April 2014 - Oberon - mailman3.inf.ethz.ch

Fwd: Oberon Technical Notes / ETH Tech Reports Site
by Jack Johnson 25 Apr '14

25 Apr '14

Thanks, Paul! ---------- Forwarded message ---------- From: Paul Reed <paulreed(a)paddedcell.com> Date: Fri, Apr 25, 2014 at 7:33 AM Subject: Re: [Oberon] Oberon Technical Notes / ETH Tech Reports Site To: Jack Johnson <knapjack(a)gmail.com> Hi Jack, > I'm trying to recall, it seems like back in the early V2 ports days, > there was something strange like everyone was implementing their own > mark and sweep garbage collector, but everything else was documented > enough to get by. Does that ring a bell, or am I crazy? No, you'll be pleased to know, you're not crazy. :) It took me a little time to find the old "Oberon Technical Notes" paper online. This was very useful to me when I was trying to build a DOS/Windows compiler based on Project Oberon in the dying years of the last century, and way before I knew anyone at ETH. The technical notes are hidden under a bushel, invisibly tacked on to the 1991 paper "On the linearization of graphs and writing symbol files", now available along with a bunch of other stuff at http://www.inf.ethz.ch/research/technical-reports/1990-1999.html which is an amazing resource I REALLY wish was available back then. Since you emailed me directly on this (you're not the only one!), I can't with respect reply including the oberon list, where it might be helpful for others. Therefore if you agree I'd be grateful if you'd forward my reply to the list (including this comment!) :) Thanks, Paul

1 0

Math.Mod / compile rrror (to many digits)
by Volkert Barr 24 Apr '14

24 Apr '14

Dear all, i tried to compile the Module "Math.mod" (12.10.2013, ../wirth/ProjectOberon/Sources/Math.Mod.txt) and got a dozen "to many digit" compile errors . Looks like some constant assignments are to long. My question: Is the bug in Math Module or is the compiler wrong here? I use the Oberon-Image (RISC.IMG, 28.03.2014) from Paul Reed´s "Project Oberon"-Website with the nice "Oberon RISC Emulator" from Peter De Wachter. BW, Volkert

5 10

Re: [Oberon] Oberon V4 availability (was: Project Oberon 2013 on GitHub?)
by Chris Burrows 23 Apr '14

23 Apr '14

> -----Original Message----- > From: Paul Thomas Melville [mailto:ptmelville@gmail.com] > Sent: Tuesday, 22 April 2014 9:54 PM > To: ETH Oberon and related systems > Subject: Re: [Oberon] Project Oberon 2013 on GitHub? > > It won't work under ARM. Oberon V4 works only on x86. ??? While I am not aware of any port of Oberon V4 to (one of the many variants of processors designed by) ARM, Oberon V4 is available on many other architectures apart from x86. These include the Amiga, Atari, DecStation, HP700, 68000 Mac, PowerMac, IBM RS6000, Silicon Graphics (SGI), Sun Sparc etc). All of these can be downloaded from the ETH Oberon V4 ftp site: ftp://ftp.ethoberon.ethz.ch/Oberon/OberonV4/ Alternatively some are available from the Linz Oberon V4 server: http://www.ssw.uni-linz.ac.at/Oberon.html Regards, Chris Burrows CFB Software http://www.cfbsoftware.com

4 4

Re: [Oberon] Compiler Modules Empty?
by Nemo 23 Apr '14

23 Apr '14

Jan: On 12 April 2014 19:48, Jan Verhoeven <jan(a)verhoeven272.nl> wrote (in part): > I have started reformatting the sources on a new webpage: > http://fruttenboel.verhoeven272.nl/projob14/index.html Are you reformatting by hand or using a pretty-printer? N.

3 2

Re: [Oberon] Pseudo-module SYSTEM (was RISC5)
by Chris Burrows 22 Apr '14

22 Apr '14

> -----Original Message----- > From: eas lab [mailto:lab.eas@gmail.com] > Sent: Monday, 21 April 2014 2:03 AM > To: ETH Oberon and related systems > Subject: Re: [Oberon] RISC5 > > SYSTEM.MOVE(SYSTEM.ADR(item.data(*.data*)[24]), > SYSTEM.ADR(item.data(*.data*)[14]), NetIP.AdrLen); > > But where did the author select this instruction from? >From the documentation supplied with the particular implementation of Oberon that they were targeting / using. > Watson doesn't show the exports of SYSTEM-module. > Watson cannot know about *pseudo-modules* like SYSTEM whose only manifestation is within the Oberon compiler itself. It can only report the information for normal external modules if a 'symbol file' is available for that module. NOTE: Watson is not available on all versions of Oberon. It is useful when it is available but you should not depend on its existence. Regards, Chris Burrows CFB Software http://www.cfbsoftware.com

2 1

Re: [Oberon] Pictures of Ceres Workstation?
by Claudio Nieder 19 Apr '14

19 Apr '14

Hi, > The only pics i found are Have a look at the images here: http://www.retrocomputing.net/parts/e/eth/ceres3/ claudio -- Claudio Nieder, Talweg 6, CH-8610 Uster, Tel +4179 357 6743, www.claudio.ch

2 1

Creating SD card image - RISC.img
by Paul Onyschuk 17 Apr '14

17 Apr '14

Since Peter De Wachter done such great work with emulator, next logical step in my opinion would be to get it fully self hosted: creating new RISC.img under emulator and exporting it via PCLink, then booting from new image. There is information about boot linker module called ORL in section 14.2 of Project Oberon, but it seems that code for it isn't provided. Are there plans to change it, is there any ETA for it? I've seen some previous discussion about it on mailing list, but that was before availability of emulator. Listing partition table gives interesting results: $ fdisk -l RISC.img Disk RISC.img: 268 MB, 268862464 bytes 255 heads, 63 sectors/track, 32 cylinders, total 525122 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0x00000000 Device Boot Start End Blocks Id System RISC.img1 * 2048 524287 261120 6 FAT16 RISC.img2 524288 655359 65536 ff BBT I would like to ask Paul Reed about details involving creation of image (since current images are provides by him as far I understand), but I'm afraid it could be asking too much at the time. -- Paul Onyschuk

5 9

_Non_sequitur_ in ConfigInput?
by peter＠easthope.ca 17 Apr '14

17 Apr '14

Hi, This is meant to be more a question than a criticism. If I'm on the right track, a tiny clarification in source code is possible. Certainly no offense to pjm is intended. This is from ETH Oberon / PC Native 05.01.2003, the last alpha release. MODULE ConfigInput; (* pjm *) ... BEGIN ... Kernel.GetConfig("PS2", mp); ps2 := mp # "0"; ... END ConfigInput. The PS2 configuration string can be "" or "0" or "1" for detection of a PS/2 mouse being unspecified or inhibited or requested. mp represents the mouse port number for a serial mouse with possible values "", "0", "1", "2" & "3". 'Kernel.GetConfig("PS2", mp);' is peculiar. Although all the values are strings or arrays of characters, the flag for detection of a PS/2 mouse being put where a port number should be, is confusing. What about Kernel.GetConfig("MP", mp); Kernel.GetConfig("PS2", ps2); IF ps2 = "" THEN ps2 := "1"; ? An additional statement is needed but the intention is more clear. If someone can offer a better explanation, thanks, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ New address Bcc: peter at easthope. ca

1 0

Re: [Oberon] RISC5
by eas lab 17 Apr '14

17 Apr '14

joerg.straube wrote:- ]Depends on the definition of "emulated" vs "implemented" :-) ] ]Couldn't you do the following analogy? ] Verliog code = PdWs emulator code ] Verilog tools = C compiler ] Spartan 3 = host environment Yes, that's how I see it too: the Verliog code is 'compiled' by Verilog tools into the <physical configuration> of the silicon, resulting in the register-set & buss/s which implement the [partial] V5 system. The hidden/mysterious part for me is the display/textFrames. Does V5 drive standard VGA? Where's the code showing; VGA -> FrameBuffer -> ETHOviewer. When I built 8 bit P-code interpreters, [wire wrapped! I've still got a 68?3xx one which I failed to get working], I just output a stream of bytes to the mechanical TTY. And the CPM-based version used the BIOS-like <outChar(char)> function, IIRC ? But that was just a toy compared to V5 with TextFrames. ==Chris Glur.

3 2

Re: [Oberon] FP
by Peter De Wachter 15 Apr '14

15 Apr '14

No, that wouldn't work. Every architecture does floating point differently. And RISC FP in particular doesn't implement the full IEEE standard, for example it doesn't support infinities and denormal numbers. The current code in the emulator isn't pretty or fast but at least it's accurate. > float *a = (float *)(&x); And don't do that, that invokes undefined behavior. To emulate C++'s reinterpret_cast, use an union: union { uint32_t x; float f; } float_hack; float_hack.x = x; return float_hack.f; On 15-04-14 13:17, Jörg wrote: > Hi all > > Indeed. The FP code in the emulator is an example of "backporting assembler > to C". PdW could have taken the "normal" FP arithmetic offered by C and let > the C compiler decide to generate the code. > The FP format NW uses is the same as C, namely IEEE single precision. > > So, PdW could have written something like: > uint32_t fp_add(uint32_t x, uint32_t y, bool u, bool v); > { > float *a = (float *)(&x); /* aka reinterpret_cast in C++ */ > float *b = (float *)(&y); > > /* do whatever you want to do with the FP numbers a and b */ > float sum, prod; > sum = *a + *b; > prod = *a * *b; > ... > } > > Jörg

5 4