Those who have commit-access to the SVN repository are few, and they are probably busy people.
This makes submitting changes by relative unknowns slow and unmotivating.
I suggest the creation of a Patches forum; this way, the community can try changes out and leave comments, so that the few with real power don't have to do so much sifting.
Patches
-
hybrid
- Admin
- Posts: 14143
- Joined: Wed Apr 19, 2006 9:20 pm
- Location: Oldenburg(Oldb), Germany
- Contact:
Good idea. That's why the bug forum and the patch page were invented 
so if you have some bug fixes provide the patch in the bug forum, I'll also host in on my patch page. Add-ons better go to the code snippets forum, but will probably make it to the patch page as well. That way the developers will see these things, can test them, and maybe finally integrate them into the core.
so if you have some bug fixes provide the patch in the bug forum, I'll also host in on my patch page. Add-ons better go to the code snippets forum, but will probably make it to the patch page as well. That way the developers will see these things, can test them, and maybe finally integrate them into the core.
-
hybrid
- Admin
- Posts: 14143
- Joined: Wed Apr 19, 2006 9:20 pm
- Location: Oldenburg(Oldb), Germany
- Contact:
But most patches fix bugs, such that there would be two threads. Furthermore, the forum does not allow to upload files, you have to host them somewhere. And now the patch page comes in 
Many large additions also go into the project forum because they are such large extensions. I think that's quite ok, maybe the forum descriptions should be more specific on that.
Many large additions also go into the project forum because they are such large extensions. I think that's quite ok, maybe the forum descriptions should be more specific on that.
Perhaps there could be an official way to upload patches.
Irrlicht seems small at first glance, but it's quite a huge project in the sense that it has a gigantic following and offical and unoffical daughter sites.
The patch and bug system are a bit immature.
It's time we got rid of Spintz and Lightfeather and other spin offs; we need to centralize everything Irrlicht related with one central, core, well-built system.
I'm not the one to dictate how to carry this off, obviously, but I think it's time the leaders showed more direction.
Irrlicht seems small at first glance, but it's quite a huge project in the sense that it has a gigantic following and offical and unoffical daughter sites.
The patch and bug system are a bit immature.
It's time we got rid of Spintz and Lightfeather and other spin offs; we need to centralize everything Irrlicht related with one central, core, well-built system.
I'm not the one to dictate how to carry this off, obviously, but I think it's time the leaders showed more direction.
-
hybrid
- Admin
- Posts: 14143
- Joined: Wed Apr 19, 2006 9:20 pm
- Location: Oldenburg(Oldb), Germany
- Contact:
I'm now part of the official developer team, so I'm really trying to
Most small patches from NX are already included in the SVN version (were only a few I extracted from them), and IrrSpintz is kind of a testbed for large enhancements which is really useful. I guess that spintz will sync with SVN and thus become closer to Irrlicht core again. But it will take some time and effort to adapt all changes fom all enhancements available already. And most of these will also require major changes before they fit into the Irrlicht style.
Maybe moving the patch page into the Wiki would be an idea (if it comes up again), but I've always hosted all patches sent to me.
Most small patches from NX are already included in the SVN version (were only a few I extracted from them), and IrrSpintz is kind of a testbed for large enhancements which is really useful. I guess that spintz will sync with SVN and thus become closer to Irrlicht core again. But it will take some time and effort to adapt all changes fom all enhancements available already. And most of these will also require major changes before they fit into the Irrlicht style.
Maybe moving the patch page into the Wiki would be an idea (if it comes up again), but I've always hosted all patches sent to me.
yes, triple ouch!
Lightfeather has nothing to do with Irrlicht nowadays. It had its code entirely rewritten. Spintz, on the other hand, is a set of improvements, patches and bugfixes to Irrlicht´s code; besides Spintz' own personal additions to the engine. I don´t see the point of not coexisting with them?
Lightfeather has nothing to do with Irrlicht nowadays. It had its code entirely rewritten. Spintz, on the other hand, is a set of improvements, patches and bugfixes to Irrlicht´s code; besides Spintz' own personal additions to the engine. I don´t see the point of not coexisting with them?
-
hybrid
- Admin
- Posts: 14143
- Joined: Wed Apr 19, 2006 9:20 pm
- Location: Oldenburg(Oldb), Germany
- Contact:
But forks are also a very common method in software development. And at least spintz is also very active in the forums (only a little lack in time recently) and still developing merely Irrlicht updates than an independent spin-off. So we should just please spintz to avoid a development on IrrSpintz++ (which was the official announcement of the independence of the LF predecessor).
I realize you guys think I'm pulling stuff out of thin air, but this doesn't require deep analysis.
There are lots of talented people working on LightFeather right now. Spintz is busily spending his time working on items that make his version intriguing.
This simply means the environment around Irrlicht isn't conducive to contribution.
Besides, there is a realworld "it looks incomplete" perception when people are spinning out their own knock-offs, because "Irrlicht is too buggy, simplified, or featureless."
Come on guys. Lose your academic need for "experts" and "authorities" to speak on such subjects.
There are lots of talented people working on LightFeather right now. Spintz is busily spending his time working on items that make his version intriguing.
This simply means the environment around Irrlicht isn't conducive to contribution.
Besides, there is a realworld "it looks incomplete" perception when people are spinning out their own knock-offs, because "Irrlicht is too buggy, simplified, or featureless."
Come on guys. Lose your academic need for "experts" and "authorities" to speak on such subjects.
-
Eternl Knight
- Posts: 313
- Joined: Tue Nov 01, 2005 5:01 am
Well, spinoffs are not only common practise in open-source development, but are often a source of new/innovative features that the original developers are either unable/unwilling to implement. For example, the terrain feature as implemented currently came from Spintz's spinoff.
LightFeather is not longer simply an Irrlicht spinoff either. For example, it had Mac OSX support long before Irrlilcht did and it's rendering subsystem is very different to Irrlicht's beinig based around "Render Passes", hardware vertex buffers, etc. When one looks through the code, it is obvious that the parent project is Irrlicht. But there is alot of architecture differences that (I believe) are superior to that in Irrlicht.
Now, I don't know what hybrid's schedule is like but I can bet he simplpy can't merge LightFeather in a short amount of time (and remember, Niko still get's a say in this!). I believe given this situation that simiply "killing" the alternatives is unwise (besides which - only the developers themselves can "kill" those projects, we don't really get a say in it)
--EK
LightFeather is not longer simply an Irrlicht spinoff either. For example, it had Mac OSX support long before Irrlilcht did and it's rendering subsystem is very different to Irrlicht's beinig based around "Render Passes", hardware vertex buffers, etc. When one looks through the code, it is obvious that the parent project is Irrlicht. But there is alot of architecture differences that (I believe) are superior to that in Irrlicht.
Now, I don't know what hybrid's schedule is like but I can bet he simplpy can't merge LightFeather in a short amount of time (and remember, Niko still get's a say in this!). I believe given this situation that simiply "killing" the alternatives is unwise (besides which - only the developers themselves can "kill" those projects, we don't really get a say in it)
--EK
First of all to Eternl Knight: thanks for the kind words!
now to lingwitt:
when irrlichtnx was started about 2 years ago, there was no cvs/svn for irrlicht, there was only one developer working on it and many useful patches were posted on the forum and not integrated by that single developer. there have also been mutliple requests for a cvs for irrlicht back then which were denied.
so tell me, please, how would you have made progress possible without forking ?
and since youre not around long enough to know: i am the one who started irrlichtnx, so i consider myself to be an expert and authority on this topic.
Now to your post saying that LF and irrspintz should be killed (i really didnt want to discuss it but it seems unavoidable):
as afecelis (and also Eternl Knight) already pointed out: you shouldnt consider LF to be an irrlicht spinoff (thats why the name was changed from irrlichtnx++ to LF).
all it has in common with irrlicht these days are the base classes (vector,string,list,array and such) ,which makes it look similar when you browse the code, since those are heavily used classes. but even those classes have been patched a lot (patches for these classes are the reason for me to visit this forum regularly) .
you can say about LF what you want, i respect every persons right to his/her own opinion, but but asking for its demise, especially when you have no idea, what youre talking about, is just too much.
now to lingwitt:
im sorry that i have to disagree, but thats exactly what youre doing.lingwitt wrote:I realize you guys think I'm pulling stuff out of thin air, but this doesn't require deep analysis.
when irrlichtnx was started about 2 years ago, there was no cvs/svn for irrlicht, there was only one developer working on it and many useful patches were posted on the forum and not integrated by that single developer. there have also been mutliple requests for a cvs for irrlicht back then which were denied.
so tell me, please, how would you have made progress possible without forking ?
thanksThere are lots of talented people working on LightFeather right now.
it has apparently gotten a bit better within the last 3 weeks or so since hybrid and bitplane got commit rights. but before that ...This simply means the environment around Irrlicht isn't conducive to contribution.
there are many features i deem necessary which irrlicht doesnt have. that doesnt make irrlicht buggy, simplified or featureless. merely not fitting my needs.Besides, there is a realworld "it looks incomplete" perception when people are spinning out their own knock-offs, because "Irrlicht is too buggy, simplified, or featureless."
i wasnt calling for any experts or authorities, i was just saying, that you are none with regard to this matter (this is not to insult you, its just that youre not around long enough to be one).Come on guys. Lose your academic need for "experts" and "authorities" to speak on such subjects.
and since youre not around long enough to know: i am the one who started irrlichtnx, so i consider myself to be an expert and authority on this topic.
Now to your post saying that LF and irrspintz should be killed (i really didnt want to discuss it but it seems unavoidable):
as afecelis (and also Eternl Knight) already pointed out: you shouldnt consider LF to be an irrlicht spinoff (thats why the name was changed from irrlichtnx++ to LF).
all it has in common with irrlicht these days are the base classes (vector,string,list,array and such) ,which makes it look similar when you browse the code, since those are heavily used classes. but even those classes have been patched a lot (patches for these classes are the reason for me to visit this forum regularly) .
you can say about LF what you want, i respect every persons right to his/her own opinion, but but asking for its demise, especially when you have no idea, what youre talking about, is just too much.