Quantcast

[SqNOS] NOS: Small test results #1 + my vague windy ideas

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[SqNOS] NOS: Small test results #1 + my vague windy ideas

Evgeniy Philippov
(11:22:29) egp_: I may hope to write a C/C++ compiler to run from inside
of (modified?) Squeak, to import Linux into Squeak... To experiment like
that... But I am windy, don't know if the desire will stay.
(11:23:44) egp_: Maybe some additional languages like (new) mathtalk
would be necessary
(11:24:02) egp_: (mathtalk is my term)
(11:24:43) egp_: (all of the above experimental thoughts are vague)

About NOS:

I downloaded a 2011.March HDD image and ran it inside a VMWare Player.

(11:47:11) egp_: Great. SqueakNOS has booted OK under "VMWare Player"
under my "Win7 Home Extended". NOS has shown a number of Squeak windows.
(11:48:32) egp_: My machine's hardware are ASUS motherboard, Intel i3
32-bit processor, and nvidia GeForce 8800 GTS.
(11:51:34) egp_: At NOS, mouse and keyboard are working OK. "Load
sources from disk" prompt seems to have loaded sources OK.
(11:53:33) egp_: Monticello browser shows an error "primitive failed".
(11:55:18) egp_: Menu -> System -> "Space left" shows 17 MB.
(11:56:26) egp_: squeak image save seem to have worked OK (I didn't try
reloading image yet.)
(11:58:33) egp_: Test Runner -> "Run Coverage" has shown an error:
MessageNotUnderstood
(12:00:31) egp_: NetworkTests-(Protocols,Url): all (i.e. 39) tests
passed OK.
(12:03:44) egp_: Net-Tests: 63 run, 18 passed OK, 9 failures, 36 errors.
(12:06:09) egp_: NetworkTests-*: 156 run, 123 passed OK, 30 "expected
failures", 3 errors.
(12:10:07) egp_: I switched from VMWare NAT mode to Bridged mode.
Re-testing.
(12:11:55) egp_: Exactly the same results for Net-Tests and Network-Tests-*.
(12:13:44) egp_: Heheh FAT32F*S*Test: 15 run, 6 passed OK, 3 failures, 6
errors. SO BEWARE OF USING ON A REAL HDD, this is a buggy code :)
(12:17:02) egp_: SqueakNOS-(FileSystems, FileSystemsFAT32): 18 run, 7
passed, 3 failures, 8 errors.
(12:19:32) egp_: Well. Personally I find NOS mode as more interesting
than OS+Squeak mode. :)
(12:22:33) egp_: Pressed "Save and quit".
(12:25:50) egp_: Pressed Reset. After a quick boot process (my machine
has SSD drive and 3.3Ghz CPU), all Squeak windows shown up OK.

WBR,
Evgeniy.
_______________________________________________
SqueakNOS mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/squeaknos
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [SqNOS] NOS: Small test results #1 + my vague windy ideas

Guido Chari
Hi,

The linux experiment sounds very interesting. We have talked lot of times about doing something like that for NOS. More than a compiler we have thought something like an emulator. Its very interesting, but is a complex and big project. 

For the tests not green well...last versions are old and, as you can see from last mails, we are trying to find time to adapt NOS to the new VMs (Cog, Stack). That first step would help a lot.

Last versions also hasn't Network support, that's why on test you have that and monticello shows errors. There are some old packages for doing some network stuff, we should look if we can make them work again.

Cheers,
Guido.




2012/11/2 Evgeniy Philippov <[hidden email]>
(11:22:29) egp_: I may hope to write a C/C++ compiler to run from inside of (modified?) Squeak, to import Linux into Squeak... To experiment like that... But I am windy, don't know if the desire will stay.
(11:23:44) egp_: Maybe some additional languages like (new) mathtalk would be necessary
(11:24:02) egp_: (mathtalk is my term)
(11:24:43) egp_: (all of the above experimental thoughts are vague)

About NOS:

I downloaded a 2011.March HDD image and ran it inside a VMWare Player.

(11:47:11) egp_: Great. SqueakNOS has booted OK under "VMWare Player" under my "Win7 Home Extended". NOS has shown a number of Squeak windows.
(11:48:32) egp_: My machine's hardware are ASUS motherboard, Intel i3 32-bit processor, and nvidia GeForce 8800 GTS.
(11:51:34) egp_: At NOS, mouse and keyboard are working OK. "Load sources from disk" prompt seems to have loaded sources OK.
(11:53:33) egp_: Monticello browser shows an error "primitive failed".
(11:55:18) egp_: Menu -> System -> "Space left" shows 17 MB.
(11:56:26) egp_: squeak image save seem to have worked OK (I didn't try reloading image yet.)
(11:58:33) egp_: Test Runner -> "Run Coverage" has shown an error: MessageNotUnderstood
(12:00:31) egp_: NetworkTests-(Protocols,Url): all (i.e. 39) tests passed OK.
(12:03:44) egp_: Net-Tests: 63 run, 18 passed OK, 9 failures, 36 errors.
(12:06:09) egp_: NetworkTests-*: 156 run, 123 passed OK, 30 "expected failures", 3 errors.
(12:10:07) egp_: I switched from VMWare NAT mode to Bridged mode. Re-testing.
(12:11:55) egp_: Exactly the same results for Net-Tests and Network-Tests-*.
(12:13:44) egp_: Heheh FAT32F*S*Test: 15 run, 6 passed OK, 3 failures, 6 errors. SO BEWARE OF USING ON A REAL HDD, this is a buggy code :)
(12:17:02) egp_: SqueakNOS-(FileSystems, FileSystemsFAT32): 18 run, 7 passed, 3 failures, 8 errors.
(12:19:32) egp_: Well. Personally I find NOS mode as more interesting than OS+Squeak mode. :)
(12:22:33) egp_: Pressed "Save and quit".
(12:25:50) egp_: Pressed Reset. After a quick boot process (my machine has SSD drive and 3.3Ghz CPU), all Squeak windows shown up OK.

WBR,
Evgeniy.
_______________________________________________
SqueakNOS mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/squeaknos


_______________________________________________
SqueakNOS mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/squeaknos
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [SqNOS] NOS: Small test results #1 + my vague windy ideas

Evgeniy Philippov
02.11.2012 21:12, Guido Chari пишет:
The linux experiment sounds very interesting. We have talked lot of times about doing something like that for NOS. More than a compiler we have thought something like an emulator. Its very interesting, but is a complex and big project. 
Agreed with all.
`Emulator' is a more precise hit than a vague and misguiding `compiler' term.

WBR,
E.

2012/11/2 Evgeniy Philippov <[hidden email]>
(11:22:29) egp_: I may hope to write a C/C++ compiler to run from inside of (modified?) Squeak, to import Linux into Squeak... To experiment like that... But I am windy, don't know if the desire will stay.
(11:23:44) egp_: Maybe some additional languages like (new) mathtalk would be necessary
(11:24:02) egp_: (mathtalk is my term)
(11:24:43) egp_: (all of the above experimental thoughts are vague)


_______________________________________________
SqueakNOS mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/squeaknos
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [SqNOS] NOS: Small test results #1 + my vague windy ideas

Evgeniy Philippov
In reply to this post by Guido Chari
02.11.2012 21:12, Guido Chari пишет:
> Hi,
>
> The linux experiment sounds very interesting. We have talked lot of
> times about doing something like that for NOS. More than a compiler we
> have thought something like an emulator. Its very interesting, but is
> a complex and big project.

(1) Well... I guess that emulating the linux userspace APIs is one of
the necessary things to acquire, to be able to run Firefox, Adobe Flash,
linux user programs... It is somewhat easier than emulating the entire
linux (with hardware drivers)...

(2) Non-emulating approach could be (A) to rewrite Linux in Smalltalk or
(B) to semi-automatically metaprogrammatically convert Linux into
Smalltalk. I think this approach (2) is more feasible than emulating
approach (1) and will think more on (2).

(2.1) It would help to write a semi-automatic metaprogrammatic converter
environment for (2B). To do (2.1), a gcc + g++ backend generating a
{Smalltalk+asm} code seems to be necessary. So gcc + g++ should be
modified for this. This is highly complex task as gcc is highly complex,
but (I think is) much more feasible than the emulator (1).

***

Meanwhile, I will try to study gcc's source code. This might require
years of study of gcc's sources.

WBR,
Eugene

> For the tests not green well...last versions are old and, as you can
> see from last mails, we are trying to find time to adapt NOS to the
> new VMs (Cog, Stack). That first step would help a lot.
>
> Last versions also hasn't Network support, that's why on test you have
> that and monticello shows errors. There are some old packages for
> doing some network stuff, we should look if we can make them work again.
>
> Cheers,
> Guido.

_______________________________________________
SqueakNOS mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/squeaknos
Loading...