[delinend]

Anyone knows, when we can test and use Rbej's Gotham builds Again, with all Add'on's from the repo. ?

[javaboyuk]

@popcornmix / @abej
(I found hdmi_force_open=1 courses my rpi to crash if I try to skip forward or jump to next track :-()

Certainly getting the equivalent of hdmi_force_open working is on my list (ideally from "keep receiver alive" option in GUI).
That's not the same as gapless (which actually does fading between tracks), but should improve music playback.

or use two players one for video one for audio only, is there some issue using the original xbmc player for
rpi for audio only, as that might be a very simple fix? ie have it loaded and compiled but only used if you
use the the 'advancedsetting.xml' to choose it for audio, assuming you could then have a 0.5 second or
mix setting or something to avoid the hdmi drop on rpi?

Going back the hdmi_force_open equivalent, to be honest that would be good enough for me!
(based from my experience of using the "hdmi_force_open=1" for play back)

Any idea when you might get some time on it?

Can I help you in some way?

JB

[doveman2]

Frodo Branch Updated

- updated firmware

- [rbp/omxplayer] Add caching of states to reduce number of round trip delays to GPU

- [rbp/omxplayer] Add wait for buffers done calls

- [rbp/omxplayer] Handle timeouts greater than a second

- [rbp/omxplayer] Callback context pointer is not needed once in the class

- [rbp/omxplayer] Add timeouts to buffer calls

- [rbp/omxplayer] Remove warnings from event logging calls

- [rbp/omxplayer] Add more event logging

- [rbp/omxplayer] Signal bad state on decode error

- [rbp/omxplayer] We should block until buffers are released

- [rbp/omxplayer] Load OMX dll just once

- [rbp/omxplayer] Remove component locking

- [rbp/omxplayer] Avoid submitting too large audio buffers

- [rbp/omxplayer] Fix bugs with jpeg parsing

http://netlir.dk/rbej/builds/

http://lysin.me/rbej

After updating to this, my RPi just gets stuck on the OpenElec splashscreen. It corrupted the SD card as updating always does for me (I'll try getting a different SD card soon and see if that helps), so I put it in my PC and copied across the files that way, as I've done before (everything from bootloader/ except I kept my config.txt) and SYSTEM and KERNEL renamed to kernel.img but whilst that normally works, it isn't now for some reason.

[popcornmix]

I've been working on some jpeg decode/encode optimisations.
https://github.com/popcornmix/xbmc/commits/newclock3
I repeat: this is for testers only. It is probably not stable.

Another update. This now uses a jpeg->resize->gl texture pipeline on the GPU, so the ARM is largely isolated from processing fanart/posters.
https://docs.google.com/file/d/0B-6zmEDJ...sp=sharing

It defaults to 16bpp textures, but there is a settings option to switch to 32bpp.
On a 256M Pi I'd recommend leaving it at 16bpp, and on a 512M pi, 32bpp with gpu_mem=256 should be fine.

I've also added a gui option to remove the 720p GUI limit. Again, I'd leave that on for a 256M board, but removing it (with gpu_mem=256) for a 512M Pi.
Note currently the switch between 720p limit and no limit doesn't fix up guisettings.xml. EIther delete it and reboot, or edit the overscan settings in the resolutions group.
(using the calibration option may well work).

See here for a video:
http://forum.xbmc.org/showthread.php?tid=174485

I repeat: this is for testers only. It is probably not stable.

[doveman2]

Hmm, it seems this time the update corrupted my USB as well as the SD, which I thought it shouldn't be able to do and is the main reason for using USB.

I'm not sure if the Linux partition is corrupted as I can't view that in Windows but the FAT32 partition before it certainly is and the RPi says "Can't access /dev/sda1".

EDIT: Well I don't understand. After spending an hour trying to recover the USB, which I think had a 100MB FAT32 partition and the rest the ext4 /Storage partition, I just have a Linux partition that starts at Cyl 32.

If I boot with
boot=/dev/mmcblk0p1 disk=/dev/mmcblk0p2 ssh quiet

and then use df -h or mount I can see the USB is /dev/sda1 mounted as /var/media/USB_16GB yet if I try and boot using

boot=/dev/mmcblk0p1 disk=/dev/sda1 ssh quiet

then it hangs on the boot screen and eventually says "Unable to mount /dev/sda1"!

EDIT: Well after spending all night trying to repair, re-format, copy back files to the USB, it seems there's just a bug with that build which prevents it booting from USB. Switched back to the 15/09 build and it boots fine.

[allan87]

Popcornmix:

Note currently the switch between 720p limit and no limit doesn't fix up guisettings.xml. EIther delete it and reboot, or edit the overscan settings in the resolutions group.
(using the calibration option may well work).

in that quote are you talking about the same thing as this?

- every build I tried since august has messed up the video calibration (the widths of interface elements are distorted) between restarts and has to be manually corrected.

Thanks.

[doveman2]

Well I got to watch 3/4 of one programme before the picture broke up/lost-sync and then XBMC completely locked up on me! http://xbmclogs.com/show.php?id=63750

When I came back about 10mins later, it was on Home but the logs show that it had actually rebooted: http://xbmclogs.com/show.php?id=63748

[Jönke]

@popcornmix
Maybe dumb question but could a feature Rpi firmware make it possible to use only usb (without sd card) as it seems in your post thats usb is faster so why not get rid of it ?

[xbs08]

You can't, RPI needs to boot from the sdcard first.

[popcornmix]

You can't, RPI needs to boot from the sdcard first.

Yes, initial boot (FAT partition) needs to be on sdcard. It can be small (e.g. 128MB) and slow, but it does need to be there.

[Wanderlei]

CEC suddenly stopped working again while I was watching something. Cannot get it working again, last time changing HDMI port tv got it working again but not this time.

[h.udo]

Hi,

@popcornmix, your latest build works fine for the most part. It really feels more snappier.

First problem encountered: Shutdown and reboot don't work. Both kill xbmc and restart it.

I'm using 900/333/450 on a 512 Rpi with 256/256 memory split.

Keep up the great work.

[Milhouse]

First problem encountered: Shutdown and reboot don't work. Both kill xbmc and restart it.

That's been broken in Gotham builds for a while - xbmc.bin segfaults when exiting and consequently it's impossible to power off/reboot. I'm sure it's on the list of things to do!

[h.udo]

Thanks @MilhouseVH

Been away from development branch(es) for quite a while.

[nuttypro69]

Hey popcornmix your latest build It's really more snappier,maybe rbji could use that on his next frodo build since his last one corruated my usbdrive now it is lock useless had to load back Sept 15 build.