I am not aware of any existing GL applications that completely break like this. I have run lots and lots of the top applications on Xoom without trouble. Applications also haven't generally broken on previous platform releases, and I am pretty sure there are many applications on Market that don't use GLSurfaceView.
To be honest, I don't have more time than to just scan through your long rant; it looks like most of what you are trying to do is let off steam than look for anything constructive, so I'll leave it at that. On Sat, Apr 23, 2011 at 6:36 PM, MichaelEGR <[email protected]> wrote: > Greets, > > I finally jumped the shark and got the G-Slate and all my 3D demos > from the simplest to the much more complex result in a blank screen. I > just got the G2x too and have the Nexus S, N1, G1, Droid, etc. and my > engine / platform works great on every OS 1.5+... Auriga3D flies (old > build too). > > TyphonRT, my soon to be released platform many years in the making, > manages it's own GL context and doesn't use the API / SDK provided > GLSurfaceView. As things go I've been in a massive refactoring > operation and can't currently compile the 3D stuff for about a week > more (currently using old builds that run fine on all devices > including G2x). This is the exact same thing that happened with > Android 2.0 w/ horrible OpenGL driver support, unannounced changed > implementations of core OpenGL ES methods (was it the implementation > of eglChooseConfig if I recall correctly you guys changed without > telling anyone). So you guys either changed an implementation of a > fundamental GL method or for some reason my GL attribute system is not > requesting a proper context or falling back correctly even though it > works on every other OS and device combination I have including the > G2x, so it's not a Tegra 2 thing. > > All I got to say is that did you guys even try to test 3D / GL support > using an example / test case that doesn't use the API / SDK supplied > GLSurfaceView. It's obvious things are way buggy or something is not > right when managing ones own context / surface. Remember exact same > code runs amazingly well on all OS versions with a wide range of > devices. > > I'm trying to pour some ice water on the back of my neck, but how can > you guys let such slipshod work get released (I really do appreciate > the work of the Android team, but not the poor quality / rushed > aspect). I hope a 3.0.1 is coming darn soon just like 2.0.1 and you > guys fix the issues just like 2.0.1.. I was delayed 2.5 - 3 weeks > finding workarounds for the 2.0 release. Granted 2.0.1 fixed things > too (6 weeks later; 3 after my workarounds), but guess what my bread > winning client canceled my contract back then due to the unplanned 3 > week delay and I had major hardship. This time I just finished a short > term contract (non GL and something that got _a lot_ of press > recently), so I'm not going to be out on the street because of this > f'up. Can I now assume there will be major GL issues every OS release > with serious GL fundamental API breakage on major point releases > (let's not bring up Froyo and the borked Java bindings, so I guess > it's not just major point releases)? > > Granted again it'll be about a week until I finish up some of the last > major refactoring here and can compile the 3D demos I have for my > engine / platform again and I'll get to the bottom of it and find the > egl error or which method call is misbehaving. I hope it's something > really simple like a wrong attribute combination. My system right now > attempts to grab a context with the following params and backs off to > lower parameters if nothing matches: > > "High Quality" (Droid + really all 2nd gen and up devices) > EGL_RED_SIZE, 5, > EGL_GREEN_SIZE, 6, > EGL_BLUE_SIZE, 5, > EGL_DEPTH_SIZE, 24, > > "Medium Quality" (G1, Hero 1st gen, etc.) > EGL_RED_SIZE, 5, > EGL_GREEN_SIZE, 6, > EGL_BLUE_SIZE, 5, > EGL_DEPTH_SIZE, 16, > > So you guys either changed the implementation to a GL method or when > requesting standard high quality settings the G-Slate freaks / GL > driver freaks out and it thinks it has a proper context, but things > are borked. As I said everything is fine on the G2x / Tegra 2 and > every other device and OS version 1.5+, so this points mostly to an > unannounced underlying GL method implementation change. I'll find it > in a week once I can compile my 3D tests / demos again, but if someone > from Google cares to pipe up and point me to the release notes that > owns up to any underlying fundamental GL API implementation changes > that would be great. You guys didn't do proper release notes for 2.0 > and the GL method implementation changes had to be yanked from yah on > this list, so I can only assume the worst and that this may be the > case again. Now the answer is _NOT_ to use the SDK / API provided > GLSurfaceView. Thanks, but no thanks... I'm glad you guys have built > in workarounds or whatever is making all of your platform 3D tests > pass without examining the larger and more fundamental GL API usage. > Raw OpenGL ES API and custom context management / usage should not > break like this! I seriously want to provide you guys with a non-SDK > GLSurfaceView implementation of a super simple spinning cube and it'd > be nice if it was added to the test suite as clearly you guys are not > testing non SDK oriented GLSurfaceView apps otherwise this would be > caught _EASILY_. > > I'd also be a bit more gracious too if this wasn't the _4TH_ major OS > release to have serious low level graphics API problems. Like I said > this time I didn't lose a client over it, I can't begin to tell you > how professionally unsatisfying it is to have a client blame oneself > for a delay when it is out of ones control and despite trying to be > patient and describe that the problem is in Google's court said client > simply didn't care as all they were looking at was the calendar (not a > very good or understanding client per se especially with early GL dev > & Android circa '09, but such is life). FIX THIS NOW AND STOP f'ing > up; seriously a vote of no confidence is close! Sorry.... fool me > once..... but fool me 4 times and I'm going to yell. While I'm at it > I'll take a dig too; sorry... !!! stuttering on the home page > switching views and scrolling the app tray... SERIOUSLY on a Tegra 2 > you guys can't make this smooth as butter... dizaam.. doh.. ;P Feels > like it's '08 and the G1 scrolling the app tray; fix that I say! If I > can get a non-trivial 3D engine running at 60 FPS on a Tegra 2 device > the app tray / homescreen can work without stuttering! > > Look I'm an Android diehard working with the GL/ES since ver 1.0 of > Android and have been in the s'ht the whole time. I'm not going > anywhere, but I can't stress it more that you guys need to tighten up > testing / release process and do a much better job. It's been half- > a*sed all the way with big problems every release (2.3 amazingly > wasn't borked; good job!) > > Yah.. A rant.. Look I'll be the first apologize if there is a good > reason or something that makes logical sense, but remember not all > clients of us Android devs use anything remotely close to logic / > rational thinking or for goodness sake I'll be launching a major > engine / platform soon I'll be charging money for and if this kind of > thing brings my whole platform down and everyones games / apps, etc. > w/ no easy / immediate fix in the future it'll seriously piss me off > (5th time, 6th, 7th)... when is it going stop and you guys get > fundamental graphics APIs rock solid? It can't be that hard to provide > basic working GL functionality (and no saying use the provided > GLSurfaceView is not the answer). Like I said I could be wrong here > and I'll admit it if I am and apologize to this list. That is about a > 1-2 weeks away before I know for sure when I bring my 3D demos back > online and can debug / step through GL init line by line and check GL > method call / error states. Until then I'm waiting to hear what > underlying GL implementation changed with Honeycomb... :: sigh :: > > I'm OK for now with Diane H's explanation of the Honeycomb / Tegra 2 / > G-Slate 2D API and slowness I see though I'm getting 14-20 FPS with > drawing 1 background bitmap full screen and one moving entity (note 45 > moving enemies w/ basic physics at 55 FPS on the G2x w/ 2D API; yes > screen size is smaller, no Neon, etc! but damn 14-20 FPS with drawing > one bitmap full screen and another bitmap; somethings fishy. Something > seems badly borked, but I'll look into it more before I pass judgement > on the apparent slow 2D API.. The broken GL situation is inexcusable; > this kind of thing would have crashed my whole platform and that is > not cool for the future after I launch and people depend on my efforts > and Android to be rock solid. > > Grrrr... And I close this with a ;P but come on folks... I love and > hate Android... Good job at eliciting such incitement, you're doing > something right, but I plead that you get things right / rock solid > from now on as other peoples businesses and livelihoods are depending > on it! If this is an LG thing, please ignore, I'll yell blindly > towards S. Korea at some point and shake my fists w/ fury; oh yes, > indeed! > > -- > > -- > You received this message because you are subscribed to the Google > Groups "Android Developers" group. > To post to this group, send email to [email protected] > To unsubscribe from this group, send email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/android-developers?hl=en > -- Dianne Hackborn Android framework engineer [email protected] Note: please don't send private questions to me, as I don't have time to provide private support, and so won't reply to such e-mails. All such questions should be posted on public forums, where I and others can see and answer them. -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/android-developers?hl=en
