From: Filip Zawadiak (philz_at_vyx.net)
Date: Tue Jul 08 2003 - 01:33:37 PDT
I started restoring my site last week, started with looking through old
backup CDs for stored website. Unfortunately, haven't found it. So I had
to look my "work" directory backups package by package, and suddenly
recalled times when I actively hacked Newtons :-)
After 5 years I'm back :-)
First of all, I would like all of you for very positive response for
Paul's mail about fund raising. Having my own Newton with necessary tool
would help me finding out "dynamic" behaviour of stores - for now I'm
using old store dump of my (sold long time ago) MP130. I'm also creating
package stores with NTK - that's good because:
1. I'm sure I correctly extract all objects
2. I know what should be inside :-)
Except soups architecture I would also like to find out how TFlashBlocks
work - and that's impossible without "live" Newton. I never used MP2100
debugging images, but if they have more symbols than what I had for
MP130 it could greatly help.
Another thing would be updating (or rewriting) my bytecode
dissassembler. Valid bytecode decoding is crucial for VM implementation,
which also is something I would like to do.
I write all new tools in Ruby, as it's much faster and helps concentrate
on problem not implementation. Hopefully after veryfying their "proper"
working I'll convert everything back to C/C++.
I also have my old NewtObjects library, written in C++, which exposes
C++ API for Newton Object System resembling that of real Newtons. It's
not finished, and some thing need HUGE corrections (GC!!!) but it allows to:
1. Emit NSOF streams
2. Dump packages (like this Checkbook included on my webpage)
Hopefully, someday I'll integrate package decoders, store decoders and
object library. That would give solid foundations for building
Newton-like software environment.
What I can do now with soups:
1. Extract objects from package stores (100% correct)
2. Extract objects from Internal flash store (Correctness unverified...)
3. Probably extract objects from linear flash cards (unchecked)
4. Read Store information from root object
5. Read any entry given it's objects ID (new!)
5a. Decode most frequently used types from entries, adding further is
no-brainer, just takes few minutes (new!)
6. Very crude dumping of any index structure
7. Dumping of 'int and 'string index data
8. 6&7 does not include traversing of inter-index links
9. Reading VBOs, data is read for TZippy, TSimple and TLZStoreCompander.
For TSimple you can actually see what's inside :-)
10. Dumping ALL vbos from extracted objects
My website is planned to stay on vyx.net "forever" - that's my own
domain. Currently it's hosted on philz.freezope.org, and there are
redirections from vyx.net. I wanted to pay for having "vyx.net" hosted
on freezope, but they only accept payments by PayPal, and PayPal rejects
all users from Poland :-/
As for being broke - I'ts not that bad, or it is, depending on point of
view. I always dreamed on flying planes, and finally could afford it :-)
It was so nice, that I spend 200% of allocated budget for it, and
currently have to stop all "non-essential" expenses, and pay back credit
line in my bank :-)
OK, now back to decoding stores :-)
-- Filip Zawadiak mailto:philz_at_vyx.net http://vyx.net/~philz ICQ#25454699 -- This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries List FAQ/Etiquette/Terms: http://www.newtontalk.net/faq.html Official Newton FAQ: http://www.chuma.org/newton/faq/
This archive was generated by hypermail 2.1.5 : Thu Jul 10 2003 - 04:30:01 PDT