January 2015 - Oberon - mailman3.inf.ethz.ch

Re: [Oberon] email usage in Oberon systems.
by peter＠easthope.ca 30 Jan '15

30 Jan '15

Chris, From: peter(a)easthope.ca Date: Thu, 29 Jan 2015 08:03:22 -0800 > MailMessages differs from mbox by ending each message with the > three characters "." <CR> <LF>. getmail can invoke a trivial external > filter to append these characters to a message. If the filter appends "." an omits <CR> <LF> there is still a line break between "." and "From " in MailMessages. As if getmail automatically appends <CR> <LF> after invoking the external filter. I don't claim to understand the details but a filter appending just "." seems to work. Also, I haven't fathomed why Oberon includes the "." rather than stick to the established mboxrd format. Also, getmail can operate with TLS. That should cover all the bases, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

1 0

email usage in Oberon systems. Was: Re: A2, targeting a Geode SC3200.
by peter＠easthope.ca 29 Jan '15

29 Jan '15

Chris, From: eas lab <lab.eas(a)gmail.com> Date: Thu, 29 Jan 2015 11:23:15 +0200 > Although I'm very happy with ETH-Linux-Oberon, I can't see how to use > it to solve my painfull problem of needing to use the standard http > interface for my gmail [alternatively to be always on-line]. According to http://en.wikipedia.org/wiki/Gmail#Security , mail can be retrieved by POP3 or IMAP, both under TLS. The POP3 case is the simplest. With UnixAos, getmail can retrieve messages and append to the Oberon MailMessages file, which is a variant of an mbox file. Ref. http://en.wikipedia.org/wiki/Mbox . MailMessages differs from mbox by ending each message with the three characters "." <CR> <LF>. getmail can invoke a trivial external filter to append these characters to a message. The filter is easily made as a shell script or awk script or something else. I've never used ETH-Linux-Oberon but getmail should work just as well there. As far as I know, ETH-Linux-Oberon and UnixAos don't lock MailMessages. (A correction is welcome.) Ref. http://en.wikipedia.org/wiki/Mbox#File_locking If getmail changes MailMessages while Oberon Mail is active, there will be trouble. Therefore getmail should be invoked just prior to starting Oberon and this can be automated using a script. To update MailMessages, you will need to quit Oberon and start it again. If another complication is added, new mail can be retrieved without quiting Oberon. Have getmail retrieve to an mbox independent of MailMessages. Then have qpopper deliver mail to Oberon. The POP3 server cited in Oberon.Text will be localhost. With this arrangement, getmail can be activated manually. Also it can be scheduled by cron. Also it can run just prior to Oberon starting, as in the preceeding paragraph. With getmail and qpopper, you should have "good old Oberon mail behaviour" with the possible additional detail of starting getmail manually. Activation of getmail by incrond might even be possible. I haven't tried it. A2 has IMAP and not POP3. But the A2 client can be configured to retrieve messages and delete from the server. In other words, IMAP can similate POP3 behaviour as a special case. But I suspect that you won't be happy with the windowing interface in A2 and in UnixAos. If you use either, you will prefer to stay in the Oberon subsystem. Therefore IMAP is really not a concern. Regards, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

1 0

A2, targeting a Geode SC3200.
by peter＠easthope.ca 29 Jan '15

29 Jan '15

Hi, This was sent originally about 12 hr ago. No evidence of receipt. Hence, trying another copy. Source repository 5919 here. I'm interested to build an installer for A2 for a machine with a Geode SC3200 cpu. This is in Release.Tool. BUILDS A2 { INCLUDE "NATIVE NATIVEORIG I386 I386ORIG" COMPILER "Compiler.Compile" COMPILEOPTIONS "--useFPU" TARGET "AMD" EXTENSION "Obx" SYMBOLEXTENSION "Obx" PATH "AOS:" EXCLUDEPACKAGES "OberonVoyager OberonAnts" DISABLED "FALSE" } I haven't changed INCLUDE from the installation default. OK for the Geode? I haven't found any mention of SSE or SSE2 in this processor. Hence --useFPU. Is TARGET "AMD" right for the Geode? Any further tips? Incidentally, Release.Tool has this comment. # INCLUDE # In this package description file it is possible to describe files that are only contained in certain builds. This is done by using indicates, e.g. WIN { list of files } means that ... That should be "... using indicators ..." or "... using build indicators ...". "package description file"? Does INCLUDE refer to packages or files? Really the explanation can be written better. The pronoun "it" isn't necessary. This perhaps. # INCLUDE # A list of packages to include when building. Build indicators can # apply packages to specific builds. For example WIN { list of files } # means that ... Thanks, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

3 3

Re (2): Re (2): A2, targeting a Geode SC3200.
by peter＠easthope.ca 28 Jan '15

28 Jan '15

From: Felix Friedrich <felix.friedrich(a)inf.ethz.ch> Date: Wed, 28 Jan 2015 22:12:22 +0100 > you can try to replace I386.PCI.Mod by BiosFreePCI.Mod. That could help. The BIOS would load and run the boot sector. So I shouldn't puzzle more about that. And in NO, the CF worked only after small revisions in ATADisks according to notes from Josef S. Depending on other demands, I can try BiosFreePCI in the next few days or in early February. Thanks for the suggestion, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

1 0

Re (2): A2, targeting a Geode SC3200.
by peter＠easthope.ca 28 Jan '15

28 Jan '15

Previously, > I'm interested to build an installer for A2 for ... a Geode SC3200 cpu. An installer was built in LinuxAos and A2 installed to IDE0#1 on a CF card. Config strings are left as installed except for "Init=" and "TraceMode=1". The display is visible at http://carnot.yi.org/TatungA2Trace.jpg . Appears familiar and OK down to, No PCI BIOS32 detected {P cpuid= 0, pid= -1 PCI: No bus found.} ATADisks: Detected devices: No devices found. BootConsole: Nount failed on BootVol1 Odd that the kernel starts from a CF card which has an ATA interface yet PCI and ATA are reported to be absent. Also, if Native Oberon is on the CF, it boots and opens tracks and viewers with no complaint. What might be done to let A2 mount the filesystem? Thanks, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

2 1

Re (2): A2, targeting a Geode SC3200.
by peter＠easthope.ca 27 Jan '15

27 Jan '15

From: eas lab <lab.eas(a)gmail.com> Date: Tue, 27 Jan 2015 20:20:13 +0200 > Trying to port ETHO-ish to OLPC? A Tatung TWN-5213 CU tablet. The XO needs the Forth PROM and display controller to operate. The TWN-5213 CU is closer to a generic PC with a touch screen. Regards, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

1 0

A2, targeting a Geode SC3200.
by peter＠easthope.ca 26 Jan '15

26 Jan '15

Hi, Source repository 5919 here. I'm interested to build an installer for A2 for a machine with a Geode SC3200 cpu. This is in Release.Tool. BUILDS A2 { INCLUDE "NATIVE NATIVEORIG I386 I386ORIG" COMPILER "Compiler.Compile" COMPILEOPTIONS "--useFPU" TARGET "AMD" EXTENSION "Obx" SYMBOLEXTENSION "Obx" PATH "AOS:" EXCLUDEPACKAGES "OberonVoyager OberonAnts" DISABLED "FALSE" } I haven't changed INCLUDE from the installation default. OK for this Geode? I haven't found any mention of SSE or SSE2 in this processor. Hence the --useFPU. Is TARGET "AMD" right for the Geode? Any further tips? Incidentally, Release.Tool has this comment. # INCLUDE # In this package description file it is possible to describe files that are only contained in certain builds. This is done by using indicates, e.g. WIN { list of files } means that ... That should be "... using indicators ..." or "... using build indicators ...". "package description file"? Does INCLUDE refer to packages or files? Really the explanation can be written better. The pronoun "it" is rarely necessary. This perhaps. # INCLUDE # A list of packages to include when building. Build indicators can # apply packages to specific builds. For example WIN { list of files } # means that ... Thanks, ... Peter E. -- 123456789 123456789 123456789 123456789 123456789 123456789 123456789 12 Tel +1 360 639 0202 http://carnot.yi.org/ Bcc: peter at easthope. ca

1 0

re. ALO query
by eas lab 24 Jan '15

24 Jan '15

Michael McGaw wrote:- ... >To launch alo in the future: > >1) log in as pi, and execute startx, bringing up the GUI. >2) open LXTerminal >3) type xhost + at the command line >4) cd over to the folder where your alo release resides >5) execute ./alo Good! --You wrote that <oberon.log showed relevant info>. So I wrote "Paste the part of the *.log which shows the transition between 'seems ok' and when the problem starts. Perhaps it shows the problem cause." But using the current amerikan-infinite-frontier/use-specailist=isolated-tools mentality, you mailed an atatchment. Which also lost your previously stated, opinion of <what the log revealed>. With my cursed web-based-gmail, I'm then forced to: 1. go on-line to fetch and read the atatchment; 2. go on-line to request getting the <forms to use Send-mode>; 3. Unlike with ETHO, I can't answer your email and ADD <ethz> to the "To:" list, because gmail is designed per amerikan-WaltDisney-PrePackaged. Before I was forced to use gmail [for socio/politico reasons] ETHO's mailer allowed me to have all of the: * received email/s; * To-be sent email/s; * Relevant text-files. eg. logs, manuals ...; open together on one-single screen, so that you don't lose short-term memory about how the various documents are related, in a COMPLEX way. ==> THIS is what ETHO is about; but nobody ever tries to describe it. Computing is NOT about computers. It's about how to help humans handle complexity. And related: wily [a killer-app] is also available for rPi. wily is the public-domain version of <acme> in the plan9 OS. Plan9 was designed by the boys who designed unix. After they'd done unix, they said "we've learned a lot from that, so now let's make a proper one". That they acknowledge the superiority of ETHO, in copying its main features, says a lot. Unfortunately wily has reversed the left/right mouse-button scrolling direction of TextFrames; which is very uncomfortable for an old ETH-user. I patched the C-source [on my x86:PC's wily] to reverse the scroll-direction, but didn't succeed in fixing the rPi version yet. Typically I have 50% more wily processes than LEO [ETHO for linux] running; which tells how I value wily. OTOH, for the most complex problems, you need ETHO's better fonts and the ability to color-code sections to text on the various TextFrames to show the relationships. Color gives you an extra dimension, to represent relationships. Since you got ALO running, why don't you take it out-on-the-road now for an adventuristic/non-trivial task? Like probing-down into ARM's absolute memory. rPi is amazingly robust [unlike my PC, which lost-a-newish-SATA]. I use rPi to hear TextToSpeech, with only a power-On/Off button [time modulation] to control the play-sequence. Ie. crashing it without unmounting the 'disks'. Together with other ETHO versions, ALO has got an amazing <Decoder> which enables examining down at the machine-code-level. Let's eg. find where/how a frame-buffer is. The FrameBuffer version of LNO, and indeed the good old N-O, enable one to avoid the monsterous X-system. == Chris Glur.

1 0

Re: [Oberon] License Question from Walter Gallegos
by David Egan Evans 19 Jan '15

19 Jan '15

> I think it would be very nice to have a clear statement about the license. > For any commercial use (as apparently envisaged by Markus, Wojtek & Walter) > such a clear statement is absolutely necessary. A lack in license means, essentially, that all rights are reserved. (Depending on country, Berne convention, etc.) I believe Wirth is aware of this, and has knowingly not licensed the Oberon-07 work, other than the obvious fair use that comes from publicly posting his work. However, there is a copy of the ETH Oberon system that available is under a simplified BSD license (among other licenses). -- D. E. Evans http://oberon07.com

1 0

License Question from Walter Gallegos
by greim 19 Jan '15

19 Jan '15

Hello Walter, i have asked this question in Volume 120 issue 5, May 5th, 2014 of this e-mail list. Unfortunately, we have got never an answer from one of the copyright holders. I will city my own question once again, maybe somebody may forward it to the responsible persons: ------------------------------------------------------------------------------- BUT first i think its important to clarify the license question. For me its still not clear what the license for Wirths and Gutknechts Oberon is. The same story with Paul Reeds Verilog code. Wirth said, as far as i understood, its free. But what is free? Free to copy, free to use in a non commercial project, free to change it..... So maybe Wirth, Gutknecht and Reed should give a statement if their code is free under the MIT license or GPL or any other common model. I would prefer MIT but GPL would be ok for me too. ------------------------------------------------------------------------------- Markus Greim > Message: 1 > Date: Sat, 17 Jan 2015 14:17:21 -0200 (UYST) > From: Walter Gallegos <waltergallegos(a)vera.com.uy> > Subject: [Oberon] Is ProjectOberon free for commercial use ? > To: oberon(a)lists.inf.ethz.ch > Message-ID: > <841527004.688560.1421511441793.JavaMail.zimbra(a)vera.com.uy> > Content-Type: text/plain; charset="utf-8" > > Can someone confirm me if ProjectOberon is free for all kind of use ? > > I'm starting a new project and could be a good opportunity to try it as alternative on industrial projects. > >

3 2