--------- Sorry by my google english-------------- The version 3.10 does not work on Linux.
Folks. This afternoon download the 2 (3.10 and 3.10-1 Versions) official versions from the official FTP but these images doesn't run on Linux (my desktop is a kubuntu.). By someone reason, world doesn't is displayed . As I did not leave any log or anything similar I cant give you more details. I Understand, that most of you have mac and this is the test platform by excellence, but take into consideration that there are an enormous number of people who do not have the resources and the only alternative we have left (without falling into piracy, off course) , is linux. Furthermore, since the personal field, I recommend give more support to open sources platforms, of which we are a part. Giving more support as Mac and Windows platforms, is a sort of betrayal. I really cant judge (in addition to recently that I am in this community) what are the motivations that each has to carry squeak ahead, but I can say is that Squeak must be faithful to their roots and ideals and become an option allowing the development of quality in the purest form possible POO. Since no environments objects "Clean" are not part of the technology of "Main Stream" I think that everyone is responsible is reached REALES consistent with what is
preaches in Squeak and not just to be supportive "dummy" to open source platforms and development is made on platforms and shut out of reach for most of us who
we live south of Rio Grande. As a last resort I had to download and use the latest images of Demian Causses (which actually runs on all platforms supported by squeak). I do not take bad, but I think that instead of discussing fribolidades as renaming a squeak or if it is only a platform for research or not, we should seek to improve this
towards a modular platform and have a solid. For this goal we need the board to take reins in the matter and will lead to the release team with a clear objective, and not be carried by winds from a list of mail or those who
IRC # pass. I believe that the inclusion of Dan Ingalls on the board would take such an effect, but it seems that we will have to wait for another leader emerges who drives the fate of squeak.
I hope you do not take my comments wrong, try to be as frank as possible. I ask forgiveness if something political mistake committed, but this is what I think. I write the mail in two languages I know to be more clear with everyone.
Honestly.
La version 3.10 no funciona en linux.
Amigos: Esta tarde baje las 2 versiones oficiales, que estan el ftp de squeak y ninguna de estas funciona en linux (mi desktop es un kubuntu.). Por algún motivo particular nunca se abre el mundo. Como no me deja ningun log no puedo darles mas detalles. Yo en tiendo que la mayoría de ustedes tenga mac, y esa sea la plataforma de prueba por excelencia, pero tengan en cuenta que hay una enorme cantidad de gente que no disponemos de los recursos y la única alternativa que nos queda (sin caer en la piratería porsupuesto), es linux. Por otro lado, y ya desde el ambito personal, les recomiendo que apoyen mas a las plataformas open sources, de la cuales somos parte. Dar mas soporte a plataformas como mac y windows, es algún tipo de traición. Realmente no puedo juzgar. (ademas que hace poco tiempo que estoy en esta comunidad) cuales son las motivaciones que cada uno tiene para formar parte de Squeak, pero lo que si
puedo decir, es que Squeak debe ser fiel a sus raices e ideales, y convertirse en una opcion de calidad que permita desarrollar en la forma mas pura posible la POO. Dado que los entornos de objetos "Limpios" no son parte de las tecnología del "Main Stream" creo que es reponsabilidad de todos, que los se llegue a realeses coherentes con lo que se
predica en Squeak y no solo que se de soporte "dummy" a las plataformas open source, y que el desarrollo se haga sobre plataformas cerradas y fuera del alcance de la mayoria de nosotros que vivimos al sur de rio grande. Como ultimo recurso tuve que bajar y usar las ultimas imagenes de Demian Causso (que realmente funcionan en todos las plataforma soportadas por squeak). No me tomen a mal, pero creo que en lugar de discutir fribolidades como cambiar el nombre a squeak o si es solo una plataforma de investigacion o no, deberíamos buscar mejorar esta
en pos de una tener una plataforma modular y solida. Para este objetivo necesitamos que el board tome acciones al respeto y conduzca al release team con un objetivo claro, y no sea llevado por los vientos de una lista de mail o los que
pase en #IRC. Yo crei que la inclusion de Dan Ingalls en el board iba a tener en ese efecto, pero parece que vamos a tener que esperar que surja otro lider que conduzca el destino de squeak.
Espero que no tomen a mal mis comentarios, trate de ser lo mas franco posible. Les pido perdon si cometi algo error politico, pero esto es lo que pienso.
Sinceramente.
You're facing a problem which I find very annoying, too. The 3.10 image has its default display depth set to 16bpp. Under linux with composite extension (if I understand it right) the colors on the display are calculated in the wrong way so every color is tranlucent. You may fix it by having a dark background. This way you can see the world and you'll be able to select World->appearance->set display depth. If you choose 32bit everything will turn back to normal.
So there is a problem with calculation of the color. But the question that I find much more important is: Why has to default color depth changed from 3.9 to 3.10??
Norbert
On Fri, 2008-05-30 at 00:20 -0300, Diogenes Moreira wrote:
--------- Sorry by my google english--------------
The version 3.10 does not work on Linux.
Folks. This afternoon download the 2 (3.10 and 3.10-1 Versions) official versions from the official FTP but these images doesn't run on Linux (my desktop is a kubuntu.). By some one reason, world doesn't is displayed . As I did not leave any log or anything similar I cant give you more details. I Understand, that most of you have mac and this is the test platform by excellence, but take into consideration that there are an enormous number of people who do not have the resources and the only alternative we have left (without falling into piracy, off course) , is linux. Furthermore, since the personal field, I recommend give more support to open sources platforms, of which we are a part. Giving more support as Mac and Windows platforms, is a sort of betrayal. I really cant judge (in addition to recently that I am in this community) what are the motivations that each has to carry squeak ahead, but I can say is that Squeak must be faithful to their roots and ideals and become an option allowing the development of quality in the purest form possible POO. Since no environments objects "Clean" are not part of the technology of "Main Stream" I think that everyone is responsible is reached REALES consistent with what is
preaches in Squeak and not just to be supportive "dummy" to open source platforms and development is made on platforms and shut out of reach for most of us who
we live south of Rio Grande. As a last resort I had to download and use the latest images of Demian Causses (which actually runs on all platforms supported by squeak). I do not take bad, but I think that instead of discussing fribolidades as renaming a squeak or if it is only a platform for research or not, we should seek to improve this
towards a modular platform and have a solid. For this goal we need the board to take reins in the matter and will lead to the release team with a clear objective, and not be carried by winds from a list of mail or those who
IRC # pass. I believe that the inclusion of Dan Ingalls on the board would take such an effect, but it seems that we will have to wait for another leader emerges who drives the fate of squeak.
I hope you do not take my comments wrong, try to be as frank as possible. I ask forgiveness if something political mistake committed, but this is what I think. I write the mail in two languages I know to be more clear with everyone.
Honestly.
La version 3.10 no funciona en linux.
Amigos: Esta tarde baje las 2 versiones oficiales, que estan el ftp de squeak y ninguna de estas funciona en linux (mi desktop es un kubuntu.). Por algún motivo particular nunca se abre el mundo. Como no me deja ningun log no puedo darles mas detalles. Yo en tiendo que la mayoría de ustedes tenga mac, y esa sea la plataforma de prueba por excelencia, pero tengan en cuenta que hay una enorme cantidad de gente que no disponemos de los recursos y la única alternativa que nos queda (sin caer en la piratería porsupuesto), es linux. Por otro lado, y ya desde el ambito personal, les recomiendo que apoyen mas a las plataformas open sources, de la cuales somos parte. Dar mas soporte a plataformas como mac y windows, es algún tipo de traición. Realmente no puedo juzgar. (ademas que hace poco tiempo que estoy en esta comunidad) cuales son las motivaciones que cada uno tiene para formar parte de Squeak, pero lo que si
puedo decir, es que Squeak debe ser fiel a sus raices e ideales, y convertirse en una opcion de calidad que permita desarrollar en la forma mas pura posible la POO. Dado que los entornos de objetos "Limpios" no son parte de las tecnología del "Main Stream" creo que es reponsabilidad de todos, que los se llegue a realeses coherentes con lo que se
predica en Squeak y no solo que se de soporte "dummy" a las plataformas open source, y que el desarrollo se haga sobre plataformas cerradas y fuera del alcance de la mayoria de nosotros que vivimos al sur de rio grande. Como ultimo recurso tuve que bajar y usar las ultimas imagenes de Demian Causso (que realmente funcionan en todos las plataforma soportadas por squeak). No me tomen a mal, pero creo que en lugar de discutir fribolidades como cambiar el nombre a squeak o si es solo una plataforma de investigacion o no, deberíamos buscar mejorar esta
en pos de una tener una plataforma modular y solida. Para este objetivo necesitamos que el board tome acciones al respeto y conduzca al release team con un objetivo claro, y no sea llevado por los vientos de una lista de mail o los que
pase en #IRC. Yo crei que la inclusion de Dan Ingalls en el board iba a tener en ese efecto, pero parece que vamos a tener que esperar que surja otro lider que conduzca el destino de squeak.
Espero que no tomen a mal mis comentarios, trate de ser lo mas franco posible. Les pido perdon si cometi algo error politico, pero esto es lo que pienso.
Sinceramente.
Norbert,
Thank for you answer, but this case this not is my problem. Before to write the previous mail, I was researching a lot of time and find the color problem in internet. after that, I tried solve the problem, such as you mention in you answer, but nothing, I can't interact with the world
But, the previous mail is a warning, please, "More atention to open source platform",
if you compare the squeak implementation in windows, mac and linux you will see alot of diferents, such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action. In that simple details or the bigger details such as the color depth, you can see the disinterest of squeak's team on open source platform (at least in the releases's testing).
please don't forget of the south people. thank you for you answer again.
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
Norbert,
Thank for you answer, but this case this not is my problem. Before to write the previous mail, I was researching a lot of time and find the color problem in internet. after that, I tried solve the problem, such as you mention in you answer, but nothing, I can't interact with the world
But, the previous mail is a warning, please, "More atention to open source platform",
if you compare the squeak implementation in windows, mac and linux you will see alot of diferents, such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action. In that simple details or the bigger details such as the color depth, you can see the disinterest of squeak's team on open source platform (at least in the releases's testing).
please don't forget of the south people. thank you for you answer again.
The release team pays attention on releasing images, not for releasing VMs. For each platform we have different people who supporting the platform code.
And besides, Squeak is open source too. Instead of blaming others who 'not paying attention to open source platform', you can take part in improving things and provide code updates/fixes.
Oki, you are rigth...I can fix and update the problems, but
Perhaps it would be more productive, work together and not the separated efforts of one person?. It would be like fighting windmills, I want to solve everything, even if I could. Therefore appointed the board, because I think of where you should leave this leadership.
Anyway forgiveness for my confusion over the release team and the VM team.
By the way, I'm using an image that I was having problems with the file changes. We note that the stamp of the methods are not stored in UTF-8, in the common cases should have no problem, but if someone puts extended in the sign ASCII characters are stored wrong. That what I discovered working with the port of Glorp, if they need a reference.
PS: I will debugging the C + + code to see if I can find the current problem. If I find something, i'll send a fixs or advices, Where can send a fixes or updates?
Best regards,
On Fri, May 30, 2008 at 10:26 AM, Igor Stasenko siguctua@gmail.com wrote:
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
Norbert,
Thank for you answer, but this case this not is my problem. Before to write the previous mail, I was researching a lot of time and find the color problem in internet. after that, I tried solve the problem, such as you mention in you answer, but nothing, I can't interact with the world
But, the previous mail is a warning, please, "More atention to open source platform",
if you compare the squeak implementation in windows, mac and linux you will see alot of diferents, such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action. In that simple details or the bigger details such as the color depth, you can see the disinterest of squeak's team on open source platform (at least in the releases's testing).
please don't forget of the south people. thank you for you answer again.
The release team pays attention on releasing images, not for releasing VMs. For each platform we have different people who supporting the platform code.
And besides, Squeak is open source too. Instead of blaming others who 'not paying attention to open source platform', you can take part in improving things and provide code updates/fixes.
-- Best regards, Igor Stasenko AKA sig.
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
Oki, you are rigth...I can fix and update the problems, but
Perhaps it would be more productive, work together and not the separated efforts of one person?. It would be like fighting windmills, I want to solve everything, even if I could. Therefore appointed the board, because I think of where you should leave this leadership.
You are right. VM evolving much slower than images. There is not many people who can get there and make any changes. And most , who can do, are very busy with other things :)
Also, keep in mind that platform VM code is much more conservative than anything else in squeak, since it should be rock solid and keep supporting all features which came from old ages.
Anyway forgiveness for my confusion over the release team and the VM team.
By the way, I'm using an image that I was having problems with the file changes. We note that the stamp of the methods are not stored in UTF-8, in the common cases should have no problem, but if someone puts extended in the sign ASCII characters are stored wrong. That what I discovered working with the port of Glorp, if they need a reference.
PS: I will debugging the C + + code to see if I can find the current problem. If I find something, i'll send a fixs or advices, Where can send a fixes or updates?
As usual, at bugs.squeak.org. Also, post in this list, or vm-dev@lists.squeakfoundation.org list.
Best regards,
Diogenes, please try to avoid identifying Ubuntu with Linux, as Ubuntu is mostly a linux distro made as fast as possible, even if it's not ready.
I've just tested ftp://ftp.squeak.org/3.10/Squeak3.10.1-7175-basic.zip on Debian testing, in an amd64 machine and it works fine. So, I think you should contact Ubuntu maintainers as the problem is on Ubuntu side, not on linux or squeak sides.
Cheers. José L.
2008/5/30 Igor Stasenko siguctua@gmail.com:
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
Oki, you are rigth...I can fix and update the problems, but
Perhaps it would be more productive, work together and not the separated efforts of one person?. It would be like fighting windmills, I want to solve everything, even if I could. Therefore appointed the board, because I think of where you should leave this leadership.
You are right. VM evolving much slower than images. There is not many people who can get there and make any changes. And most , who can do, are very busy with other things :)
Also, keep in mind that platform VM code is much more conservative than anything else in squeak, since it should be rock solid and keep supporting all features which came from old ages.
Anyway forgiveness for my confusion over the release team and the VM
team.
By the way, I'm using an image that I was having problems with the file changes. We note that the stamp of the methods are not stored in UTF-8, in the common cases should have no problem, but if someone puts extended in the sign ASCII characters are stored wrong. That what I discovered working with the port of Glorp, if they need a reference.
PS: I will debugging the C + + code to see if I can find the current problem. If I find something, i'll send a fixs or advices, Where can send a fixes or updates?
As usual, at bugs.squeak.org. Also, post in this list, or vm-dev@lists.squeakfoundation.org list.
Best regards,
-- Best regards, Igor Stasenko AKA sig.
By the way, I'm using an image that I was having problems with the file changes. We note that the stamp of the methods are not stored in UTF-8, in the common cases should have no problem, but if someone puts extended in the sign ASCII characters are stored wrong. That what I discovered working with the port of Glorp, if they need a reference.
I think that the stamp is stored in UTF-8 in 3.10 as well (at least it does in the OLPC Etoys image). But #getPreambleFrom:at: is looking at the sequence backward so that it cannot decode it properly.
-- Yoshiki
Igor Stasenko ha scritto:
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
The release team pays attention on releasing images, not for releasing VMs. For each platform we have different people who supporting the platform code.
Am I the only one who thinks that this is wrong?
Giovanni
On Fri, May 30, 2008 at 6:16 PM, Giovanni Corriga giovanni@corriga.net wrote:
Igor Stasenko ha scritto:
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
The release team pays attention on releasing images, not for releasing VMs. For each platform we have different people who supporting the platform code.
Am I the only one who thinks that this is wrong?
It seems appropriate that people with the specific skills of the platform should maintain the VM code. Doesn't it?
brad
Brad Fuller ha scritto:
On Fri, May 30, 2008 at 6:16 PM, Giovanni Corriga giovanni@corriga.net wrote:
Igor Stasenko ha scritto:
2008/5/30 Diogenes Moreira diogenes.moreira@gmail.com:
The release team pays attention on releasing images, not for releasing VMs. For each platform we have different people who supporting the platform code.
Am I the only one who thinks that this is wrong?
It seems appropriate that people with the specific skills of the platform should maintain the VM code. Doesn't it?
I'm not talking about mantaining, I'm talking about releasing. Squeak is the only system that I know of where the VM/interpreter is not released at the same time as "the runtime". We should try and have new VMs released when we release the new image, if only to make things more understandable for the beginners.
Giovanni
Giovanni
Giovanni Corriga wrote:
I'm not talking about mantaining, I'm talking about releasing. Squeak is the only system that I know of where the VM/interpreter is not released at the same time as "the runtime". We should try and have new VMs released when we release the new image, if only to make things more understandable for the beginners.
Can you explain what exactly you mean by "more understandable for beginners"? The current situation is such that we have updated 3.10 VMs available at http://ftp.squeak.org/3.10/ which (outside of John's naming scheme) reflect the 3.10 heritage accurately. These VMs were available in alpha and beta versions earlier and are now available alongside the appropriate image release. What's not to understand about that?
Cheers, - Andreas
Andreas Raab ha scritto:
Giovanni Corriga wrote:
I'm not talking about mantaining, I'm talking about releasing. Squeak is the only system that I know of where the VM/interpreter is not released at the same time as "the runtime". We should try and have new VMs released when we release the new image, if only to make things more understandable for the beginners.
Can you explain what exactly you mean by "more understandable for beginners"? The current situation is such that we have updated 3.10 VMs available at http://ftp.squeak.org/3.10/ which (outside of John's naming scheme) reflect the 3.10 heritage accurately. These VMs were available in alpha and beta versions earlier and are now available alongside the appropriate image release. What's not to understand about that?
At the moment, the current version of the Windows vm is 3.10.6 . If I where a newbie, I would expect to have a 3.10.6 image too, to use with that vm.
Giovanni
On Sat, May 31, 2008 at 6:48 PM, Giovanni Corriga giovanni@corriga.net wrote:
Andreas Raab ha scritto:
Giovanni Corriga wrote:
I'm not talking about mantaining, I'm talking about releasing. Squeak is the only system that I know of where the VM/interpreter is not released at the same time as "the runtime". We should try and have new VMs released when we release the new image, if only to make things more understandable for the beginners.
Can you explain what exactly you mean by "more understandable for beginners"? The current situation is such that we have updated 3.10 VMs available at http://ftp.squeak.org/3.10/ which (outside of John's naming scheme) reflect the 3.10 heritage accurately. These VMs were available in alpha and beta versions earlier and are now available alongside the appropriate image release. What's not to understand about that?
At the moment, the current version of the Windows vm is 3.10.6 . If I where a newbie, I would expect to have a 3.10.6 image too, to use with that vm.
maybe we just need to make it clearer that version numbers are different for vm's and images. Or, convert the version numbers to a word (like "leopard") and use that for the complete package of image and vm that's released (at a major releases)
Giovanni Corriga wrote:
At the moment, the current version of the Windows vm is 3.10.6 . If I where a newbie, I would expect to have a 3.10.6 image too, to use with that vm.
I see. It seems difficult to change that though, because the VMs do evolve differently. Any ideas how to improve the situation?
Cheers, - Andreas
Well actually we had a huge difference of opinion in versioning within the Sophie team and I learned versioning is fairly irrational and a political football at times.
As background, in 1996 era the entire VM source code *was* distributed within the image. Because of this for years the VM version was tightly coupled to the image version, since each VM change for the macintosh required pushing out a new changeset and linking that somehow to the Image version number. So I'll assume guilt for keeping that cycle going for years.
Lately (read years) the VMs have moved away from daily/weekly/monthly changes, well perhaps not the macintosh VM, but that is because Apple keeps pushing out different versions of OSX and changing the rules, even beyond fostering migration from powerpc to macintel. Vista users are you happy? (couldn't resist, thought not... ).
Really I suspect the VM version numbers should fully decouple from the image version numbers, really nothing requires one to use a 3.10 image to build a VM that will work with a 3.10, I'm quite happy to continue working with my 3.8 VMMaker image...
At the moment I'm fully occupied with a VM rewrite, post Sophie era, so I'll take hints on what the new macintosh version should be called...
On May 31, 2008, at 10:46 PM, Andreas Raab wrote:
Giovanni Corriga wrote:
At the moment, the current version of the Windows vm is 3.10.6 . If I where a newbie, I would expect to have a 3.10.6 image too, to use with that vm.
I see. It seems difficult to change that though, because the VMs do evolve differently. Any ideas how to improve the situation?
Cheers,
- Andreas
-- = = = ======================================================================== John M. McIntosh johnmci@smalltalkconsulting.com Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================
John M McIntosh wrote:
Really I suspect the VM version numbers should fully decouple from the image version numbers, really nothing requires one to use a 3.10 image to build a VM that will work with a 3.10, I'm quite happy to continue working with my 3.8 VMMaker image...
And of course the story comes full circle once people realize that the reason for naming the VMs according to image versions was precisely in response to complaints about how confusing it is that the versions differ ;-)
Cheers, - Andreas
At the moment I'm fully occupied with a VM rewrite, post Sophie era, so I'll take hints on what the new macintosh version should be called...
On May 31, 2008, at 10:46 PM, Andreas Raab wrote:
Giovanni Corriga wrote:
At the moment, the current version of the Windows vm is 3.10.6 . If I where a newbie, I would expect to have a 3.10.6 image too, to use with that vm.
I see. It seems difficult to change that though, because the VMs do evolve differently. Any ideas how to improve the situation?
Cheers,
- Andreas
--
John M. McIntosh johnmci@smalltalkconsulting.com Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================
Andreas Raab ha scritto:
John M McIntosh wrote:
Really I suspect the VM version numbers should fully decouple from the image version numbers, really nothing requires one to use a 3.10 image to build a VM that will work with a 3.10, I'm quite happy to continue working with my 3.8 VMMaker image...
And of course the story comes full circle once people realize that the reason for naming the VMs according to image versions was precisely in response to complaints about how confusing it is that the versions differ ;-)
Heh. We could keep the current version as an "internal" version, and add an "external" version that will stay in sync with the image version.
What do you think?
Giovanni
We can use version strings like following format:
Squeak VM xx.yy.zz (images ver aa.bb - cc.dd)
This makes it absolutely understandable, that VM having own version , AND works/compatible with squeak images of given versions.
2008/6/1 Giovanni Corriga giovanni@corriga.net:
Andreas Raab ha scritto:
John M McIntosh wrote:
Really I suspect the VM version numbers should fully decouple from the image version numbers, really nothing requires one to use a 3.10 image to build a VM that will work with a 3.10, I'm quite happy to continue working with my 3.8 VMMaker image...
And of course the story comes full circle once people realize that the reason for naming the VMs according to image versions was precisely in response to complaints about how confusing it is that the versions differ ;-)
Heh. We could keep the current version as an "internal" version, and add an "external" version that will stay in sync with the image version.
What do you think?
Giovanni
On 1-Jun-08, at 6:12 AM, Igor Stasenko wrote:
We can use version strings like following format:
Squeak VM xx.yy.zz (images ver aa.bb - cc.dd)
This makes it absolutely understandable, that VM having own version , AND works/compatible with squeak images of given versions.
You couldn't actually specify the *last* version of image that the vm would support since you can't know that in advance.
We could however number VMs based (for example) on the revision number of the VMMaker release and the SVN repository (so something like 112-675) for development systems and then simply rename them for putting on the squeak.org download section when ready. Actually, it would require an easy way to change the version number returned to the image for getAttributes as well, so a way of recompiling only the version info and relinking would be best. Perhaps some simply header file that has just the version number string? Normal make would create it with vmmaker-svn and release make would just do 'squeakrelease'. Or whatever. I actually used to something rather like that for GBH back in the 80s.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Half the people you know are below average.
On 31-May-08, at 11:27 PM, John M McIntosh wrote:
At the moment I'm fully occupied with a VM rewrite, post Sophie era, so I'll take hints on what the new macintosh version should be called...
Oh?
Well, I'll be happy to help you figure out a name. It should be something appropriate to the new VM you're writing... what's new and cool about it compared to the old VM? What's the motivation for the rewrite? ;-)
Colin
But, the previous mail is a warning, please, "More atention to open source platform",
if you compare the squeak implementation in windows, mac and linux you will see alot of diferents, such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action.
another difference is the MIDI support which has been lacking on linux for ages.
but, on the other hand, OSProcess/CommandShell seems much more efficient on linux than on Windows.
I don't think there is a deliberate lack of interest towards a specific OS...
Stef
On Fri, May 30, 2008 at 5:57 AM, Diogenes Moreira diogenes.moreira@gmail.com wrote:
such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action.
There was discussion regarding this issue about a year ago. I thought someone, (it might have been Ian), came up with a potential solution that got us part of the way. You might check the archives. I can't remember the final outcome.
Obviously, you should close squeak from the world menu to circumvent the problem.
brad
On 30-May-08, at 9:25 AM, Brad Fuller wrote:
On Fri, May 30, 2008 at 5:57 AM, Diogenes Moreira diogenes.moreira@gmail.com wrote:
such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action.
There was discussion regarding this issue about a year ago. I thought someone, (it might have been Ian), came up with a potential solution that got us part of the way. You might check the archives. I can't remember the final outcome.
IIRC the conclusion was that there is no very useful way to do it that would work for most, let alone all, versions of unix. Or something like that.
More generally there is no particular organised bias towards any platform amongst any of the most productive developers. There *is* a slight bias these days towards 'more modern OSs' in the sense that we expect more support from the OS than we might have done 10 (or 20, or 30) years ago. This is one of the things that lead me to cease work on the Risc OS port, and it might well make life a touch more complex for anyone in the future wanting to port to a very bare platform. But really, if we want to take advantage of things like Cairo, GPUs, sockets... well pretty much anything beyond the truly basic.. then we have to put up with it.
What you have to remember is that almost everyone involved is volunteering their time. That means you have to tolerate the vagaries of their own interests, time availability and tolerance for whiners. Report problems clearly, precisely, without stupid complaints about how unfair everyone is being to your narrow little world, with examples to help recreate the problem and with debug logs whenever possible. Then occasionally remind us of the problem, politely. And if the problem goes away because of some OS fix or machine change, let us know so we don't waste still more time.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: LTT: Lose Timing Track
On 30.05.2008, at 18:25, Brad Fuller wrote:
On Fri, May 30, 2008 at 5:57 AM, Diogenes Moreira diogenes.moreira@gmail.com wrote:
such as, when you close the world form the window menu. In linux squeak is killed without question, in windows the squeak ask you if you are sure of the action.
There was discussion regarding this issue about a year ago. I thought someone, (it might have been Ian), came up with a potential solution that got us part of the way. You might check the archives. I can't remember the final outcome.
Obviously, you should close squeak from the world menu to circumvent the problem.
There is support for this in the latest Linux VM (3.10). Clicking the title-bar close icon does not immediately close the window anymore. Instead, a window-close-event is sent to the image (the same one that the host window support uses). This works the same in the OS X Carbon VM if you patch its Info.plist file to not disable the red window close button.
The new Etoys image then shows a quit dialog (unless it is run in Sugar, where it does a save-and-quit without a dialog).
I don't think anybody added that support to a regular 3.10 image. The change set is here:
http://tinlizzie.org/updates/etoys/updates/1574WindowEvents-JMM-bf.cs
- Bert -
squeak-dev@lists.squeakfoundation.org
-
Andreas Raab -
Bert Freudenberg -
Brad Fuller -
Colin Putney -
Diogenes Moreira -
Giovanni Corriga -
Igor Stasenko -
John M McIntosh -
José Luis Redrejo -
Norbert Hartl -
Stéphane Rollandin -
tim Rowledge -
Yoshiki Ohshima