Ogre Forums

the GLES2 RS is now running on Desktop GL using the ES Context. This allows to make a comparison GL vs GL3Plus vs GLES2:
https://ogrecave.github.io/ogre/gl_status/

If you are interested what the differences between the bitbucket and the github version are, see here:
https://github.com/OGRECave/ogre/blob/m ... angeLog.md

omg- evil commits on main-stream default-branch broke hydrax, crippeled ogitor and cost me 2 days work...
on ogreFrustrum enableReflection... @depriciated means old code is left for backward compability, whats the point of redirecting to the new buggy code?
i'm starting using this fork...
please consider adding paroj to the ogre-team- i feel he has a lot to offer at least for ogre1.10( which will probably stay with us for the next couple of years )
obviously he has the skills, motivation, and maturaty( responsible unbiased etc... )
ogre team now seems to actualy have only 3 active people...and one of them is superbz with ogre 2.1...

btw it will be super extra cool and a big time saver if we can download the artifacts from the cl-builds- instead of building...
thanks...

I plan to upgrade to 1.10 in the near future, so I definitely want to see it maintained.

Thanks for the work, paroj.

Am I the only one seeing this fork as a bad thing (not paroj's work itself, which is great, but the existence of a fork that gains popularity)? The whole "git vs hg" sentiment in this thread is a red herring. Who makes the decision on who becomes a team member, or who has rights over pushing to the source? Is there a decision to be made here? If made a team member, is paroj willing to work within the code versioning system chosen by the team? These are the questions that interest me. I've been around long enough to see projects get split apart by forks that begin exactly like this ... not saying that is what is happening, just saying it's giving me that itchy feeling. Doing nothing about this fork is not the right move. Doing nothing is how forks like this happen in the first place and contributors are left wondering where the fuck to put their contributions.

I do not have a problem with his fork. Thats the great thing about opensource you can take it, fork it and transform into the shape you like it.
What paroj does in his fork is ok and why not he likes to refactoring.
But the refactoring are a way to massive we need to sort the the bad merge from simigon and clean up the repo and than 1.10 will be done (with 4-5 extra things) - we will provide LTS for it.
Also our focus is 2.X its already in a good shape and needs some mobile love and some features here and there but its the future (Metal, Vulkan,...)

But to clear one thing I do not have to like it, what he does or how he does it nor we are not in a situation where somebody is taking over a project.

Anyway thumbs up for paroj providing us with PR which are all in my pipeline (to review) I am reviewing all of them - but as you guys know David and I got babies which need love, care and time also Matias is pretty busy
with his awesome new game you should check it out , so things slowed down.

Just you guys there is no offence to anybody!

Peace out

Wolfmanfx wrote:also Matias is pretty busy
with his awesome new game you should check it out , so things slowed down.

I would if i could!
where is it? what is it?

matias's game https://twitter.com/AllianceAirWar

i also don't see a problem with this fork...( although i do wish it was a direct fork sharing same repository... )
i will alway be loyal to the main branch and forever grateful to all contributers and maintainers, bless you, all comits even evil ones are sacred...
like murat said- this fork mainly represent 3 major changes to the main branch( called 1.10-unstable for a reason... )
1)it is like a stable-1.10 branch( aka LTS ) with backward compatibility to 1.9( which makes porting massive projects much easier.. )
2)it's supposed to have better mobile support( building ios is fixed, commits only reviewed on dx11 are pulled out..., opengl unified etc... )
3)CI integration...indeed a neccessirty...i think this is the main reason for paroj choosing github - i guess that's the tools he already knew how to use...

murat has a good point, ogre-2.1 is the "not so far" future
so it would be great if it will recieve the proper attention/attraction/tools/whatever it needs to be the best...

with one cute toddler... then two...hopefully three...things will only get slower( on ogre's side... ) and anyway you guy have already have your names carved in Ogres "hall of eternal fame", thats why i think others should also step-in if they wish, not instead of but in addition-to...extra juce...just my opinion....

PEACE IN...

Wolfmanfx wrote: Anyway thumbs up for paroj providing us with PR which are all in my pipeline (to review) I am reviewing all of them - but as you guys know David and I got babies which need love, care and time also Matias is pretty busy
with his awesome new game you should check it out , so things slowed down.

Cool, I know how that is! Good luck and thanks to all who put their time into maintaining Ogre

frostbyte wrote:i also don't see a problem with this fork...( although i do wish it was a direct fork sharing same repository... )

There's no problem with this specific fork, but forks can become a problem over time. But yeah, yeah same here on the repo front / repo war junk. But paroj's fork, paroj's choice This is clearly the best code for new users targeting mobile right now, or even users in later stages of production with 1.9 who want a painless bump to 1.10

Murat pretty much summarized my line of thought.

I'm also ok with a fork. That's the beauty of Open Source. If you don't like the direction, you can move and start your own.

However paroj is still actively contributing back and personally I think it's awesome his fork keeps getting in sync and works as a playtest of "what happens if the PR gets merged?" with automated testing checking if it compiles everywhere.

We're also getting to see the "what if" if we would have taken a different route, particularly not merging a controversial PR.

Murat and I are looking into reverting that commit. It's causing far more headaches than anything.

Thyrion wrote:

Wolfmanfx wrote:also Matias is pretty busy
with his awesome new game you should check it out , so things slowed down.

I would if i could!
where is it? what is it?

Alliance AirWar.
It's an iPhone/iPad game running on Metal. It's not running on Ogre so I couldn't advertise it here with its own post, though we reused a lot of Ogre components: Math, ParticleFX plugin (revamped for performance and memory usage to suite our needs), some of the Ribbon Trail math, the code to convert meshes in to 16-bit floating point and normals into QTangents, Offline Hlms shader generation (to deal with all the feature permutations).

@pavel

I like what you did with the tests and the CI maybe you want help me todo the same for our repo?

to clear some things up: the git vs. hg discussion is not the driving force behind the fork.
I chose git because of github because of the Travis CI integration which is really the backbone of my CI efforts.
It reliably handles the iOS/ OSX/ Linux and Android builds which you can not say about Appveor which regularly manages to timeout the MSVC build.

What really drove me to call my efforts a fork is the mentioned controversial PR (originally #555): it was immediately noted by me and by others that this would
cause trouble, but it was merged nevertheless.

I really appreciate that Murat reviews each of my commits before they get pulled, however it is rather pointless if others just merge huge changes without
even thinking about it.

It was the right thing to back out #588 - however there were also many bad commits in #579, and the GL crashing after #580 is still there AFAIK.
But now that you merged #562 on top of everything resolving this with hg will be pain..

Regarding CI, here you go: https://bitbucket.org/paroj/ogre/commit ... at=default
but as I said above; what you really want is Travis..

I think you missread the hg log - we took the steps and went back in the hg history to a commit which is safe and from this commit Eugene cherry picked the good parts and merged his work.
So all the bad things will go away - also I did not merged Eugene's work it was Eugene our newest team member

Wolfmanfx wrote:Eugene our newest team member

The team page could do with updating on the main site

I think this fork has proved the github eco system is better for the project. I think if anything did move it should only be 2.1+ though.

Wolfmanfx wrote:I think you missread the hg log - we took the steps and went back in the hg history to a commit which is safe and from this commit Eugene cherry picked the good parts and merged his work.

yes, I indeed misread the log. Eugene went back to the same commit the github fork is based on, so I will try to rebase the github version on what is in default now.

@Zonder right now I am working on the new website...

Zonder wrote:

Wolfmanfx wrote:Eugene our newest team member

The team page could do with updating on the main site

He's been a team member for 2 days
And... oh yeah. We'll be updating the website soon. So far Murat's been doing most of the work

Eugene has been doing a great job so far. I have rebased my fork on current default and created PRs as appropriate.

I have however not created PRs for

1. remove dependency on OIS for SampleBrowser/ Tests and use SDL2 instead if available
2. GLSupport module that can create GLES contexts everywhere

because I assume Murat would not merge that. If I am wrong I can still provide those

Hi,

2 can be relevant. 1 was not complete as I remember (mobile were missing) or I am wrong?

created pull requests for both.

however both need more testing as they affect more platforms then I can test on my own (even with CI).

the github fork comes now with batteries included:
by default the core Ogre Dependencies are build: zlib, zzip, freetype. No need to mess with the ogredeps repro and as a bonus you get the latest upstream versions with all the security fixes.

furthermore the ubuntu build was updated from 12.04 to 14.04.

To prevent further confusion I am closing this topic now.

You can check updates from the fork by following it if you like.

To conclude: paroj created a fork which is not an official one, but a personal fork which is not supported by the team.

For official versions, check this page http://www.ogre3d.org/about/what-version-to-choose

[Edit Comments]
Why I closed the topic?
As matias pointed out that ppl could take it personal. I will explain why:
+ We already stated that we are ok with forks (there hundreds out there)
+ I am not ok to start a thread as stable replacement and use this thread as dev blog (posting commit logs there) because its just not true. Imagaine 10 ppl start a thread
everybody states our fork is the stable one and starting posting their refactoring work - what user should think? Think of the newcomers.