July 2007 - Oberon - mailman3.inf.ethz.ch

New BB/AOS release & AosUsbOhci
by PETER EASTHOPE 24 Jul '07

24 Jul '07

Folk, At Mon Jul 23 10:56:45 2007 Ulrike wrote, a new crazy-fresh release of BB/AOS ... OK, Thanks. Previously I installed the 2007-04-03 crazy and found it failed to support either of my USB flash stores. Sven made a small revision to turn on bus mastering. It allowed both of my USBs to work. Thanks Sven! Disappointingly, the revision is absent from the crazy. Good news would be that the omission was an oversight. Medium news would be that "My dog chewed up the only copy we have." Bad news would be that the revision could not be used because it caused another problem. Patching is a nuisance. Can the revision be in the next crazy? Thanks, ... Peter E. http://carnot.pathology.ubc.ca/

1 0

Has BB/Aos got this quirk too?
by Chris Glur 23 Jul '07

23 Jul '07

It's the trivial 'falling in a hole' which eventually makes potential new users of ETH-oberon abort, before they have sufficient fluency to need to keep using it for the superb text manipulation facilities. Has BB/Aos got the following quirk too:- you might use a non-gadgets-editors as your default; eg. Edit.Open, ET.Open because it's needed for TextMail.Cite, TextMail.Send, News.Send, you don't want auto-indent ...etc. ? Then you will often change frames' titles to save as a 'different' file. This reflex action, of eg. just wiping & pasting a new frame-title, doesn't work for gadget-type-frames. The newly entered frame-title needs to be 'terminated'. eg. by <F9>. Some say it's like terminating a command line with <enter> ? Another apparent triviality of N-O which is actually an important principle of HCI, which is part of making N-O a winner, is that 'all commands should give a feed-back-confirmation'. Therefore normally the failure of the frame's name to change will be shown in the log during Store. OTOH if the old/pre-changed-attempt file has no *.Bak, you may have lost a file by an unintended overwrite !! Also when this quirk is combined with eg. learning how to use the ChangeFontTool, which happens to cover the log, it's just all too much. So what do we learn from this: this boring triviality is difficult to describe, so it won't be mentioned; so potentially new users will fall-in-the -hole and decide to abort using ETH-oberon. My money says Aos/BB's advanced technology hasn't fixed this ? BTW. the need for a pre-emptive OS is mostly needed to internet access. Linux-ETH-oberon is a great improvement over N-O for inet access. So what other advantage would I get from Aos/BB ? What is the size difference between N-O & Aos/BB ? I had a shock to read that just one of the file of the Aos-port-to-ARM was over 11MB. Yes, *.pdf descriptions can be big, and the images are not always justified. Thanks for any info, == Chris Glur.

1 0

New BB/AOS release
by Ulrike Glavitsch 23 Jul '07

23 Jul '07

Dear BB/AOS-users, a new crazy-fresh release of BB/AOS is ready for download under http://www.bluebottle.ethz.ch. It contains improvements in the new GC. The awaiting queues of active objects will now be marked during GC and a major bug has been fixed. Please still use this release with caution and report any problems you encounter back to me. I wish everybody good luck with this release! Ulla Glavitsch -- ETH Zürich Ulrike Glavitsch Eggler Institut für Computersysteme Clausiusstrasse 59 CH-8092 Zürich Tel.: +41 44 632 70 49 Fax : +41 44 632 13 07

1 0

oberon literate programming tool
by Eric 15 Jul '07

15 Jul '07

Hi all, Does anyone know of a literate programming tool for Oberon? E.

2 1

oberon on nokia n800, n770 internet tablets
by Antony Tersol 15 Jul '07

15 Jul '07

The nokia internet tablets look intriguing as platforms for Oberon, especially as they are linux based. What would be involved with getting Oberon to run on them? Perhaps among the possibilities - 1. a port of the version for Strong-ARM processors. 2. a port of x86 Linux oberon 3. Native Oberon for Linux 4. OberonScript (http://www.ralphsommerer.com/obn.htm) - Ralph Sommerer's "scripting language and runtime system for building interactive Web Client applications." Anyone thinking / working on this? Regards, Antony n770 cpu: ARM-based TI 1710 OMAP, ( ARM9 <http://en.wikipedia.org/wiki/ARM_architecture> + C55x <http://en.wikipedia.org/wiki/Texas_Instruments_TMS320> DSP) running at 250MHz os: Maemo <http://www.maemo.org>, a custom version of Debian <http://www.debian.org/> GNU/Linux n800 cpu: OMAP2420 - 330 MHz ARM11 <http://en.wikipedia.org/wiki/ARM_architecture> + 220 MHz C55x <http://en.wikipedia.org/wiki/Texas_Instruments_TMS320> DSP + PowerVR <http://en.wikipedia.org/wiki/PowerVR> MBX 2D/3D Graphics Accelerator + IVA os: Maemo <http://www.maemo.org>, a custom version of Debian <http://www.debian.org/> GNU/Linux

1 0

New BB release II
by Ulrike Glavitsch 12 Jul '07

12 Jul '07

Dear BB/AOS users here is the test program that I forgot to attach to the previous mail ... Best regards, Ulla Glavitsch -- ETH Zürich Ulrike Glavitsch Eggler Institut für Computersysteme Clausiusstrasse 59 CH-8092 Zürich Tel.: +41 44 632 70 49 Fax : +41 44 632 13 07

2 1

Re: [Oberon] Unix Port of Aos
by Necati Ecevit 02 Jul '07

02 Jul '07

One of the ways without compiling: make a symbolic link; $> sudo ln -s /lib/libc.so.6 /lib/libc.so -Necati Ecevit.

1 0

Unix Port of Aos (Linux/Solaris/MacOS)
by Guenter Feldmann 01 Jul '07

01 Jul '07

Hi Oberoners, The second alpha release of UnixAos is now ready for download. It exists in three versions: LinuxAos, SolarisAos and DarwinAos (MacOS X). You will find them in ftp://ftp.informatik.uni-bremen.de/home/fld/oberon/UnixAos/* This release is based on the Bluebottle sources of 21.04.2007 and uses the new object format. Currently the Oberon system is used as user interface. The Bluebottle windowmanager modules are also present but the windowmanager does not get installed yet. Knows problems: - The code for invoking user supplied exception handlers is missing. You get a standard trap view with a corresponding message instead. - The Paco compiler gets slower with every compilation. - Module AosUPD contains mostly dummy routines and is unuseable. It's only there because it was needed for compiling other modules. But the old module NetSystem works and ssh and telnet are usable. -- Guenter

2 2

Unix Port of Aos
by R. Clayton 01 Jul '07

01 Jul '07

On an x86 debian testing system (updated weekly), the second alpha release of UnixAos doesn't find libc.so: $ ~/projects/Aos/package/bin/aos UnixAos Boot Loader 03.04.2007 Unix.dlopen: loading library /lib/libc.so failed Kernel.CallBodies: Initializing Threads Kernel.CallBodies: Initializing Reals [ and so on ] Given that $ locate libc.so /lib/i686/cmov/libc.so.6 /lib/libc.so.6 /usr/lib/libc.so $ poking around in src/x86/Linux/Unix.Mod suggests that replacing libc := Dlopen("/lib/libc.so", 2); with libc := Dlopen("libc.so", 2); should fix it, but I don't know how to incoproate the change into the executable to test it.

3 2

Finding libc.os for aos on linux.
by R. Clayton 01 Jul '07

01 Jul '07

Edit the binary. I was looking for a technique that's a little more convenient, one that doesn't require that I wash my hands when I'm done. But let's give it a try: $ od -c aos.linux | grep 'b /' 0000420 004 \0 \0 \0 / l i b / l d - l i n u 0002740 022 \0 \0 \0 ó 002 \0 \0 $ 223 004 \b / \0 \0 \0 $ od -c aos.linux | grep '/ l' 0000420 004 \0 \0 \0 / l i b / l d - l i n u $ strings aos.linux | grep libc libc.so.6 __libc_start_main $ It seems that "/lib/libc.so" isn't anywhere (I'll admit I haven't done this sort of thing in a while and I'm out of practice). However, it seems that libc has been found anyway, which makes sense because aos runs, which I wouldn't expect it to do with undefined symbols from libc. And, just in case you were wondering: $ nm aos.linux nm: aos.linux: no symbols $

1 0