Dingoonity.org

Quote from: dcervi on May 04, 2013, 10:54:54 am

Well, if the OS flashing routines are causing the bug, there should be no problem in making a proper working SD card image, and replicate it, while the developers work on solving the bug for future OS updates.

This bug shouldn't cause more delays with delivery.

That could have worked, but from what I know the SD cards have already been inserted in all the devices.

It looks like all that needs to be done, is for the user to manually replace the damaged rootfs file with a working one.
Is that doable for your general end-user? It sounds like it can be done with FTP.

The Kickstarter units had the SD cards installed in China. The SE's arrived with 16gb cards but the 32gb cards needed to be swapped. That is what we were told. I don't know if the 32gb cards have been swapped out yet or not.

What if the flashing process requires them to open up the units and remove the cards anyway, regardless of installation?

pcercuei, my most recent post was made with Justin and it reflects his feelings as well as mine, or to phrase that in a different way- it reflects the feelings of your CEO and COO. As the Chief Operations Officer, once I have an assignment from my Chief Executive Officer it is my duty to follow through with it because that is the person I answer to and I don't require any other discussion after that.

Um this is a community project aimed at developers, a statement like that is about the most damaging PR possible. If you're trying to play grown-ups be a PR man then you made just about the worst mistake possible.

If you want to scare away developers and piss off volunteers that's the way to do it, not to mention that you just claimed you'd say anything you were told to which now makes your voice completely meaningless.

No. The flashing process is done through USB.

Does this not mean the SD cards can be repaired via USB anyway? I'm just curious as I've become pretty invested with anticipation.

Another update from the debugging front:

I think I found the source of the corruption. That is, I found a suspect piece of code and if I change it to look more like other code, the corruption disappears. However, I want to fully understand what the code does before declaring that I fixed the bug. I've never had to dive so deep into the ext4 code before, since usually it just works.

Regarding the audio stutters, it seems they are not related to the corruption, but caused by the known problem in the SD controller driver (hogging the CPU on writes) combined with an ext4 feature that does delayed initialisation in the background. Normally those background writes are not very noticable, but because of our driver we get stutters on writes. We can work around this by disabling the delayed initialisation. This means the first boot after flashing takes about 5 minutes longer to resize the data partition (measured with a fast 32 GB card), but once it's done resizing it's actually done and doesn't have to write stuff in the background anymore.

A little longer spent setting up the system for the first boot should be worth it, if it saves a few headaches later on.
The extra wait doesn't look like it'll concern the end-user when they receive their units though, just those setting up the units for export.

Ooo, looks like I've missed some drama lately! Good to know it sounds like things should be resolved quickly - just hurry up and ship the damn units already! XD

does this mean what i think it means?

Why are there no "official" updates from PR when good things happen?  PR isn't just Damage control.


Anyway devs you rock.  Those YouTube videos are not quite 60fps (I think YouTube videos are 30) but it looks butter smooth.  Can't wait to play it.

Quote from: Sensei on May 08, 2013, 02:07:35 pm

Why are there no "official" updates from PR when good things happen?  PR isn't just Damage control.


Anyway devs you rock.  Those YouTube videos are not quite 60fps (I think YouTube videos are 30) but it looks butter smooth.  Can't wait to play it.

As far as the 16bpp demonstration video goes, Justin probably won't release an official update about that through the kickstarter because of the nature of emulators, he is much more focused on displaying legitimate emulation like being able to buy the BIOS files you need for some emulators, or the runtimes for games like Duke Nukem 3D Atomic Edition through GOG.com.

True but it's still a HUGE accomplishment along with USB and the sound stutter being fixed. Why there hasn't been any official announcement on any of this is beyond me. Now that the issues are all fixed, let's get those units flashed and sent out eh?