December 2009 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] [update 1.1] #11106

by StÃ©phane Ducasse

;) On Dec 17, 2009, at 11:27 AM, Marcus Denker wrote: > 11106 > ----- > > Issue 1410: _ in selectors > > > -- > Marcus Denker -- http://www.marcusdenker.de > Please Note: New Email Adress! marcus.denker(a)inria.fr > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009

Re: [Pharo-project] misconception addendum :)

by StÃ©phane Ducasse

Good idea. I hope to be able to organize some other sprints before. Stef On Dec 17, 2009, at 11:26 AM, CÃ©drick BÃ©ler wrote: > I think we should do a one or two day brainstorming about issue/vision/direction. There are some experiences to share and see what we can do. ESUG10 could be a nice place for that kind of workshop. I personnaly have some ideas/insigths but I cannot really communicate them easily by email (moreover, most of the time they'll be ignored ;-) ). I think it's better to sit around a table a discuss. > > What do you think ? > > 2009/12/17 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> > Hi > > this night I was thinking about a frequent misconception about pharo. > It is easy to think that pharo wants just plain good engineering to open > grey window. Pharo really wants to be innovative and this at all level. > We want better/faster/leaner and more powerful code at all level: UI, network... > > I heard that some people think that we are against MVC but we are not > We are against isMorphic ifTrue: ifFalse:. Pharo does not have to be in > the seaside niche and it will not be. But if this makes the life easier to some > perfect. Now pharoers do not have to either pick web or the rest. > We want to make sure that if someone wants to build an alternative > windowing system he can do it easily. (BTW alain showed to us > some fun and powerful sketches of using cairo (via rome) to have multiple > windowing skeleton.) > Now we certainly do not have the knowledge for everything but this > is not by a lack of vision. So if you want to build this future don't > get us wrong. > > Stef > > PS: I still have in mind the vision of Smalltalk by dan: a system that a simple > person could understand (for me it means beauty, documentation and elegance > of the solutions). > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > > -- > CÃ©drick > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009

Re: [Pharo-project] Fwd: [squeak-dev] Re: [Vm-dev] Re: squeak VM 5.0

by StÃ©phane Ducasse

On Dec 17, 2009, at 10:22 AM, Henrik Johansen wrote: > I agree it seems to belong more in SmalltalkImage current than Smalltalk . > > TBH, I wouldn't mind if we split off a SmalltalkVM class as well :) :) What would you be an extreme abstractor :) BTW henrik it is snowing at your place because here eveyrthing is white but not deep enough to have fun with it. > I'm wondering about the comment in vmParameterAt: though, which states 40 is "VM word size - 4 or 8 (read only)" > > 1) Jannik provided a wordSize: method for setting... Does this work? > 2) In Interpreter >> primitiveVMParameter, the comment reads "40 BytesPerWord for this image" > > I'd recon the VM comment is the correct one, and this is dependant on the image and not the VM, seeing as how it returns 4 for a 32bit image on the 64bit VM. > In which case, I definately feel wordSize semantically belongs to SmalltalkImage (and thus should be cached there rather than SystemDictionary) ;) We will have to have a really neat and clean namespace class instead of systemDictionary. argh too many todos so one by one.... > > Cheers, > Henry > > On Dec 17, 2009, at 8:44 46AM, StÃ©phane Ducasse wrote: > >>> excellent >>> can you sync with marcus? >>> We will have to clean SystemDictionary because this is not a namespace but a kind of clean trash. (when I compared to what is was). >>> I will start to think about what we want. but in a nutshell >>> >>> - a namespace class that only deals with binding (may be class Smalltalk) >>> - a class that deals with image and VM code (may be called SmalltalkImage) >>> >>> Stef >>> >>> On Dec 17, 2009, at 7:44 AM, Laval Jannik wrote: >>> >>>> Hi, >>>> >>>> I modify the variable in the method "clonePreStartup". >>>> So, it works fine. >>>> >>>> ===== >>>> !SystemTracer2 methodsFor: 'clone startup' stamp: 'ajh 8/22/2002 11:15'! >>>> clonePreStartup >>>> "This will be executed right away when the new clone starts up, before processStartup. Subclasses may want to rehash all objects or something" >>>> SystemDictionary wordSize: self wordSize.! ! >>>> ===== >>>> >>>> Now, there are some tests which does not pass. >>>> So, I will publish the code and the pharoImage64, and with the community, we can check this. >>>> >>>> Cheers, >>>> Jannik >>>> >>>> >>>> On Dec 17, 2009, at 02:32 , John M McIntosh wrote: >>>> >>>>> Ok, well I created Mantis >>>>> http://bugs.squeak.org/view.php?id=7430 >>>>> to document. >>>>> >>>>> Personally I would do >>>>> Smalltalk wordSize >>>>> versus >>>>> SystemDictionary wordSize >>>>> >>>>> Also the issue is when to change the WordSize variable. >>>>> >>>>> I think in >>>>> SmalltalkImage>>snapshot: andQuit: embedded: >>>>> right at the >>>>> ifTrue: [self quitPrimitive]. >>>>> Cursor normal show. >>>>> you want to set things so that wordSize is reset. >>>>> >>>>> At quit time you could do the reset before the >>>>> "self quitPrimitive" >>>>> then you know the value is reset and needs to be recalculated at startup time. >>>>> or you can reset it before the >>>>> Cursor normal show. >>>>> which runs after startup time >>>>> >>>>> Somehow I think there is more risk resetting it after startup since I'm not sure >>>>> when something could leap in wanting a valid value. >>>>> >>>>> >>>>> >>>>> On 2009-12-16, at 2:44 PM, Laval Jannik wrote: >>>>> >>>>>> Hi Eliot, hi John, >>>>>> >>>>>> This is what I do: >>>>>> - create a class var in SystemDictionary >>>>>> - accessors (wordSize and wordSize:), the first one can initialize the variable if nil, the second is to modify the value in case of 64bits image. >>>>>> Maybe the second one could be integrated in System-Tracing file. >>>>>> - in CompiledMethod, initialPC use it. >>>>>> >>>>>> Maybe we could integrate it in pharoCore. >>>>>> >>>>> >>>>> -- >>>>> =========================================================================== >>>>> John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 >>>>> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com >>>>> =========================================================================== >>>>> >>>>> >>>>> >>>>> >>>> >>>> >>>> >>> >>> >>> >>> >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009

Re: [Pharo-project] File Browser (was: Re: [newbee] How do I save the contents of a Workspace?)

by StÃ©phane Ducasse

On Dec 17, 2009, at 10:09 AM, Levente Uzonyi wrote: > On Thu, 17 Dec 2009, Mariano Martinez Peck wrote: > >> 2009/12/17 Levente Uzonyi <leves(a)elte.hu> >> >>> On Wed, 16 Dec 2009, csrabak(a)bol.com.br wrote: >>> >>> If you go to more... you get this in a Windows machine. >>>> >>> >>> Note that other items a repeated too, like "decompress to file", >>> "changelist browser", "fileIn entire file", etc >>> >>> >> weird....maybe we should use a Set for that collection ? > > What about finding the cause and fixing the bug? Yes I was thinking the same :) I have only macs around me so I cannot try. May be we get a match in the file list for ZIP, zip, ZiP or something like that. Stef

Dec. 17, 2009

Re: [Pharo-project] File Browser (was: Re: [newbee] How do I save the contents of a Workspace?)

by StÃ©phane Ducasse

normally you do not need to have one. may be somebody will infirm that (this would be stupid that people are forced to have a gmail account). Stef On Dec 17, 2009, at 8:27 AM, Levente Uzonyi wrote: > On Wed, 16 Dec 2009, StÃ©phane Ducasse wrote: > >> Thanks this is really strange. >> I do not see why on windows it would be different but good catch. >> Can you enter a bug entry ? > > I don't have a google account, so someone else will have to. > > > Levente > >> >> Stef >> >> On Dec 16, 2009, at 10:12 PM, Levente Uzonyi wrote: >> >>> Quoting StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: >>> >>>> >>>> Can you show what you get? >>>> >>> >>> Sure, I attached it. The screenshot was taken from an 1.1-11105 image. (Note that the mouse pointer is on the first list item, but you can't see it, because it's a windows machine...) >>> >>> >>> Levente<pharofilebrowser.png>_______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009

Re: [Pharo-project] Fwd: [squeak-dev] Stuff

by StÃ©phane Ducasse

**sorry guys it may be long to read but I wrote it also for you the lurkers :)** Keith some people may decide to ignore and frankly I cannot because I ***UNDERSTAND*** your frustration and I do not know what to do. So I reply because I consider you human and I know that you and me have the same vision for a build system. > I am not better than others, I can't work without others. It is precisely because I need the contribution of others that I am so pissed off with people for splitting up the community we used to have. Particularly the Pharo team for explicitly choosing NOT to participate or use ANY shared community projects that have been ongoing for 3/4 years, when those projects were originally conceived with the same goals as pharo in mind. Half of those projects were Stefanes idea in the first place at a time when he said he didn't have time to do any coding (2006-7) Rio is a prime example, as was interest in SSpec, I need other members of the team to test and feedback on my work, but that teamwork is NOT available, because of the NIH attitude demonstrated by several parties who will not in any circumstance contribute to any externally organized project but insist on working as individuals or as cliques. Keith people have different agenda. I ***understand*** your frustration. Now contrary to most people I did a look at rio (I read it several times in fact) I did a look at Installer. I was pushing it (ask jannik I forced him gently to write a chapter on it). I did a look at sake but could not turn my head around (I should have see the videos before). Now I ***manage*** a team of 9 people during the day. I have to write fucking proposal. My agenda today is 8h30 meeting 9 h30 meeting with RH 12-14 guess what meeting for hiring 14-18 meeting as team head with the others. this is not everyday like that but the other days are - pushing PhD - brainstorming on our research = nothing to do with pharo - writing deliverables - writing projects to get money - fighting to be able to finish coding small stuff ....... Now we created pharo because we could not breath in Squeak. We had enough of insult, bashing, people talking all the time. Pharo is a vision: better infrastructure to change the world! What I suggest is that either you join or you will be ignored and your frustration will grow. > Do you think I have a clue what I am doing when I write something like Sake? No I haven't got a clue, I have never even understood how 'make' works. So I have a go, I learn on the job, sake is on iteration 2.5 as it is, and it would be nice to have some help, I look forward to others joining in and being community to help were I am weak. However what do those others do, they do their own thing because they are too elitist and hotshot to consider contributing to an existing public project that couldn't possibly be worthy of their esteemed input. Some people prefer to do their own. They have this luxury. I do not have it. I'm rewriting packageInfo and it is taking a lot of time Now Pharo is not against what you do. Read my previous answers to you. > Installer is an OPEN community project, I did version 1, Matthew did most of version 2. Along comes Gofer with some new ideas, does gofer contribute to the community supporting vision of a unified extendable DSL for installing things, or does it compete? But lukas has the right to do something that he prefers. You can commit to his repository there is no problem. > Do you think that Bob could have benefitted from Colins experience? I think it could. Why is it acceptable for me to spend $30k of my own time and resources developing something as a project for the community to contribute to and participate in, only to have it flushed down the toliet by someone who is being paid a packet, and can't even be bothered to email out a question like "does Bob do this already?", or "can I help?". Same goes for Dale, he is being paid to do this stuff, so is Stef and Marcus. Oh NO!!! We are NOT PAID to do pharo. Don't dream. Pharo is not my job. My wife is asking me why I do it believe me. Because my research would be probably easier in Ruby or Lua. We do pharo because we want a better substrate for our research. May be we are fool and should do it in Python. If we would be paid we would go MUCH faster to change pharo. May be squeakers are paid but we are not. > I always use other peoples stuff. Everything I do uses other people's stuff, I have always fixed bugs and fed back every problem I have ever found, contributing to their projects as best I can. If I have an idea I look to see if anyone else is doing it and I seek to contribute to that project not to compete with it. I choose to enter dialog, simply because it is easier if someone else has already done the work. Every contribution that I have made has been for the benefit of everyone else, and built on the work of others, and I have sacrificed a lot of paid work to make those contributions. Every contribution I have ever made has been aimed for use by all users of squeak/pharo whatever the image they are currently stuck using, even if it is 3.7. Most of the audience here and over on squeak-dev is only developing for one image and one set of users, so perhaps I do have the moral high ground here, and I feel I may be joined here by the likes of Nick Cellier who has also been contributing equally for everyone. We are not against other but we do not believe that a solution should work to any forks. Now apparently andreas wants competition: I think that this is not good but what can we do? I will not discuss wrong arguments about why pharo could be use as a basis for squeak (pharo goals is to be small and flexible and frankly this would be simpler to have a multimedia layer on top of it. Apparently what we did is so trivial that they can redo it. Perfect to me. We will continue to build our vision. Peeble by peebles... > Now if I can make numerous contributions for 4 years that everyone can use, such as Rio, Logging, Sake, Sake/Packages, Installer, MC1.5, Bob, Sake/Scheduler, Jetsam, Beach etc then with a minor modicum of thought so can everyone else on some projects at least. For example HTTPClient, Network, Collections, NewCompiler, Preferences, DS, changes, closures etc. This partisan politics of pharo camps, squeak camps, is all utter rubbish. It is laziness, simply because no one can be bothered to talk to each other, or to plan what they are doing, or to define an API, and no one in charge has the backbone or political will to say "We will contribute to, and use this project X as a common resource for all". Come on there was nobody beside you and matthew was doing anything in squeak when we started pharo. We got bashed and insulted by people after 3.9 so why would we like to talk with people that do not consider us (or only like shit). For me Squeak was over. > Effort is needed to develop integration and testing servers that would underpin such approaches and this is needed upfront in advance of relentless image hacking. (Bob2 was essentially finished in February 2009) That was the basis of the proposal to the squeak board you know all about, that the board has subsequently abandoned, so now the squeak release team is also now only developing for their own clique. Pharo or Squeak progress isn't planned, you can't tell me what Stefane or Andreas is going to announce next! This isn't really software engineering, it is hacking about reminiscent of what we used to do as schoolboys. We wrote our next milestones check adrian blog (we plan to move it to the project site) > I thought that if you want to plan you make an announcement, like we used to do. "We want to move such and such a feature forward, the repository is here, would you like to join us". What I want to know is why when someone attempts to say "We will open up project X as a common resource for all" the default reaction is not to contribute but to compete and start your own project? Adrian did one for 1.1 Alien or FFI + external package management + beta of new compiler Of course we are not directly working on that but soon we will start > When in 2006 the consensus opinion is that progress was needed I decided to volunteer in some small way. Recognizing that the image needed to move towards a kernel image, one small step, I started with some small projects. I have 32 years programming experience, and I have been using SUnit since 1999ish, my first project ran to 2000 tests in ST/X so I know a little bit about how SUnit could be improved with the ability to categorize tests. I am not some fresh faced coder from graduate school, I felt that if anyone was qualified to make a small contribution to SUnit I was. So we announced some ideas, and made everything public and open for contributions we moved SUnit into a separate repository (squeaksource/Testing) and looked towards combining it with SSpec, we added lots of new features for categorizing tests and a file output based TestReporter which generates results for http display and download using apache. We also added hooks to invoke tests suites from the command line. > > This was offered as the planned community way forward for SUnit in Sept 2006ish FOR EVERYONE as encouraged by Stefane in the first place and at the time it all worked reasonably well in squeak 3.9. > > So then stefane starts Pharo and the pharo team ignore squeaksource/Testing completely and their fresh faced graduate students start making things incompatible, they are actively UNDERMINING all progress that had been made. you are losing credibility... I already said that we should have a look at SUnit. I do not really decide. Now nobody ever said yes this is cool let us use it. We have no feeling about that besides urgency related to time. Now give me one engineer and this will be different! > This isn't just rude, this is patently hostile to any efforts to enable collaboration to happen. It also completely scuppers me because I need the input of others particularly experienced Pharo people contributing to the team in order for it to function. MC1.6 atomic loading needed pharo input to be able to load traits, and that input never happened, and never could happen, because pharo began using a MC that was 2 years behind. This is basically an approach which is destructive to the community, and it is a result of political decision making, whether to support a project as an external resource for everyone to use, or whether to act like control freaks and to retain ownership of a package, (and break it for everyone else) that we really need to be developed for the benefit of all as a matter of importance to be common between all squeak/pharo camps. > So then finally to add insult to injury the processes adopted by both the Squeak and Pharo communities effectively shut out every contribution I have made, not least because all of my contributions are supposed to be for both. For example, my contribution to Squeak is to restructure some packages, in particular Monticello and PackageInfo. Now you try loading a restructured set of packages into the trunk repository. It wont work because the trunk process starts with the MC tools loaded and the package structure already decided, not least of which is the problem that MC cant load a new version of MC into itself. So I spend 3 months of effort working on that problem (LevelPlayingField), for what? Anyhow the net result is a process which effectively to shut out my contribution that I spent a lot of time and resources on. Thanks for that! > > Having a process that could harness everyones contributions was a key philosophical starting point. You have to start any project with your philosophical values, and build on that. You cannot jump ship on your values without causing major upheaval. That is what happened, Andreas explicitly replaced the process that was designed to enable fundamental and important changes to happen in a planned manner with a process that makes such contributions completely impossible. While Elliot is talking about the need to build up an image from a minimal source/kernel, others are talking about reducing the image to a kernel, along comes Andreas and proposes a process based around a monolithic image. It's lunatic stuff it really is. Indeed. To collaborate we should be two and trust each others. We spent too many days with some squeak big names to want to avoid that as much as possible. Once you get burned once you try to think twice. I will not argue. I talk to you because I respect you even when you are mad at me :) I do not talk to others because this is simply useless. > So yes I judge people, when you make decisions that effect other people without consulting them, I think its fair to speak up. Stefane was the one in 2006 listing what features squeak definitely needed, so when I got to work I think it is the least he could do, to build on this work rather than compete with it. Our decision was not competition. I was ok let us take what we know and do something. Probably not clever and more a survival reaction. > I used to believe that the only hope for this squeak/pharo community was for there to be a "sensible and thoughtful" squeak-dev development team following on 6-12 months behind the "forge right on ahead and innovate at all costs" pharo team. This was really only possible if pharo used common tools, which the pharo team point blank refused to do. I thought we could work around this in the short term, and pharo team would eventually see sense, being wowed by the effectiveness of Bob and the new release often process cycle. But now we have a "forge on ahead" without tools squeak-dev team as well, the joint community thing is not ever going to happen. keith I think that one key aspect is that we should meet. ESUG this year will have free entry slot. Request one and come. > While we have a squeak-board process that allows a newly elected person who hasn't been making any recent contributions himself to the community to then scupper 3 years of consistent beneficial work without even discussing his plans with those it effects, and the board then spouts "the end justifies the means", when it doesn't even know what the end is going to be. The squeak world is really broken because it is institutionally unpredictable and I really advise people to stay well clear. I really understand your frustration and we left also because of that kind of attitude. Remember we did (me marcus goran mike) the first foundation. > I will probably use pharo for future work now, just because I will have to use something. I might look into returning to lovely peaceful professional world of ST/X that I am beginning to miss. I am withdrawing my personal input to any of my contributions that does not contribute in some way to the bottom line, I simply cannot afford not to. I am also a full time carer, so I cant just earn nothing indefinitely. If I had time and income I might be persuaded to be nice, document stuff and play the competition game for a bit. I never wanted to compete, and I refuse to compete, all I ever wanted was for people to work together and contribute. Yes I understand that. But it does not work to have a solution working for all the forks. > Work on 3.11 and Bob stopped the instant that Andreas announced his competing approach, because I don't have any will to compete, this isn't a competition, we are supposed to be collaborating and communicating on a professional basis of mutual respect. lol > > Although Bob is finished and operational, I won't make any further contribution until that contribution is towards an actual vision of benefit to the community. While Andreas is competing with pharo, his vision is not helpful to anyone, while pharo's vision is to ignore everyone else, their vision is not helpful either. We do not ignore. We evaluate and sometimes it takes too much time. I was really thorn apart about rio because I like part of it and other I dont; like the law of demeter style. You know it. I know that you remove the Null Pattern. So our goal is not to be arrogant sometimes taking decision is difficult. > The previous squeak 3.11 vision was purposefully orthogonal to the pharo one, and so the two would have been able to fit together with the minimal of effort. yes! > If someone wants to support the vision of continuing to develop tools that we can use to bring the communities together then we stand a chance of moving forward and I would be happy to advise, but I am unable to contribute to this vision any further for financial reasons. I understand. > So to answer your questions... yes I think I can judge because at least I have made an effort to contribute and harmonize all along. Yes and this is why I reply to you. Silence would be horrible in this case. > As far as listening is concerned there is nothing I can do. I no longer have any resources to do anything with what I hear. > > My ranting merely serves to inform those who want to compete rather than to contribute that they have won by default by walking over everything that has gone before. I can't compete, so you have a choice, you either use the extensive work which with has been done on your behalf or you throw it away. We do not want to compete. I do not understand why people can neglect the work with started in May 2008 and want to kill us. Now what can I do? We will really make sure that we will be hard to kill because our vision is really important. > I am no longer in a position to effect things either way, since I have been unable due circumstances to write a single line of code since the squeak board acted as they did. Ranting is all I have left that I can do. Get a job. Did you saw the job offers of JPMorgan. > sorry I'm too. > > Keith > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009

[Pharo-project] [update 1.1] #11106

by Marcus Denker

11106 ----- Issue 1410: _ in selectors -- Marcus Denker -- http://www.marcusdenker.de Please Note: New Email Adress! marcus.denker(a)inria.fr

Dec. 17, 2009

Re: [Pharo-project] misconception addendum :)

by CÃ©drick BÃ©ler

I think we should do a one or two day brainstorming about issue/vision/direction. There are some experiences to share and see what we can do. ESUG10 could be a nice place for that kind of workshop. I personnaly have some ideas/insigths but I cannot really communicate them easily by email (moreover, most of the time they'll be ignored ;-) ). I think it's better to sit around a table a discuss. What do you think ? 2009/12/17 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> > Hi > > this night I was thinking about a frequent misconception about pharo. > It is easy to think that pharo wants just plain good engineering to open > grey window. Pharo really wants to be innovative and this at all level. > We want better/faster/leaner and more powerful code at all level: UI, > network... > > I heard that some people think that we are against MVC but we are not > We are against isMorphic ifTrue: ifFalse:. Pharo does not have to be in > the seaside niche and it will not be. But if this makes the life easier to > some > perfect. Now pharoers do not have to either pick web or the rest. > We want to make sure that if someone wants to build an alternative > windowing system he can do it easily. (BTW alain showed to us > some fun and powerful sketches of using cairo (via rome) to have multiple > windowing skeleton.) > Now we certainly do not have the knowledge for everything but this > is not by a lack of vision. So if you want to build this future don't > get us wrong. > > Stef > > PS: I still have in mind the vision of Smalltalk by dan: a system that a > simple > person could understand (for me it means beauty, documentation and elegance > of the solutions). > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- CÃ©drick

Dec. 17, 2009

[Pharo-project] freetype plugin on os-x squeak v5.0

by John M McIntosh

Ok, I compiled up a universal plugin 32bit PPC, 32/64Bit i386 FreeType plugin. Its in the experimental/COCOA folder on my iDisk http://homepage.mac.com/johnmci/FileSharing.html FT2Plugin.2.3.11.TESTONLYNOTWORKINGFOR64.bundle.zip UNzip and place that in the 64 bit VM for 32bit images Resource folder. YOu can confirm it works by running the VM in 32bit mode, then switch to 64bit mode and explore why it's broken. Now I need someone to tell me where it fails when running as a 64bit VM with a 32bit image. No doubt there is a primitive call that needs to be changed, but I'm a bit too tired now to sort thru how the heck the FreeType class works to find out where it fails. (Silently I might add...) Once we do that then we'll look at the plugin for a 64bit VM and 64bit Image -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

Dec. 17, 2009

Re: [Pharo-project] Fwd: [squeak-dev] Re: [Vm-dev] Re: squeak VM 5.0

by Laval Jannik

Hi, On Dec 17, 2009, at 10:22 , Henrik Johansen wrote: > I agree it seems to belong more in SmalltalkImage current than Smalltalk . > > TBH, I wouldn't mind if we split off a SmalltalkVM class as well :) > > I'm wondering about the comment in vmParameterAt: though, which states 40 is "VM word size - 4 or 8 (read only)" > > 1) Jannik provided a wordSize: method for setting... Does this work? Yes it works :) > 2) In Interpreter >> primitiveVMParameter, the comment reads "40 BytesPerWord for this image" > > I'd recon the VM comment is the correct one, and this is dependant on the image and not the VM, seeing as how it returns 4 for a 32bit image on the 64bit VM. > In which case, I definately feel wordSize semantically belongs to SmalltalkImage (and thus should be cached there rather than SystemDictionary) ;) So, we could do a cache class variable in SmalltalkImage instead of SystemDictonary. This is just a structural problem, not technical. I would do the better for pharo :) Jannik > > Cheers, > Henry > > On Dec 17, 2009, at 8:44 46AM, StÃ©phane Ducasse wrote: > >>> excellent >>> can you sync with marcus? >>> We will have to clean SystemDictionary because this is not a namespace but a kind of clean trash. (when I compared to what is was). >>> I will start to think about what we want. but in a nutshell >>> >>> - a namespace class that only deals with binding (may be class Smalltalk) >>> - a class that deals with image and VM code (may be called SmalltalkImage) >>> >>> Stef >>> >>> On Dec 17, 2009, at 7:44 AM, Laval Jannik wrote: >>> >>>> Hi, >>>> >>>> I modify the variable in the method "clonePreStartup". >>>> So, it works fine. >>>> >>>> ===== >>>> !SystemTracer2 methodsFor: 'clone startup' stamp: 'ajh 8/22/2002 11:15'! >>>> clonePreStartup >>>> "This will be executed right away when the new clone starts up, before processStartup. Subclasses may want to rehash all objects or something" >>>> SystemDictionary wordSize: self wordSize.! ! >>>> ===== >>>> >>>> Now, there are some tests which does not pass. >>>> So, I will publish the code and the pharoImage64, and with the community, we can check this. >>>> >>>> Cheers, >>>> Jannik >>>> >>>> >>>> On Dec 17, 2009, at 02:32 , John M McIntosh wrote: >>>> >>>>> Ok, well I created Mantis >>>>> http://bugs.squeak.org/view.php?id=7430 >>>>> to document. >>>>> >>>>> Personally I would do >>>>> Smalltalk wordSize >>>>> versus >>>>> SystemDictionary wordSize >>>>> >>>>> Also the issue is when to change the WordSize variable. >>>>> >>>>> I think in >>>>> SmalltalkImage>>snapshot: andQuit: embedded: >>>>> right at the >>>>> ifTrue: [self quitPrimitive]. >>>>> Cursor normal show. >>>>> you want to set things so that wordSize is reset. >>>>> >>>>> At quit time you could do the reset before the >>>>> "self quitPrimitive" >>>>> then you know the value is reset and needs to be recalculated at startup time. >>>>> or you can reset it before the >>>>> Cursor normal show. >>>>> which runs after startup time >>>>> >>>>> Somehow I think there is more risk resetting it after startup since I'm not sure >>>>> when something could leap in wanting a valid value. >>>>> >>>>> >>>>> >>>>> On 2009-12-16, at 2:44 PM, Laval Jannik wrote: >>>>> >>>>>> Hi Eliot, hi John, >>>>>> >>>>>> This is what I do: >>>>>> - create a class var in SystemDictionary >>>>>> - accessors (wordSize and wordSize:), the first one can initialize the variable if nil, the second is to modify the value in case of 64bits image. >>>>>> Maybe the second one could be integrated in System-Tracing file. >>>>>> - in CompiledMethod, initialPC use it. >>>>>> >>>>>> Maybe we could integrate it in pharoCore. >>>>>> >>>>> >>>>> -- >>>>> =========================================================================== >>>>> John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 >>>>> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com >>>>> =========================================================================== >>>>> >>>>> >>>>> >>>>> >>>> >>>> >>>> >>> >>> >>> >>> >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dec. 17, 2009