installers (eg, NSIS) going on all  
platforms.

I am not too concerned about any other bugs, unless they have been
introduced since the last release or are "show stoppers".

It has been such a long time since the last release, that I do not
really see the point in a "beta". Just do a release, and then do
another release later.

My main concern is getting the install-tool syntax settled, since once
this is released, it will be harder to go back.

Hugh

> There's been a huge jump in activity on this list about nme- many thanks  
> to Hugh and Joshua and everyone who's been making the project so  
> appealing recently!
>
> I've made a few commits and tried to keep track of any features I add in  
> the googlecode tracker.  There's a few in there I'm working on, a few  
> I'm stumped on, and a few from other people that need attention.  Plus  
> quite a few issues showing up in Haxelist but not googlecode.
>
> Would anyone be interested in this plan:
>
> 1- rebuild all binaries in svn and invite the list to try things out,  
> based on Josh g's new how to
>
> 2- call for any pending issues to be submitted to the googlecode issues  
> list
>
> 3- schedule a time for an irc meeting that works for the contributors +  
> any interested parties.  Agenda:
>
> i- review each issue and triage- accept, defer or assign to a comitter
>
> ii- open discussion of feature plans / code issues
>
> iii- q and a session
>
> iv- consider some dates and milestones- (loosely since this a free time  
> project for everyone )
>
> Alex
>
> Sent from my Phone
> --
> 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: Proposed irc triage for nme

Alex Liebert
I think that sounds great.  I appreciate all the great work (and the
fixes for some of the things I requested like the font problem, and
the activation/deactivation handlers!) and look forward to a release.
Also keen to help with testing however I can.  I'm doing active
development towards a production release on multiple platforms right
now, so I've been finding alot of real world issues.

Now speaking just from my personal point of view:  I'm 95% of the way
towards having everything in nme working to the point I can release a
game publicly on android and ios.  None of these are showstoppers for
me really (because I can forget about Android 'til a later date if I
have to) but these are the issues currently giving me trouble:

iOS: Text won't render on iOS 3.1.3.  I think backwards compatibility
is really just a 'nice to have', so I'm okay making my app require
4.0+ to get around this.  I tried some debugging but got stumped.
http://code.google.com/p/nekonme/issues/detail?id=112

iOS: orientation can't be set to landscape at start on 3.1.3.  I've
been able to fix this locally but only by changing the cpp code
manually to force it.  The issue there is that the first time
'shouldAutoRotate' is called in the iOS lifecycle, the haxe/NME side
is not yet initialized so the user hasn't been able to override that
method from haXe.  (the reason it's not an issue on 4.0 is because
theres a plist value that set initial orientation.)

-Android: when app loses focus, all sound samples fail to play until
app restart.  This is because the OS seems to drain the SoundPool when
an app is backgrounded.  The solution is to rebuild the pool, but it's
more complicated work in C++ then what I've done so far adding
SharedObject etc, and I'm not confident to do it safely.  I can
probably figure it out eventually because I'm getting a little more
familiar with C++ every day as I invest more into nme.
http://code.google.com/p/nekonme/issues/detail?id=109

-Android: Roll out events do not fire.  I think this can be resolved
on the haxe side without touching c++, I just don't completely
understand the Event handling flow for nme yet.
http://code.google.com/p/nekonme/issues/detail?id=111

Anything else I've had luck fixing myself, or will get to shortly (I
added sound transform handling for iOS and Android, added SharedObject
for iOS and android, and also Lib.getURL on both and committed.)  I
plan to look into the event propogation issue this weekend.  I also
plan to fill out the windows and mac platform support for the features
I added to mobile soon.

So that's my list :)  I know a few other people are working on apps
and having some issues, some of which are also in googlecode issue
list.

I'll also be looking at the install tool / templated resources access
stuff we discussed a few weeks ago very soon.

Alex

On Fri, Jul 15, 2011 at 9:51 PM, Gamehaxe <[hidden email]> wrote:

> Hi,
> Well, my idea on this is:
>
> 1. Release new (minor) version of HXCPP.  This will include webOS.
>   -> This depends on me being able to test basic functionality on webOS.
>      I have been moving towards this this week, since I now have a webOS
> device.
>   -> Depends on having the install-tool working for webOS - this is
> happening.
>
> Once HXCPP is released, I will release NME.
> My main criteria for this is are:
> 1. That it has not gone backwards (all samples compile + work)
> 2. The install tool works for most use cases.
>   -> This has been in flux this week.
>
> The main thing that was holding me up was getting icons working on all
> platforms,
> which I have now mostly done.
> I have deferred the idea of getting installers (eg, NSIS) going on all
> platforms.
>
> I am not too concerned about any other bugs, unless they have been
> introduced since the last release or are "show stoppers".
>
> It has been such a long time since the last release, that I do not
> really see the point in a "beta". Just do a release, and then do
> another release later.
>
> My main concern is getting the install-tool syntax settled, since once
> this is released, it will be harder to go back.
>
> Hugh
>
>> There's been a huge jump in activity on this list about nme- many thanks
>> to Hugh and Joshua and everyone who's been making the project so appealing
>> recently!
>>
>> I've made a few commits and tried to keep track of any features I add in
>> the googlecode tracker.  There's a few in there I'm working on, a few I'm
>> stumped on, and a few from other people that need attention.  Plus quite a
>> few issues showing up in Haxelist but not googlecode.
>>
>> Would anyone be interested in this plan:
>>
>> 1- rebuild all binaries in svn and invite the list to try things out,
>> based on Josh g's new how to
>>
>> 2- call for any pending issues to be submitted to the googlecode issues
>> list
>>
>> 3- schedule a time for an irc meeting that works for the contributors +
>> any interested parties.  Agenda:
>>
>> i- review each issue and triage- accept, defer or assign to a comitter
>>
>> ii- open discussion of feature plans / code issues
>>
>> iii- q and a session
>>
>> iv- consider some dates and milestones- (loosely since this a free time
>> project for everyone )
>>
>> Alex
>>
>> Sent from my Phone
>> --
>> 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: Proposed irc triage for nme

Gamehaxe
Hi,

I have most of the things I wanted to get into NME going now - mainly
the new install-tool, icons and webos.

So I'm now ready to look at any "must fix" issues before release.
Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
people are hanging out for?

I know there are quite a few outstanding issues ATM, but if they come in
faster than I can fix them, then there will never be a release.

Hugh

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

Re: Proposed irc triage for nme

Marcelo de Moraes Serpa
Hey Hugh,

I'd like to help somehow, in fact, I think it'd be a nice way for me to get deeper into C and C++, and I'm willing to. Where are the current issues listed?

As a side note, why don't we move NME over to github? It'd facilitate collaboration since we could have many forks and then you could accept the patches (through pull requests) as needed.

Cheers,

Marcelo.

On Tue, Aug 2, 2011 at 12:04 PM, Gamehaxe <[hidden email]> wrote:
Hi,

I have most of the things I wanted to get into NME going now - mainly
the new install-tool, icons and webos.

So I'm now ready to look at any "must fix" issues before release.
Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
people are hanging out for?

I know there are quite a few outstanding issues ATM, but if they come in
faster than I can fix them, then there will never be a release.

Hugh

--
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: Proposed irc triage for nme

Joshua Harlan Lifton
In reply to this post by Gamehaxe
Hi Hugh,

The biggest NME issue for me at the moment is how polygons are filled,
specifically, how concave polygons are incorrectly filled. My
understanding from a previous thread on this topic is that it will
take a considerable amount of work to get beyond this problem. Is that
still the case? I've had to cut some important features from the
product I'm working on because of this bug, to the point where I think
it makes sense to offer a bounty of some sort. What's the proper forum
for bounties?

The other two big issues I've had are SSL, which I've solved for
myself but haven't yet integrated into trunk, and getting the new
install tool working on Mac OS 10.6.8, though I haven't tried the
latest SVN revisions in a week or two and probably won't have time to
do so for at least another week.

Other than that, I've noticed a lot of small TextField issues, some of
which have already been fixed, but nothing that should prevent a
release.

Thanks,
Josh

On Tue, Aug 2, 2011 at 1:04 PM, Gamehaxe <[hidden email]> wrote:

> Hi,
>
> I have most of the things I wanted to get into NME going now - mainly
> the new install-tool, icons and webos.
>
> So I'm now ready to look at any "must fix" issues before release.
> Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
> people are hanging out for?
>
> I know there are quite a few outstanding issues ATM, but if they come in
> faster than I can fix them, then there will never be a release.
>
> Hugh
>
> --
> 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: Proposed irc triage for nme

Gamehaxe
It is not impossible by any means, but not in the scope of this release.
It is on the roadmap, though - possible along with render-to-texture and
stage3d stuff.

Hugh

> The biggest NME issue for me at the moment is how polygons are filled,
> specifically, how concave polygons are incorrectly filled. My
> understanding from a previous thread on this topic is that it will
> take a considerable amount of work to get beyond this problem.

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

Re: Proposed irc triage for nme

Michael Baczynski-2
In reply to this post by Joshua Harlan Lifton
just wondering - why can't you decompose a concave polygon into convex ones and then render those?
There are very simple algorithms for this (e.g. ear clipping)

best,
michael

On 03.08.2011 10:23, Joshua Harlan Lifton wrote:

> Hi Hugh,
>
> The biggest NME issue for me at the moment is how polygons are filled,
> specifically, how concave polygons are incorrectly filled. My
> understanding from a previous thread on this topic is that it will
> take a considerable amount of work to get beyond this problem. Is that
> still the case? I've had to cut some important features from the
> product I'm working on because of this bug, to the point where I think
> it makes sense to offer a bounty of some sort. What's the proper forum
> for bounties?
>
> The other two big issues I've had are SSL, which I've solved for
> myself but haven't yet integrated into trunk, and getting the new
> install tool working on Mac OS 10.6.8, though I haven't tried the
> latest SVN revisions in a week or two and probably won't have time to
> do so for at least another week.
>
> Other than that, I've noticed a lot of small TextField issues, some of
> which have already been fixed, but nothing that should prevent a
> release.
>
> Thanks,
> Josh
>
> On Tue, Aug 2, 2011 at 1:04 PM, Gamehaxe<[hidden email]>  wrote:
>> Hi,
>>
>> I have most of the things I wanted to get into NME going now - mainly
>> the new install-tool, icons and webos.
>>
>> So I'm now ready to look at any "must fix" issues before release.
>> Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
>> people are hanging out for?
>>
>> I know there are quite a few outstanding issues ATM, but if they come in
>> faster than I can fix them, then there will never be a release.
>>
>> Hugh
>>
>> --
>> 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: Proposed irc triage for nme

davidedc
In reply to this post by Gamehaxe
Hi Hugh,

what's the approximate date of the release?

Have two suggestions. One is to have an option for the software renderer to draw in aliased mode rather than anti-aliased (the option would be linked to the quality settings of the stage). Maybe not a must-fix to be fair.

Also, It would be very cool to get a "draw on top of the previous frame" mode. To be fair, I think that really that's only interesting for translating Processing programs - so it's hardly a must fix, but I had quite some progress since my last mail on the subject and I think the results so far are quite pretty.

iPhone has a shortcut for that (I think I sent a patch but I never properly integrated it with the system of flags and properties of the stage).

Other platforms probably need to explicitly save the buffer as texture and then paint such texture on a quad on the next frame.

Cheers,
Davide

On Tue, Aug 2, 2011 at 6:04 PM, Gamehaxe <[hidden email]> wrote:
Hi,

I have most of the things I wanted to get into NME going now - mainly
the new install-tool, icons and webos.

So I'm now ready to look at any "must fix" issues before release.
Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
people are hanging out for?

I know there are quite a few outstanding issues ATM, but if they come in
faster than I can fix them, then there will never be a release.

Hugh

--
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: Proposed irc triage for nme

davidedc
In reply to this post by Michael Baczynski-2
The algorithms are all there (even though if we wanted something really solid then the algorithms are not so simple - Hugh pointed to a FIST implementation he did and it's no stroll in the park IMHO).

I personally didn't have the spine to marry any algorithm with the existing data structures though - manipulating those takes some focus and quality time (I think).

D

On Wed, Aug 3, 2011 at 12:31 PM, Michael Baczynski <[hidden email]> wrote:
just wondering - why can't you decompose a concave polygon into convex ones and then render those? There are very simple algorithms for this (e.g. ear clipping)

best,
michael


On 03.08.2011 10:23, Joshua Harlan Lifton wrote:
Hi Hugh,

The biggest NME issue for me at the moment is how polygons are filled,
specifically, how concave polygons are incorrectly filled. My
understanding from a previous thread on this topic is that it will
take a considerable amount of work to get beyond this problem. Is that
still the case? I've had to cut some important features from the
product I'm working on because of this bug, to the point where I think
it makes sense to offer a bounty of some sort. What's the proper forum
for bounties?

The other two big issues I've had are SSL, which I've solved for
myself but haven't yet integrated into trunk, and getting the new
install tool working on Mac OS 10.6.8, though I haven't tried the
latest SVN revisions in a week or two and probably won't have time to
do so for at least another week.

Other than that, I've noticed a lot of small TextField issues, some of
which have already been fixed, but nothing that should prevent a
release.

Thanks,
Josh

On Tue, Aug 2, 2011 at 1:04 PM, Gamehaxe<[hidden email]>  wrote:
Hi,

I have most of the things I wanted to get into NME going now - mainly
the new install-tool, icons and webos.

So I'm now ready to look at any "must fix" issues before release.
Fonts on 3.1 would be nice, but I can't test it.  Is there anything else
people are hanging out for?

I know there are quite a few outstanding issues ATM, but if they come in
faster than I can fix them, then there will never be a release.

Hugh

--
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: Proposed irc triage for nme

Gamehaxe
In reply to this post by Michael Baczynski-2
Mainly because the existing solution of using "cacheAsBitmap" solved the
problem (and others as well) for my use cases.
As I said, it's on the roadmap - I'll accept a patch if you have one :)

Hugh


> just wondering - why can't you decompose a concave polygon into convex  
> ones and then render those? There are very simple algorithms for this  
> (e.g. ear clipping)
>
> best,
> michael
>
> On 03.08.2011 10:23, Joshua Harlan Lifton wrote:
>> Hi Hugh,
>>
>> The biggest NME issue for me at the moment is how polygons are filled,
>> specifically, how concave polygons are incorrectly filled. My
>> understanding from a previous thread on this topic is that it will
>> take a considerable amount of work to get beyond this problem. Is that
>> still the case? I've had to cut some important features from the
>> product I'm working on because of this bug, to the point where I think
>> it makes sense to offer a bounty of some sort. What's the proper forum
>> for bounties?

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

Re: Proposed irc triage for nme

Gamehaxe
In reply to this post by davidedc
>
> what's the approximate date of the release?

I'm thinking about 1 week (that's half of 2 weeks if you have been  
following).
I'm just going for Low hanging fruit" and then I'll do the builds.


> Have two suggestions. One is to have an option for the software renderer  
> to
> draw in aliased mode rather than anti-aliased (the option would be  
> linked to
> the quality settings of the stage). Maybe not a must-fix to be fair.

I'll look to see if it is easy.  I bug report on
http://code.google.com/p/nekonme/issues/list
with source code is the best way to get my attention.


> Also, It would be very cool to get a "draw on top of the previous frame"
> mode. To be fair, I think that really that's only interesting for
> translating Processing programs - so it's hardly a must fix, but I had  
> quite
> some progress since my last mail on the subject and I think the results  
> so
> far are quite pretty.

This one strays quite far from a simple fix - may think about it later -
probably via manual render-opengl-to-texture, and then a display.Bitmap,
which would be a more general fix.
This would then be a manual process (probably works now, but would be very  
slow)

Hugh



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

Re: Proposed irc triage for nme

Joshua Harlan Lifton
In reply to this post by Gamehaxe
> Mainly because the existing solution of using "cacheAsBitmap" solved the
> problem (and others as well) for my use cases.

I haven't heard about this solution. Can you provide more details?

Thanks,
Josh

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

Re: Proposed irc triage for nme

davidedc
If you draw your concave polygon on a sprite and set the sprite's "cacheAsBitmap" to true, then the sw renderer draws the polygon instead of opengl, and the sw renderer draws the polygon correctly.

Only thing though is that there is a bitmap transfer from cpu to gpu every time you change the polygon. So if the polygon is static it doesn't matter, but if the polygon is big and changes shape every frame then it might get slow.

Also, the sw renderer always draws anti-aliased, while every other non-cached geometry you draw via opengl will be jagged (unless you are using multi-sampling, which might introduce its own performance drawbacks).

If you want to run a quick test you can hack this example on the left (just set the cacheAsBitmap flag to true): http://is.gd/ZsyMHj

Cheers,
Davide

~~~~~~~~~~~~~~~~~~~~~
"Unless Hugh says differently"


On Thu, Aug 4, 2011 at 9:07 AM, Joshua Harlan Lifton <[hidden email]> wrote:
> Mainly because the existing solution of using "cacheAsBitmap" solved the
> problem (and others as well) for my use cases.

I haven't heard about this solution. Can you provide more details?

Thanks,
Josh

--
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: Proposed irc triage for nme

Gamehaxe
In reply to this post by Joshua Harlan Lifton
cacheAsBitmap causes the software renderer to be used, and it creates a  
texture.
The software renderer also handles things like AA and gradients better,
but is is slower since you must upload a texture every time something  
changes.
If the content in relatively static, then it actually may be faster.

Hugh

>> Mainly because the existing solution of using "cacheAsBitmap" solved the
>> problem (and others as well) for my use cases.
>
> I haven't heard about this solution. Can you provide more details?
>
> Thanks,
> Josh

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

Re: Proposed irc triage for nme

Gamehaxe
In reply to this post by Gamehaxe
Actually, I think I fixed this a little while ago - have you checked  
recently?

Hugh

>  Have two suggestions. One is to have an option for the software  
> renderer to
> draw in aliased mode rather than anti-aliased (the option would be  
> linked to
> the quality settings of the stage). Maybe not a must-fix to be fair.
>  

--
haXe - an open source web programming language
http://haxe.org
12