While talking to Craig about spoon, I learned that Tim made a new Compiled Method format using normal objects, and (I think) a VM that ran it. I tried searching for it and came up with:
New Compiled Method Format and 3.0 Image http://wiki.squeak.org/squeak/750 VI4 http://wiki.squeak.org/squeak/2119 Version 4 http://wiki.squeak.org/squeak/3716
I have been discussing what would go into Squeak 4 since we now need to create such a system, due to the urgency of the license situation, and I am now leaning toward Spoon rather than KernelImage as a basis for Squeak4, after learning the magnitude of the relicense effort on Monday.
So, where could I find this code? Does it work? Is there just a CompiledMethod format change, or is there a whole other VM project I never heard of? I can't really tell, but now would be the time to use it if it works.
On Thu, Mar 27, 2008 at 7:06 AM, Matthew Fulmer tapplek@gmail.com wrote:
While talking to Craig about spoon, I learned that Tim made a new Compiled Method format using normal objects, and (I think) a VM that ran it. I tried searching for it and came up with:
New Compiled Method Format and 3.0 Image http://wiki.squeak.org/squeak/750 VI4 http://wiki.squeak.org/squeak/750VI4 http://wiki.squeak.org/squeak/2119 Version 4 http://wiki.squeak.org/squeak/3716
I have been discussing what would go into Squeak 4 since we now need to create such a system, due to the urgency of the license situation, and I am now leaning toward Spoon rather than KernelImage as a basis for Squeak4, after learning the magnitude of the relicense effort on Monday.
Tell us about the magnitude of this relicense effort... is it unachievable? Would we spend less time on the relicensing if we "just did it" using vanilla 3.10, or would it be less time in total to include this refactoring first and then do the relicensing?
Personally, I will eventually (1-2 years maybe?) need a refactored and tidied up CompiledMethod format for my own project, so I have a vested interest in this.
Gulik.
On Thu, Mar 27, 2008 at 09:27:45AM +1300, Michael van der Gulik wrote:
On Thu, Mar 27, 2008 at 7:06 AM, Matthew Fulmer tapplek@gmail.com wrote:
While talking to Craig about spoon, I learned that Tim made a new Compiled Method format using normal objects, and (I think) a VM that ran it. I tried searching for it and came up with: New Compiled Method Format and 3.0 Image http://wiki.squeak.org/squeak/750 VI4 http://wiki.squeak.org/squeak/2119 Version 4 http://wiki.squeak.org/squeak/3716 I have been discussing what would go into Squeak 4 since we now need to create such a system, due to the urgency of the license situation, and I am now leaning toward Spoon rather than KernelImage as a basis for Squeak4, after learning the magnitude of the relicense effort on Monday.Tell us about the magnitude of this relicense effort... is it unachievable? Would we spend less time on the relicensing if we "just did it" using vanilla 3.10, or would it be less time in total to include this refactoring first and then do the relicensing?
Currently, August 2008 is looking like a reasonable release time. For details, see http://installer.pbwiki.org/LicenseAuditing
We have 2 months or so in which no relicensing work can happen (in Plan B), and that time could be spent getting a VM bug-fix into the image.
Personally, I will eventually (1-2 years maybe?) need a refactored and tidied up CompiledMethod format for my own project, so I have a vested interest in this.
Gulik.
-- http://people.squeakfoundation.org/person/mikevdg http://gulik.pbwiki.com/
Can we get a clear statement about what a rewrite process would look like?
Can the new author read the old method? Can someone else produce a requirements document that specifies what the method needs to do? Can someone write tests from the old method that need to pass with the new method?
I still do not understand what we are allowed to do in order to reverse engineer an unclean method.
Thanks, Ron Teitelbaum
-----Original Message----- From: Matthew Fulmer
On Thu, Mar 27, 2008 at 09:27:45AM +1300, Michael van der Gulik wrote:
On Thu, Mar 27, 2008 at 7:06 AM, Matthew Fulmer tapplek@gmail.com
wrote:
While talking to Craig about spoon, I learned that Tim made a new Compiled Method format using normal objects, and (I think) a VM that ran it. I tried searching for it and came up with: New Compiled Method Format and 3.0 Image http://wiki.squeak.org/squeak/750 VI4 http://wiki.squeak.org/squeak/2119 Version 4 http://wiki.squeak.org/squeak/3716 I have been discussing what would go into Squeak 4 since we now need to create such a system, due to the urgency of the license situation, and I am now leaning toward Spoon rather than KernelImage as a basis for Squeak4, after learning the magnitude of the relicense effort on Monday.Tell us about the magnitude of this relicense effort... is it unachievable? Would we spend less time on the relicensing if we "just
did
it" using vanilla 3.10, or would it be less time in total to include
this
refactoring first and then do the relicensing?
Currently, August 2008 is looking like a reasonable release time. For details, see http://installer.pbwiki.org/LicenseAuditing
We have 2 months or so in which no relicensing work can happen (in Plan B), and that time could be spent getting a VM bug-fix into the image.
Personally, I will eventually (1-2 years maybe?) need a refactored
and
tidied up CompiledMethod format for my own project, so I have a
vested
interest in this.
Gulik.
-- http://people.squeakfoundation.org/person/mikevdg http://gulik.pbwiki.com/
-- Matthew Fulmer -- http://mtfulmer.wordpress.com/
On Thu, Mar 27, 2008 at 09:00:48AM -0400, Ron Teitelbaum wrote:
Can we get a clear statement about what a rewrite process would look like?
Can the new author read the old method? Can someone else produce a requirements document that specifies what the method needs to do? Can someone write tests from the old method that need to pass with the new method?
I still do not understand what we are allowed to do in order to reverse engineer an unclean method.
A draft of "the rules" is at http://installer.pbwiki.org/SqueakRelicenseParty
This has not yet been reviewed. Also, a rewrite effort will not happen until July if we indeed go with spoon.
If you could, please wait until I write down the results of meeting four (which was 3 days ago, on Monday). I'm compiling a web site with all this stuff, but I've been pretty busy this week. I'll announce on squeak-dev when it's done.
-----Original Message----- From: Matthew Fulmer A draft of "the rules" is at http://installer.pbwiki.org/SqueakRelicenseParty
Thank you that was exactly what I was looking for. Very nice.
Ron
On Thu, 2008-03-27 at 06:28 -0700, Matthew Fulmer wrote:
On Thu, Mar 27, 2008 at 09:00:48AM -0400, Ron Teitelbaum wrote:
Can we get a clear statement about what a rewrite process would look like?
Can the new author read the old method? Can someone else produce a requirements document that specifies what the method needs to do? Can someone write tests from the old method that need to pass with the new method?
I still do not understand what we are allowed to do in order to reverse engineer an unclean method.
A draft of "the rules" is at http://installer.pbwiki.org/SqueakRelicenseParty
This has not yet been reviewed. Also, a rewrite effort will not happen until July if we indeed go with spoon.
If you could, please wait until I write down the results of meeting four (which was 3 days ago, on Monday). I'm compiling a web site with all this stuff, but I've been pretty busy this week. I'll announce on squeak-dev when it's done.
The mailing list that is mentioned on the page does not exist.
Norbert
vm-dev@lists.squeakfoundation.org
-
Matthew Fulmer -
Michael van der Gulik -
Norbert Hartl -
Ron Teitelbaum