March 2010 - Oberon - mailman3.inf.ethz.ch

XML Components
by Yaroslav Romanchenko 14 Mar '10

14 Mar '10

Hi, Community! I've made some XML components hierarchy. E.g. there's component-container named DrakonDocument. It may has several instances of other child containers named DrakonSchema. Every DrakonSchema container may has only elements (not containers) named DrakonElementHead, DrakonElementAction, ... and so on. Then I've built at runtime simple document and stored it to XML successfully. But when I try to deserialize XML back to component structure in memory it seems fails. At first I've tried to make repository, load it and then load XML. Without success... Actually, then I've tried experiment with standard repository too, but result is strange: Repositories.Load Standard ~ Loading repository 'Standard' ... done. TestComponents.Test Button ~ Component 'Button' not found Is it bug??? Then I've rewrote code of WMComponents.Load such way to use own instance of ElementRegistry: PROCEDURE RegisterAndLoad*(CONST filename: ARRAY OF CHAR; VAR document: DrakonComponents.DrakonDocument); VAR content: XML.Content; registry: XML.ElementRegistry; name: Strings.String; BEGIN (* Try to register elements' generators *) NEW(registry); registry.RegisterElementByName("DrakonDocument", "DrakonComponents", "GenDrakonDocument"); registry.RegisterElementByName("DrakonSchema", "DrakonComponents", "GenDrakonSchema"); registry.RegisterElementByName("DrakonElementHead", "DrakonComponents", "GenDrakonElementHead"); registry.RegisterElementByName("DrakonElementAction", "DrakonComponents", "GenDrakonElementAction"); registry.RegisterElementByName("DrakonElementEnd", "DrakonComponents", "GenDrakonElementEnd"); registry.RegisterElementByName("DrakonElementNode", "DrakonComponents", "GenDrakonElementNode"); document := NIL; content := LoadWithRegistry(filename, registry); IF content # NIL THEN KernelLog.String("xml loaded!"); KernelLog.Ln END; IF content IS XML.Element THEN name := content(XML.Element).GetName(); KernelLog.String(name^); KernelLog.Ln END; IF content IS DrakonComponents.DrakonDocument THEN KernelLog.String("content IS DrakonComponents.DrakonDocument"); KernelLog.Ln; document := content(DrakonComponents.DrakonDocument) END END RegisterAndLoad; It produces output: xml loaded! DrakonDocument So, the root element of XML is XML.Element and element's name is correct, but element isn't of DrakonComponents.DrakonDocument type. Modules attached below. To create XML, execute: DrakonTest.TestCreate schema.XML ~ To load: DrakonTest.TestLoad schema.XML ~ --- Cheers, SAGE http://sage.com.ua/

1 0

Re (2): [Oberon] abbreviation NO in http://en.wikipedia.org/wiki/Native_Oberon
by peasthope＠shaw.ca 05 Mar '10

05 Mar '10

Pieter & everyone, Date: Fri, 05 Mar 2010 22:04:49 +0100, pm> Actually, have we ever considered the abbreviation "ETHNO" for "ETH > Native Oberon"? Good to me. The Wikipedia page is revised. Incidentally, hardware requirements, "133MHz Pentium ... " should be "Intel 80386 or later or equivalent" shouldn't it? I still use the Alpha release on a Toshiba T2100, 486. Regards, Peter E. -- FF, IE, IW, OP, SA: Google "pathology workshop" NO: Desktops.OpenDoc "http://www.google.ca/search?q=pathology+workshop" [click here ^]

1 0

abbreviation NO in http://en.wikipedia.org/wiki/Native_Oberon
by peasthope＠shaw.ca 05 Mar '10

05 Mar '10

I've added the abbreviation NO in the first sentence in http://en.wikipedia.org/wiki/Native_Oberon. Any objection or suggestion or recommendation? Thanks, ... Peter E. -- FF, IE, IW, OP, SA: Google "pathology workshop" NO: Desktops.OpenDoc "http://www.google.ca/search?q=pathology+workshop" [click here ^]

2 1

Assembler: change in release
by Søren Renner 03 Mar '10

03 Mar '10

Assembler code that was accepted by the previous compiler no longer is: MOV EDI, 8[EBP] must be rewritten as MOV EDI, [EBP+8] and so forth.

1 0

compiler trap update
by Søren Renner 03 Mar '10

03 Mar '10

Curiously it seems that the compiler trap results from the attempt to compile a module which has an old OBX object file present on the partition, the deletion of which repairs the bug.

1 0

Release Bug: Compiler Trap
by Søren Renner 03 Mar '10

03 Mar '10

The compiler runs from the boot CD, but after installation to a partition there is a trap in PCT. On the advice of Thomas Frey I tried to recompile PCT.Mod from another boot partition (where the old compiler works) but the resulting code is not compatible with the other new code and will not run.

1 0

New A2 native release
by Ulrike Glavitsch 03 Mar '10

03 Mar '10

Dear A2 users, a new native A2 release (rev. 2731 of Dec. 11, 2009) is ready for download at www.bluebottle.ethz.ch. It contains the new interruptible garbage collector that now functions properly for multi-core machines. I want to thank all those who helped finding errors during the process of making this release stable. I wish everybody good luck with the new release. Best regards Ulla Glavitsch -- ETH Zurich Ulrike Glavitsch Eggler Institute for Computer Systems Clausiusstrasse 59 8092 Zurich ulrike.glavitsch(a)inf.ethz.ch +41 44 632 70 49 phone +41 44 632 13 07 fax

3 4