Any updates about RiscOS on the Raspberry Pi?

That’s odd, because the bootcode.bin in the 23Jun2012 image had the pullups enabled.

bootcode.bin start.elf and loader.bin are unchanged between the 23Jun2012 and 25Jun2012 images. As is the HardDisc structure – only the rom image was updated.

Has anyone tried to get a VNC server running on RISC OS/RPi (without a TV/monitor plugged in to the RPi) so you can connect to it from a laptop?

I realise that this is a pre-alpha and problems are to be expected, but I am getting a bit frustrated with my lack of progress. It (the Pi) extremely inconsistent in its behaviour. Whether this is the Pi itself, the PSU, the assorted SDHC cards I’ve tried (three now), my Apple keyboard, I don’t know. If there was a pattern to it all, I’d be much happier, but I fire it up once and it does one thing; I fire it up again (usually after it hangs up) and it does something entirely different.

I imaged Chris Hall’s latest image on to a fresh SDHC card (thanks for that, Chris, and for all the other ones – most appreciated), turned everything on and it got up to the point of probing DHCP. I pressed ESCAPE at this point to get it to skip and got an abort error of some description instead of a desktop. I rebooted and changed the file type of “SetupNet” to skip running it. One reboot later and I was in the desktop! Good progress, or so I thought. I then tried to get my troublesome Microsoft mouse to work, but it wasn’t playing ball. Unlike the last release I tried, it doesn’t lock up the machine – it just doesn’t respond to any input from it.

I’m not sure whether this is of any interest or might help someone more knowledgable than myself, but I wrote down the USB information returned by my recalcitrant rodent:

USBDevInfo returns:

Device class: 00
Device sub-class: 00
Device protocol: 00
Max packet size: 08
Vendor ID: 045E
Product ID: 077D
Device ID: 0144

1. of configs: 1
   Manufacturer: ‘Microsoft’
   Product: ‘Comfort Mouse 6000’
   Serial number: ’’

USBConfInfo returns:

Current config: 1

1. of interfaces: 1
   Config value: 1
   Name: ‘’
   Attributes: Bus powered, Remote wakeup
   Maximum power: 100mA
   Interface 0.0 class 3.1:2 ’’
   1 HID111 descriptors, country 0
   Type 22; length 364
   3 IN Interrupt 8 bytes 10 frames

Sometimes on booting, it gets stuck when intialising the Display Manager module. If it takes more than a second or so to get past this, I know it’s not going to boot. I also encountered problems with the initialisation of the DHCP module when I had two (yes, two) mice plugged in (I was trying to do some comparison of USB characteristics). If I plugged my Apple mouse in at all, it would load everything up to the DHCP module and then slow down to an absolute crawl (I could switch monitor input to OS X, copy the Debian breezy image from a USB stick and start the ‘dd’ process to image it to another SDHC card and RISC OS would still be booting).

Then, all of a sudden, it started exhibiting the same behaviour as with the previous images I’d tried (an earlier one of Chris Hall’s and Chris Gransden’s) – hanging on loading !+Resources. On one of the rare occasions I got it to get past this point, I had a look inside this, and it looks fairly straightforward – just loads a few modules. I’m not sure why it gets stuck on this (on the three SDHC cards I’ve tried, it’s always this that causes the most trouble).

Ho hum.

@John K, I have been using a Dell keyboard and wired mouse, aside from it being a US keyboard and having @ and " swapped, it works without any problems, although the mouse is quite “straightforward” in that it only has the left, right and center scroll wheel button. Do you have a “simpler” mouse you can test with?

I suspect it’s the apple keyboard – I had the same issue at my demo. Tesco value keyboard and mouse from a fellow meetup geek worked perfectly, same card, same treatment, this was Chris’ first image too. I suggest trying your setup with alternative keyboard and mouse, basic as possible, and seeing if over several attempts it’s more consistent?

Has anyone tried to get a VNC server running on RISC OS/RPi (without a TV/monitor plugged in to the RPi) so you can connect to it from a laptop?

It works but the screen update is very slow. If you want to try it out just drop the contents of this zip file into your ‘PreDesk’ folder. You’ll need to set the default screen mode to 32000 colours.

Then e.g. from linux,

'vncviewer -bgr233 <host>'

I got a blank screen unless I specified the ‘-bgr233’.

@John K, i’ve experienced a similar problem with the !+Resources hang when using a cerulian wireless keyboard and mouse combo. I’ve found that if I boot without the USB dongle plugged in, then plug it in once in the desktop, the mouse and keyboard both work fine. Might be worth a try?

@John K, i’ve experienced a similar problem with the !+Resources hang

but only occasionally and almost always it then proceeds after a few tens of seconds extending the boot up time from 19 seconds to 40 seconds or so. Marginally quicker to switch off and on as it is only occasionally that it happens.

@Vanny @John K

Another !+Resources hang here – and that’s with no USB devices plugged in. The last image I played with (Chris’ 20-June or thereabouts) didn’t exhibit this behaviour as regularly. Very bizarre!

Still, playing around with this and getting to re-learn RISC OS is absolutely brilliant – I started my development career playing around with WIMP BASIC programs in the late 90’s, so I’m thrilled to see so much effort going into getting this up and running.

Sadly I have no C/C++ dev experience, which is a shame, because I’d love to be more productive around Pi development efforts.

Actually, in response to my last, I just left it on !+Resources and it eventually gets to the desktop. It’s still not quite there, but I think a 12 minute boot time (so far) has got to be a record!

“An alternative would be to pick a more modern version of tar (e.g. the GNU one) and add the various RISC OS extensions to that.”

I have been using my port of tar within a frontend that organises bz, gz and xz ports for decompressing and compressing as well. It also has multipart tar archiving, and I think I have the bugs out of that now.
I would like to hear of any problems if you find them.

http://homepages.woosh.co.nz/ron.may/tarctrl.zip

“I have also not been able to get Nettle working (complains about Socketwatch – not sure how to fix). "
You need to install the module.
http://www.archifishal.co.uk/~aardvark/sockwatch.zip

New !NettleSSH is at riscos.info
http://www.riscos.info/packages/NetworkDetails.html#Nettle

Hi all,

I’m really struggling to do something basic – run NetSurf. It all seems to be down to the card I/O issues.

Even using my best SD card that 95% of the time boots to desktop OK, using Chris’ latest Distro, I struggle to get !Boot and !System merged on the SD card.
Then, I struggle to copy NetSurf from the Zip file to the SD card. It seems that whenever I copy a more complex file structure, the system will hang after successfully copying a number of files. It’s not completely random – it will stop on the same files if you copy the same structure – but it’s not the files themselves as I can then manually copy the individual file and start again, until I have duplicated the whole structure.

Then, having merged !Boot and !System, and unzipped !NetSurf to either SD or USB stick, I cannot run the programme. When I click it, it just hangs for several seconds, and then I can navigate the desktop as normal. Nothing in task display, nor icon bar.

Any ideas?? I have a (hopefully) class 10 Transcend SD card on the boat from China to see if this improves things… I really want to be able to get to the web from my RISC OS R-Pi.

I’m really struggling to do something basic – run NetSurf. It all seems to be down to the card I/O issues.

If you’re having that many problems why not switch to running the ‘root’ filesystem from a USB stick/HDD?

That’s how I have mine up and running and since Jeffery’s latest changes to the USB code its been very stable, whereas the SD support on the RPi is known to have issues. Then once the SD support is more mature you can switch back to running from the SD card.

I have been using my port of tar within a frontend

I’ll try that tonight and see how it goes.

"presentation right in front of a group of people at the London Raspberry Jam"

A mention on this blog, for benefit of other non-ROUGOL members who may not have seen it before.

Hi Trevor, sounds like the blogger really liked RISC OS – nice to see its potential for the future when seen through fresh eyes.

If anyone’s interested -here are a couple of blog entries from people attending the Oxfordshire Raspberry Pi Meetup – I was pleasantly surprised that a) my demo worked – thanks for the image Chris ;) and b) the positive reaction to RISC OS – it did blow people’s expectations to see a non-linux OS on the Pi…

http://thekeywordgeek.blogspot.co.uk/2012/06/oxon-raspberrypi-meetup-2.html
http://www.designspark.com/comment/7936#comment-7936

Hello all, I back tracked the link from my blog.

regards.

EDIT – details of second London jam are not as confirmed as I thought. updates fill follow.

If anyone needs it, This image boots from SCSI disk 0 (also the NV messages are removed)

Distro updated to improve Welcome message on start up, see screen shot:

Oooh, that looks nice.

I bought myself a new keyboard and another mouse in an attempt to get my Pi to respond to user input. I went to my local PC World and bought the cheapest, simplest ones I could find (aside from the PC World own-brand stuff). And (drum roll) they both work perfectly with RISC OS. Phew! The mouse is a Logitech M100 (a wired affair, with two normal buttons and scroll wheel); the keyboard is a Microsoft Wired Keyboard 600 (it has volume up/down/mute buttons, plus additional keys for play/pause and a calculator). So thank you to those who recommended ditching my Apple keyboard.

The new keyboard/mouse make no difference to the problems I’ve had with ! + Resource. I can quite easily go ten times in a row with the boot sequence getting stuck there. I decided to do some testing, so got to a command prompt, changed the type of Boot.Choices.Boot.Predesk. ! + Resource to text, rebooted and it went straight to the desktop in no time at all. I then tweaked the !Run file within Boot.Resources. ! + Resource to add the “-verbose” switch to the “Repeat” command to see where it was getting stuck. Unfortunately that doesn’t work at all – it gives a “SWI &4258B not known error”. If I do a simple “Repeat -verbose” from the F12 command line, I get the same error – removing the -verbose switch works fine.

Next, I tried to run each module manually. The one where it gets consistently stuck is “WimpPool” – the others load in next to no time. I did “RMLoad WimpPool” whilst starting to write this message on my Mac, and it did nothing for ages and eventually gave me a “Command timeout” error. Looking at the file information in the Filer, the “WimpPool” file is quite a bit larger than the others – around 3MB.

One other thing I noticed – pressing CTRL+BREAK either causes RISC OS to freeeze up, or it gives me an error dialogue and then freezes up.

One issue I’ve had is that if the SD card doesn’t work properly, the file system can get corrupted such that it then doesn’t work even if the SD card is then fixed. In my case that was aborts in loading !+Resource (because that has a 3.3MB WimpPool file in it which stresses the SD system a bit). When a fresh boot sequence is started for the first time, it does lots of writes to initialise parts of the structure. So worth re-imaging your SD card to ensure it hasn’t got corrupted somewhere along the way.

I have been using my port of tar

Well I had to go and install ConfiX first (probably needs an explicit check for that, as the error I was seeing wasn’t very helpful)

Trying to extract the BCM2835 source archive it seemed to be moving along nicely, long files were handled okay and filetypes were being set correctly. Unfortunatly after it had been running for quite a while it finally crashed with a ‘seg fault’ error.

The crash output was….

Stack backtrace:

Running thread 0xadd98
  (  207f38) pc:    93658 lr:    93be0 sp:   207f3c  __write_backtrace()
  (  207fa0) pc:    9377c lr:    6d0f4 sp:   207fa4  __unixlib_raise_signal()
  (  207fb0) pc:    6cff8 lr:    6ddac sp:   206e7c  __h_cback()

Stack backtrace:

Running thread 0xadd98
  (  207f38) pc:    93658 lr:    93be0 sp:   207f3c  __write_backtrace()
  (  207fa0) pc:    9377c lr:    6d0f4 sp:   207fa4  __unixlib_raise_signal()
  (  207fb0) pc:    6cff8 lr:    6ddac sp:   206e7c  __h_cback()

  Register dump at 00207fb4:

    a1:    9ef01 a2:        0 a3:        0 a4: fe000000
    v1: e59b1020 v2:    ac9a8 v3:    a4964 v4:        1
    v5:        0 v6:    b6a00 sl:   206208 fp:   206e90
    ip:   206e94 sp:   206e7c lr:    6ddac pc:    6eab0
    cpsr: a0000110

  0006ea9c : .... : 1a000003 : BNE     &0006EAB0
  0006eaa0 : ...ê : ea00000b : B       &0006EAD4
  0006eaa4 : P@“å : e5944050 : LDR     R4,[R4,#80]
  0006eaa8 : ..Tã : e3540000 : CMP     R4,#0
  0006eaac : .... : 0a000008 : BEQ     &0006EAD4
  0006eab0 : .0“å : e5943000 : LDR     R3,[R4,#0]
  0006eab4 : þ.Sã : e35304fe : CMP     R3,#&FE000000
  0006eab8 : ùÿÿ. : 0afffff9 : BEQ     &0006EAA4
  0006eabc : .. á : e1a00004 : MOV     R0,R4

  (  206e90) pc:    6ea78 lr:    6ddac sp:   206e94  __newstream()
  (  206eb4) pc:    6dd08 lr:    654a8 sp:   206eb8  fopen()
  (  206ed0) pc:    65480 lr:    21b38 sp:   206ed4  getpwnam()
  (  206eec) pc:    21ae0 lr:    1e81c sp:   206ef0  uname_to_uid()
  (  206f30) pc:    1e32c lr:    1f470 sp:   206f34  decode_header()
  (  206f88) pc:    1f2c4 lr:    29e24 sp:   206f8c  read_and()
  (  206fec) pc:    2928c lr:    7b27c sp:   206ff0  main()
  Register dump at 00207fb4:

    a1:    9ef01 a2:        0 a3:        0 a4: fe000000
    v1: e59b1020 v2:    ac9a8 v3:    a4964 v4:        1
    v5:        0 v6:    b6a00 sl:   206208 fp:   206e90
    ip:   206e94 sp:   206e7c lr:    6ddac pc:    6eab0
    cpsr: a0000110

  0006ea9c : .... : 1a000003 : BNE     &0006EAB0
  0006eaa0 : ...ê : ea00000b : B       &0006EAD4
  0006eaa4 : P@“å : e5944050 : LDR     R4,[R4,#80]
  0006eaa8 : ..Tã : e3540000 : CMP     R4,#0
  0006eaac : .... : 0a000008 : BEQ     &0006EAD4
  0006eab0 : .0“å : e5943000 : LDR     R3,[R4,#0]
  0006eab4 : þ.Sã : e35304fe : CMP     R3,#&FE000000
  0006eab8 : ùÿÿ. : 0afffff9 : BEQ     &0006EAA4
  0006eabc : .. á : e1a00004 : MOV     R0,R4

  (  206e90) pc:    6ea78 lr:    6ddac sp:   206e94  __newstream()
  (  206eb4) pc:    6dd08 lr:    654a8 sp:   206eb8  fopen()
  (  206ed0) pc:    65480 lr:    21b38 sp:   206ed4  getpwnam()
  (  206eec) pc:    21ae0 lr:    1e81c sp:   206ef0  uname_to_uid()
  (  206f30) pc:    1e32c lr:    1f470 sp:   206f34  decode_header()
  (  206f88) pc:    1f2c4 lr:    29e24 sp:   206f8c  read_and()
  (  206fec) pc:    2928c lr:    7b27c sp:   206ff0  main()

I hadn’t tried the BCM2835 archive until now, but I’m only getting as far as
BCM2835Dev/castle/RiscOS/Sources/FileSys/FSLock/rm/.cvstag
and then getting a Segmentation fault crash also
It has been handling .cvstag OK previously so I dont think it is the naming.
I’ll try a few things and see if I can find the problem.
Thanks for letting me know.

OK, regarding !TarCtrl and the BCM2835 archive, I doubled the wimpslot size from 2048K to 4096K in the !Run file and it went OK.
The resulting textwindow after !StrongEd exited the taskwindow is 1,575,043 bytes, so this may have something to do with it.
I have had smaller jobs run OK in as little as 800K wimpslot, but Multipart tars needed at least double that so I thought 2048K would have been ample. I’ll set it at 4096K and make a note in a ReadMe for someone who might like to use it on an old machine with limited memory.
Note that at the moment !TarCtrl uses the system CSD for direction so any changes to the CSD while it is running would upset the applecart.
Sometime I may try DDEUtils Prefix out, as it can set a working directory on a per task basis.