Hi everyone!
This *unofficial* report is just to keep everyone a bit updated. Unofficial means that this report is written by me personally and the other Coordinators (Doug, Marcus, Cees) haven't read it and may disagree :).
Things in this report:
1. Squeak 3.8 Real Soon Now. :) 2. Teams churning. 3. The planning of 3.9 and beyond. 4. New VMs! 64 bits here we come!
Squeak 3.8 ------------- The release has now slipped a whole month, but hey :), Michael Rueger will hopefully soon get it out. AFAIK it is ready to go and only some image last-minute-preparation is left to do. Michael? Stressed out? ;)
Teams ------- We now have a bunch of Teams working on separate tasks (see http://anakin.bluefish.se/castaways/6) , and the team model seems quite promising so far. The teams all have their own mailinglists so a lot of that traffic has left squeak-dev - so don't be fooled and think things aren't moving along.
The following is my personal possibly incorrect :) perception of what is happening in the Teams at the moment (feel free to correct me guys). I am not following all teams, but have skimmed the mls of those I don't follow:
- WebSite (9 subscried to ml). This team has set up a Smallwiki for collaboration on box1 and seems to be starting to itch to get the new website deployed, but I have not yet seen the current version (I think it is temporarily offline). Looking forward to it though!
- MorphicSplitters (15 subscribed to ml). Seems to be lots of activity by looking at the ml acrhive. This team is also quite busy exploring tools for doing analysis, including Dan's code and Daniel Vainsenchers MudPie. Juan Vuletich seems to be the main heavylifter at the moment and wow, I look forward to see what comes out of this team!
- ToolBuilder (32 subscribed to ml). This project started out with a bang, but the ml has been mostly silent during march. I don't think that means anything in particular but I personally look forward to get ToolBuilder installed as a package in Basic in the 3.9 alpha stream ASAP, Brian, you hear me? :) The reasoning for that is to make the API available for all of us so that we can start to program against it under 3.9 which in turn is important because in later releases (4.0 or 4.1 or whatever) Morphic will turn into a package (MorphicSplitters) and then tools that hasn't been ported to using ToolBuilder will become inoperational without Morphic installed.
- Packages (25 subscribed to ml). This is a team I am part of and we have made some initial deicisons (for example to use PackageInfo as it stands today) and also discussed and decided on a way on how SqueakMap and Universes can be united/merged/whatever. For details on that, check the archive. Avi has also created a page on the Swiki where Squeakers can sign up if interested in helping with splitting out or maintaining (or co-maintaining) a part of current Squeak Basic (http://minnow.cc.gatech.edu/squeak/5641). Please people, go there and sign up! The actual "doing it" hasn't begun yet, but I feel it will soon. Avi? Whaddayasay, time to push some PIs into 3.9alpha? :)
- Modules (51 subscribed to ml). This team is led by Dan Ingalls and is focused on delivering a module system "for tomorrow". The team started with a very good inventory of available technology and ideas (http://minnow.cc.gatech.edu/squeak/5608) and is now moving over from analysis to synthesis. :) It will be interesting to see what happens now - IMHO there are a LOT of different aspects here, and to find common ground can get real tricksy. But I have trust in Dan. ;) The inventory in itself is a very good part way result and you can learn a lot just by reading those.
- Janitors (11 subscribed to ml). The janitors are doing us a great service and with Ken at the helms I feel safe that the process of fixes and bugs will gradually move to a better state. One of the things the janitors are doing is that they catch all reports and fixes posted to squeak-dev and move them over to Mantis. So people, remember, it is Mantis we should use from now on. :) The team has also written a good guide on how to report stuff in Mantis.
- v 3.9 (9 subscribed to ml). Doug Way is the release leader for 3.9 and there is a plan for it on the Swiki (http://minnow.cc.gatech.edu/squeak/5645). The work on 3.9 hasn't kicked into gear just yet, but as the teams (Packages, ToolBuilder) get moving and we get the Harvesting started again (Janitors, what are your thoughts on that? Or Doug?) I think it will pick up. And with Doug leading it I am not worried. :) Planned release is in june.
- FileArea and Box1Administration are small teams and nothing much to report. Box1 seems to be humming along fine and Bruce is doing a great job as always keeping the file area up to date. If you haven't noticed this is the current place to find Squeak: http://box1.squeakfoundation.org/files/ ...and if you want to mirror it, read: http://box1.squeakfoundation.org/files/MIRRORING.txt
In my opinion all teams seem to be doing quite good. The archives of the mls are available here: http://discuss.squeakfoundation.org
I don't think we need more teams at the moment - enough work going on to keep us all busy - but if there are new teams brewing out there, then feel free to contact us Coordinators and see if we can form an official team. I know of at least one such team idea that has been discussed.
3.9 and beyond ----------------- The plan for 3.9 is mostly set and 3.9 will turn out as a more or less classic Squeak release with the exception of the result from the Packages team which makes it different from previous releases. But what happens then? What about 4.0 or 4.1? As we described earlier the result of MorphicSplitters was meant to go in *first thing* in the release after 3.9. Since then the discussion about the new image format was brought up - and given that discussion at least I am personally leaning towards this plan:
4.0 = Let's make a release more or less immediately after 3.9 which is identical to 3.9 but with the new image format and related changes. It seems like a logical move. Tim? Reasonable? How about max 1 month after 3.9, which would mean late july?
4.1 = This will then be the release with the result from MorphicSplitters and thus we are buying them a little bit more time. The plan was to push that stuff in as the first thing and if we open the alpha stream of 4.1 after release of 4.0 then that would mean... say 1st august? What do the splitters say about that? Sounds doable?
At the moment I can't find the thread discussing this, so please correct me if you think the above is wrong.
New VMs! ----------- Those of you that track vm-dev (or read the announcement today on squeak-dev :)) are aware of that Ian Piumarta has been working hard lately.
First we got a fresh release of 3.7 with a fix for the annoying 2Gb bug (manifesting as trouble loading ImageSegments (like SqueakMap for example) on selected platforms) and some rearrangements in the source tree that now includes a full generated default src thus removing the need for manually setting up VMMaker if you just want to build a tagged release directly from SVN. Very nice work Ian!
Then today - as you saw in his posting "64-bit VMs" - Ian posted that the 64-bit stuff finally went into the trunk! Yiha. So there are now 64-bit goodness ready for download! Find it all here:
http://squeak.hpl.hp.com/squeak64/dist3/
Hmmm, now I just have to find a 64-bit machine to use it on. And let's just hope this isn't a 1st april joke. ;)
Over and out, Göran
On Fri, 1 Apr 2005 14:57:05 +0100, goran.krampe@bluefish.se wrote:
4.1 = This will then be the release with the result from MorphicSplitters and thus we are buying them a little bit more time. The plan was to push that stuff in as the first thing and if we open the alpha stream of 4.1 after release of 4.0 then that would mean... say 1st august? What do the splitters say about that? Sounds doable?
Doable, yes. But the MorphicSplitters work will result in LARGE changes (the .mcz files are around 2.5Mb together - we're dealing with a huge chunk of code here), so putting it in 4.1 will result in a large changes file needing to be distributed with all of the 4.x series...
Hi Folks,
The first part of the MorphicSplitters project is basically PackgeInfo, i.e. reclassification of stuff. Very little methods were actually modified, and they sould not generate any problems.
But if we wait until 4.1 to get the work in, we'll have to re-check any Morphic or Etoys stuff that is modified during that time. This is especially true for any ToolBuilder stuff.
So I prefer to get this first MorphicSplitters stuff in 3.9, to serve as a base to other efforts, most important ToolBuilder. At some time, MorphicSplitters will need actual refactoring to remove bad dependencies between BaseSqueak and CoreMorphic and between CoreMorphic and FullMorphic/Etoys. The first of these (BaseSqueak -> CoreMorphic) is also needed for ToolBuilder. So, we'll have to coordinate that work. That will be much easier if we start from the same base. This means, please start serious ToolBuilder refactoring with as much as MorphicSplitters in as possible.
Cees, Dan, do you agree?
Cheers, Juan
----- Original Message ----- From: "Cees de Groot" cg@cdegroot.com To: "The general-purpose Squeak developers list" squeak-dev@lists.squeakfoundation.org Sent: Friday, April 01, 2005 11:26 AM Subject: Re: Unofficial status report
On Fri, 1 Apr 2005 14:57:05 +0100, goran.krampe@bluefish.se wrote:
4.1 = This will then be the release with the result from MorphicSplitters and thus we are buying them a little bit more time. The plan was to push that stuff in as the first thing and if we open the alpha stream of 4.1 after release of 4.0 then that would mean... say 1st august? What do the splitters say about that? Sounds doable?
Doable, yes. But the MorphicSplitters work will result in LARGE changes (the .mcz files are around 2.5Mb together - we're dealing with a huge chunk of code here), so putting it in 4.1 will result in a large changes file needing to be distributed with all of the 4.x series...
On Fri, 1 Apr 2005 11:46:14 -0300, Juan Vuletich jmvsqueak@uolsinectis.com.ar wrote:
The first part of the MorphicSplitters project is basically PackgeInfo, i.e. reclassification of stuff. Very little methods were actually modified, and they sould not generate any problems.
Yup. The hard part is only starting :)
Cees, Dan, do you agree?
I agree that the sooner we put stuff into base, the better.
I think that any large amount of work that can rapidly rot should be pushed into 3.9 or current unstable version. Else you will get frustrated. But I'm a bit extreme on that :) (a KCP guy talking).
Hi Folks,
The first part of the MorphicSplitters project is basically PackgeInfo, i.e. reclassification of stuff. Very little methods were actually modified, and they sould not generate any problems.
But if we wait until 4.1 to get the work in, we'll have to re-check any Morphic or Etoys stuff that is modified during that time. This is especially true for any ToolBuilder stuff.
So I prefer to get this first MorphicSplitters stuff in 3.9, to serve as a base to other efforts, most important ToolBuilder. At some time, MorphicSplitters will need actual refactoring to remove bad dependencies between BaseSqueak and CoreMorphic and between CoreMorphic and FullMorphic/Etoys. The first of these (BaseSqueak -> CoreMorphic) is also needed for ToolBuilder. So, we'll have to coordinate that work. That will be much easier if we start from the same base. This means, please start serious ToolBuilder refactoring with as much as MorphicSplitters in as possible.
Cees, Dan, do you agree?
Cheers, Juan
----- Original Message ----- From: "Cees de Groot" cg@cdegroot.com To: "The general-purpose Squeak developers list" squeak-dev@lists.squeakfoundation.org Sent: Friday, April 01, 2005 11:26 AM Subject: Re: Unofficial status report
On Fri, 1 Apr 2005 14:57:05 +0100, goran.krampe@bluefish.se wrote:
4.1 = This will then be the release with the result from MorphicSplitters and thus we are buying them a little bit more time. The plan was to push that stuff in as the first thing and if we open the alpha stream of 4.1 after release of 4.0 then that would mean... say 1st august? What do the splitters say about that? Sounds doable?
Doable, yes. But the MorphicSplitters work will result in LARGE changes (the .mcz files are around 2.5Mb together - we're dealing with a huge chunk of code here), so putting it in 4.1 will result in a large changes file needing to be distributed with all of the 4.x series...
On Apr 1, 2005, at 7:46 AM, Juan Vuletich wrote:
Hi Folks,
The first part of the MorphicSplitters project is basically PackgeInfo, i.e. reclassification of stuff. Very little methods were actually modified, and they sould not generate any problems.
But if we wait until 4.1 to get the work in, we'll have to re-check any Morphic or Etoys stuff that is modified during that time. This is especially true for any ToolBuilder stuff.
Well, so far the TooBuilder stuff has been to remove dependencies on specific GUI classes out and classes that aren't part of a GUI class. :) (Say that three times fast)
So there isn't at this point much overlap is there?
Brian
That's great. But getting this MorphicSplitter stuff in the stream will save other conflicts, and I see no reason for not doing it.
Cheers, Juan ----- Original Message ----- From: "Brian Brown" rbb@techgame.net To: "The general-purpose Squeak developers list" squeak-dev@lists.squeakfoundation.org Sent: Friday, April 01, 2005 1:29 PM Subject: Re: Unofficial status report
Well, so far the TooBuilder stuff has been to remove dependencies on specific GUI classes out and classes that aren't part of a GUI class. :) (Say that three times fast)
So there isn't at this point much overlap is there?
Brian
On [DATE], "[NAME]" <[ADDRESS]> wrote:
Well, so far the TooBuilder stuff has been to remove dependencies on specific GUI classes out and classes that aren't part of a GUI class. :) (Say that three times fast)
So there isn't at this point much overlap is there?
Brian
Yes I was thinking in subscribe to TooBuilder list too, as I do test
Edgar
Goran
I think that we will never have 3.9 for june but for October. Because I see how days are passing fast.
Stef
Hi everyone!
This *unofficial* report is just to keep everyone a bit updated. Unofficial means that this report is written by me personally and the other Coordinators (Doug, Marcus, Cees) haven't read it and may disagree :).
Things in this report:
- Squeak 3.8 Real Soon Now. :)
- Teams churning.
- The planning of 3.9 and beyond.
- New VMs! 64 bits here we come!
Squeak 3.8
The release has now slipped a whole month, but hey :), Michael Rueger will hopefully soon get it out. AFAIK it is ready to go and only some image last-minute-preparation is left to do. Michael? Stressed out? ;)
Teams
We now have a bunch of Teams working on separate tasks (see http://anakin.bluefish.se/castaways/6) , and the team model seems quite promising so far. The teams all have their own mailinglists so a lot of that traffic has left squeak-dev - so don't be fooled and think things aren't moving along.
The following is my personal possibly incorrect :) perception of what is happening in the Teams at the moment (feel free to correct me guys). I am not following all teams, but have skimmed the mls of those I don't follow:
- WebSite (9 subscried to ml). This team has set up a Smallwiki for
collaboration on box1 and seems to be starting to itch to get the new website deployed, but I have not yet seen the current version (I think it is temporarily offline). Looking forward to it though!
- MorphicSplitters (15 subscribed to ml). Seems to be lots of activity
by looking at the ml acrhive. This team is also quite busy exploring tools for doing analysis, including Dan's code and Daniel Vainsenchers MudPie. Juan Vuletich seems to be the main heavylifter at the moment and wow, I look forward to see what comes out of this team!
- ToolBuilder (32 subscribed to ml). This project started out with a
bang, but the ml has been mostly silent during march. I don't think that means anything in particular but I personally look forward to get ToolBuilder installed as a package in Basic in the 3.9 alpha stream ASAP, Brian, you hear me? :) The reasoning for that is to make the API available for all of us so that we can start to program against it under 3.9 which in turn is important because in later releases (4.0 or 4.1 or whatever) Morphic will turn into a package (MorphicSplitters) and then tools that hasn't been ported to using ToolBuilder will become inoperational without Morphic installed.
- Packages (25 subscribed to ml). This is a team I am part of and we
have made some initial deicisons (for example to use PackageInfo as it stands today) and also discussed and decided on a way on how SqueakMap and Universes can be united/merged/whatever. For details on that, check the archive. Avi has also created a page on the Swiki where Squeakers can sign up if interested in helping with splitting out or maintaining (or co-maintaining) a part of current Squeak Basic (http://minnow.cc.gatech.edu/squeak/5641). Please people, go there and sign up! The actual "doing it" hasn't begun yet, but I feel it will soon. Avi? Whaddayasay, time to push some PIs into 3.9alpha? :)
- Modules (51 subscribed to ml). This team is led by Dan Ingalls and
is focused on delivering a module system "for tomorrow". The team started with a very good inventory of available technology and ideas (http://minnow.cc.gatech.edu/squeak/5608) and is now moving over from analysis to synthesis. :) It will be interesting to see what happens now
- IMHO there are a LOT of different aspects here, and to find common
ground can get real tricksy. But I have trust in Dan. ;) The inventory in itself is a very good part way result and you can learn a lot just by reading those.
- Janitors (11 subscribed to ml). The janitors are doing us a great
service and with Ken at the helms I feel safe that the process of fixes and bugs will gradually move to a better state. One of the things the janitors are doing is that they catch all reports and fixes posted to squeak-dev and move them over to Mantis. So people, remember, it is Mantis we should use from now on. :) The team has also written a good guide on how to report stuff in Mantis.
- v 3.9 (9 subscribed to ml). Doug Way is the release leader for 3.9
and there is a plan for it on the Swiki (http://minnow.cc.gatech.edu/squeak/5645). The work on 3.9 hasn't kicked into gear just yet, but as the teams (Packages, ToolBuilder) get moving and we get the Harvesting started again (Janitors, what are your thoughts on that? Or Doug?) I think it will pick up. And with Doug leading it I am not worried. :) Planned release is in june.
- FileArea and Box1Administration are small teams and nothing much to
report. Box1 seems to be humming along fine and Bruce is doing a great job as always keeping the file area up to date. If you haven't noticed this is the current place to find Squeak: http://box1.squeakfoundation.org/files/ ...and if you want to mirror it, read: http://box1.squeakfoundation.org/files/MIRRORING.txt
In my opinion all teams seem to be doing quite good. The archives of the mls are available here: http://discuss.squeakfoundation.org
I don't think we need more teams at the moment - enough work going on to keep us all busy - but if there are new teams brewing out there, then feel free to contact us Coordinators and see if we can form an official team. I know of at least one such team idea that has been discussed.
3.9 and beyond
The plan for 3.9 is mostly set and 3.9 will turn out as a more or less classic Squeak release with the exception of the result from the Packages team which makes it different from previous releases. But what happens then? What about 4.0 or 4.1? As we described earlier the result of MorphicSplitters was meant to go in *first thing* in the release after 3.9. Since then the discussion about the new image format was brought up - and given that discussion at least I am personally leaning towards this plan:
4.0 = Let's make a release more or less immediately after 3.9 which is identical to 3.9 but with the new image format and related changes. It seems like a logical move. Tim? Reasonable? How about max 1 month after 3.9, which would mean late july?
4.1 = This will then be the release with the result from MorphicSplitters and thus we are buying them a little bit more time. The plan was to push that stuff in as the first thing and if we open the alpha stream of 4.1 after release of 4.0 then that would mean... say 1st august? What do the splitters say about that? Sounds doable?
At the moment I can't find the thread discussing this, so please correct me if you think the above is wrong.
New VMs!
Those of you that track vm-dev (or read the announcement today on squeak-dev :)) are aware of that Ian Piumarta has been working hard lately.
First we got a fresh release of 3.7 with a fix for the annoying 2Gb bug (manifesting as trouble loading ImageSegments (like SqueakMap for example) on selected platforms) and some rearrangements in the source tree that now includes a full generated default src thus removing the need for manually setting up VMMaker if you just want to build a tagged release directly from SVN. Very nice work Ian!
Then today - as you saw in his posting "64-bit VMs" - Ian posted that the 64-bit stuff finally went into the trunk! Yiha. So there are now 64-bit goodness ready for download! Find it all here:
http://squeak.hpl.hp.com/squeak64/dist3/
Hmmm, now I just have to find a 64-bit machine to use it on. And let's just hope this isn't a 1st april joke. ;)
Over and out, Göran
goran.krampe@bluefish.se wrote:
- ToolBuilder (32 subscribed to ml). This project started out with a
bang, but the ml has been mostly silent during march. I don't think that means anything in particular but I personally look forward to get ToolBuilder installed as a package in Basic in the 3.9 alpha stream ASAP, Brian, you hear me? :) The reasoning for that is to make the API available for all of us so that we can start to program against it under 3.9 which in turn is important because in later releases (4.0 or 4.1 or whatever) Morphic will turn into a package (MorphicSplitters) and then tools that hasn't been ported to using ToolBuilder will become inoperational without Morphic installed.
The ToolBuilder *api* has been available on SqueakSource for quite a while already. The changes Brian is working on are the places that need to be fixed for ToolBuilder *integration*, e.g., using it in places where we currently have hardcoded references to specific UI frameworks. If you want to look at and play with the ToolBuilder API all you need to load is the packages from Squeaksource (see the Wiki for more information). BTW, these packages are non-intrusive; they contain no modifications but only additions (we keep Brian's changes separate from the core packages).
Regards, - Andreas
Andreas
Would it make sense (after feedabck period) to push the ToolBuilder Api of SqueakSource on 3.9?
Stef
On 2 avr. 05, at 1:51, Andreas Raab wrote:
goran.krampe@bluefish.se wrote:
- ToolBuilder (32 subscribed to ml). This project started out with a
bang, but the ml has been mostly silent during march. I don't think that means anything in particular but I personally look forward to get ToolBuilder installed as a package in Basic in the 3.9 alpha stream ASAP, Brian, you hear me? :) The reasoning for that is to make the API available for all of us so that we can start to program against it under 3.9 which in turn is important because in later releases (4.0 or 4.1 or whatever) Morphic will turn into a package (MorphicSplitters) and then tools that hasn't been ported to using ToolBuilder will become inoperational without Morphic installed.
The ToolBuilder *api* has been available on SqueakSource for quite a while already. The changes Brian is working on are the places that need to be fixed for ToolBuilder *integration*, e.g., using it in places where we currently have hardcoded references to specific UI frameworks. If you want to look at and play with the ToolBuilder API all you need to load is the packages from Squeaksource (see the Wiki for more information). BTW, these packages are non-intrusive; they contain no modifications but only additions (we keep Brian's changes separate from the core packages).
Regards,
- Andreas
Andreas Raab andreas.raab@gmx.de wrote:
goran.krampe@bluefish.se wrote:
- ToolBuilder (32 subscribed to ml). This project started out with a
bang, but the ml has been mostly silent during march. I don't think that means anything in particular but I personally look forward to get ToolBuilder installed as a package in Basic in the 3.9 alpha stream ASAP, Brian, you hear me? :) The reasoning for that is to make the API available for all of us so that we can start to program against it under 3.9 which in turn is important because in later releases (4.0 or 4.1 or whatever) Morphic will turn into a package (MorphicSplitters) and then tools that hasn't been ported to using ToolBuilder will become inoperational without Morphic installed.
The ToolBuilder *api* has been available on SqueakSource for quite a while already. The changes Brian is working on are the places that need
I know that, what I am talking about is *inclusion* in Basic. A lot more people will find it and look at it if it is in Basic, and of course also make the decision to depend on it.
But I agree - I could have mentioned that it has been available all along on SqueakSource, sorry for that. :)
regards, Göran
squeak-dev@lists.squeakfoundation.org
-
Andreas Raab -
Brian Brown -
Cees de Groot -
goran.krampe＠bluefish.se -
Juan Vuletich -
Lic. Edgar J. De Cleene -
stéphane ducasse