January 2013 - Pharo-users - lists.pharo.org

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by StÃ©phane Ducasse

andy thanks for this discussion. It would be good to raise a kind of consensus because this is an effort that the association could support. Stef On Jan 26, 2013, at 11:41 AM, Andy Burnett wrote: >> he, sorry for insist, but cloudfoundry already does that... and is opensource too :) >> at ESUG I seen James present it to deploy pharo images... so is viable :) >> >> cheers, >> Esteban >> > > I agree. And, I was really quite optimistic about cloud foundry when > it first came out. I particularly like the way you can have a local > micro cloud, and then transition to the hosted cloud. My only > disappointment was that VMware are not - seemingly - going to offer > gemstone as an option on cloudfoundry.com. > Building on what Janko said, I think it is really important to join > the cloud 'wave'. > > The question is how might we - collectively - make this happen? I am > perfectly happy to go with cloud foundry, or open shift, or any other > option if more knowledgeable people can show that X is the way to go. > My only requirement is that I don't want to deal with anyone lower > than pharo. All OS patches need to be done automagically by the > hosting provider. I think cloudfoundry.com and open shift do that, but > my understanding of cloudfoundry.org is that this is a 'roll it > yourself' solution. Obviously, that would require someone to step up > and do that for us. > My preference would be for us to create a suitably configured module > for one of e platforms, and we can then start getting experience with > it. Stephane, is this something that the consortium might take a lead > on? My company is happy to pay some of the costs, if this will help. > > Cheer > Andy >

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Francois Stephany

On 26/01/13 02:02, Andy Burnett wrote: > Our requirements are pretty simple. We need a way of uploading an image, > and some standard for of persistence. At this point, we would be > perfectly happy to make compromises by agreeing that it had to be e.g. > Mongo or whatever (although ideally it would be gemstone). We just want > to make some progress on this. I would prefer to upload a MetacelloConfiguration (or something similar) instead of an image. I like to think about the image as a temporary runtime. Not something where I store code or someting I deploy. I would love to have something like: $ pharo -v 1.4-summer myProject.ston where myProject.ston would contain all the necessary information to load my project (dependencies, repositories, etc). This command would create an image with my project on the fly from a stock Pharo 1.4-summer and launch it. It would recreate the image everytime I want to work on my project (or when it is deployed). The problem right now is the loading time for dependencies (try to load seaside in a pristine image, you'll get the point). All the other bricks are there.

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Andy Burnett

> he, sorry for insist, but cloudfoundry already does that... and is opensource too :) > at ESUG I seen James present it to deploy pharo images... so is viable :) > > cheers, > Esteban > I agree. And, I was really quite optimistic about cloud foundry when it first came out. I particularly like the way you can have a local micro cloud, and then transition to the hosted cloud. My only disappointment was that VMware are not - seemingly - going to offer gemstone as an option on cloudfoundry.com. Building on what Janko said, I think it is really important to join the cloud 'wave'. The question is how might we - collectively - make this happen? I am perfectly happy to go with cloud foundry, or open shift, or any other option if more knowledgeable people can show that X is the way to go. My only requirement is that I don't want to deal with anyone lower than pharo. All OS patches need to be done automagically by the hosting provider. I think cloudfoundry.com and open shift do that, but my understanding of cloudfoundry.org is that this is a 'roll it yourself' solution. Obviously, that would require someone to step up and do that for us. My preference would be for us to create a suitably configured module for one of e platforms, and we can then start getting experience with it. Stephane, is this something that the consortium might take a lead on? My company is happy to pay some of the costs, if this will help. Cheer Andy

Jan. 26, 2013

Re: [Pharo-users] How do i list the execution order of statements inside a method?

by mircea

Visitor pattern. I read about this pattern, and i can understand it. Again i know the "rules and conventions" but i'm a novice player. I won't detail my thought pattern againâ¦ it's based on trial and errorâ¦ I found a match, an example that i can relate to from experience. "Consider the design of a 2D CAD system. At its core there are several types to represent basic geometric shapes like circles, lines and arcs. The entities are ordered into layers, and at the top of the type hierarchy is the drawing, which is simply a list of layers, plus some additional properties. A fundamental operation on this type hierarchy is saving the drawing to the system's native file format. At first glance it may seem acceptable to add local save methods to all types in the hierarchy. But then we also want to be able to save drawings to other file formats, and adding more and more methods for saving into lots of different file formats soon clutters the relatively pure geometric data structure we started out with. A naive way to solve this would be to maintain separate functions for each file format. Such a save function would take a drawing as input, traverse it and encode into that specific file format. But if you do this for several different formats, you soon begin to see lots of duplication between the functions, e.g. lots of type-of if statements and traversal loops. Another problem with this approach is how easy it is to miss a certain shape in some saver. Instead, you could apply the Visitor pattern. The Visitor pattern encodes a logical operation on the whole hierarchy into a single class containing one method per type. In our CAD example, each save function would be implemented as a separate Visitor subclass. This would remove all duplication of type checks and traversal steps. It would also make the compiler complain if you leave out a shape. Another motivation is to reuse iteration code." Based on the above text i can extract the basic pattern of interaction and apply it to our example. Essentially "passing a visitor" as you say would involve me creating a class that has a method or methods for each type of object i want to "draw a picture" for. Per from your text, i would pass this visitor to the AST to produce these images from the input that AST gets from some example code or something. Please be so kind as to point out anything that i might have misunderstood. Pe 26.01.2013, la 13:41, StÃ©phane Ducasse [via Smalltalk] <ml-node+s1294792n4665526h11(a)n4.nabble.com> a scris: > > Print all the methods on small pieces of paper and try to understand the thing as a whole first. Then focus on understanding the patterns and reason behind their use. > > > > Do you know what is a visitor? > If not you should read the pattern because on a AST you can pass a visitor and it can produce many different shapes -- View this message in context: http://forum.world.st/How-do-i-list-the-execution-order-of-statements-insid… Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Andy Burnett

> > you could use cloudfoundry for that :) > > http://www.cloudfoundry.org/ > > Esteban Yup, cloudfoundry could be great too. Particularly if it made it easier to get gemstone running. I think it is important that we try to reach a consensus on this, purely to minimise duplicated work. It would be great if - for example - the pharo consortium were able to issue a new cartridge when they issue a new pharo release. So that the hosted version of pharo is automatically kept up to date. My vision is that we end up with a one-click hosting model in just the same way we have one click pharo. That would be fantastic :-)

Jan. 26, 2013

Re: [Pharo-users] How do i list the execution order of statements inside a method?

by mircea

I will reply for each. Detailing my thought pattern. I know all the "rules and conventions" of smalltalk but i "can't play" very well. At first (Point>>#x) parseTree explore just looks weird. Then things start to settle. () Paranthesis means i don't actually have to care what is in there, conclusion is whatever happens the product of () will be an object. I'm actually scared of >>, since i don't know what it doesâ¦ i've seen it in old code snippets of class definitions. parseTree seems to be a method, and so does explore. Going onâ¦ it seems to be a chain of unary messages. Based on first trial conclusion it might be a statement even though the "." at the end is missing. I paste in the Workspace and right click 'Do It' on the whole thing. A window opens with lots of information in it. I can't make sense of it since i have no idea what i just did. I need more information on what i just did. Restart from the beginning. "Point" might be a point like in a cardinal systemâ¦ x is there so i might be right. I replace x with y and 'Do It' againâ¦ another window pop openâ¦ seems to be quite similar to the first one only it has y in it instead of x. I continue the assumption and replace y with z. I get a KeyNotFound:, i don't understand this but it's similar to MessageNotUnderstood:. Since success would have meant i would get a similar window as in the x y case, getting a different window means the test failed. I can only use x and y in the statement as it is now. Point might be a class, i search for itâ¦ it seems to have instance variable names by that name. I add z to its list of instance variable names, save, replace it in the commandâ¦ error. Assumption was incorrect. I look at the Point class again notice it also has accessor methods by that name, just x and y â¦ there is no z. Agrees with the failure of earlier experiment, i might be on the right track. I choose another method of Point. Execute the command againâ¦ get the "success" window with lots of information on it. I choose another class and one of its methods, execute the command againâ¦ get the "success" windows with lots of information on it. Conclusionâ¦ I can put any class name and choose any method it has and the command will pop up information about if. I don't understand the information from the pop up. Restart from the beginning. I search for parseTree in the browser and there seems to be a lot of implementors of that method. I search for explore in the browser and Object seems to implement it. I 'Inspect It' on just the (Point>>#x) part and it seems to be producing a CompiledMethod object. Based on assumption from the beginning whatever happens between () is an object this object then responds to parseTree. I 'Inspect It' on just (Point>>#x) parseTree and get an RBMethodNode object. Seems the parseTree gets CompiledMethod as input and outputs this different kind of object called RBMethodNode. Adding explore, based on the fact that explore is a method of object, seems to be just a generic printout of this entire object. The end. Time elapsed, a little over 3 min. Took longer to write the email. Pending questions: What exactly does >> do here? I know # is followed by a string literal normally in '' but since there are no spaces just the characters will do. Pe 26.01.2013, la 13:41, StÃ©phane Ducasse [via Smalltalk] <ml-node+s1294792n4665526h11(a)n4.nabble.com> a scris: > > It's like being a two-eyed man in a dark room with no flashlight. Your eyes are there, but without the flashlight there might as well be no room. > > > > Without previous knowledge of how this AST is implemented my best bet is do the "architectural plan" example i gave Dale H. > > (Point>>#x) parseTree explore > will show you the tree then you can navigate in it. -- View this message in context: http://forum.world.st/How-do-i-list-the-execution-order-of-statements-insid… Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Janko MivÅ¡ek

Hi Andy, Dne 26. 01. 2013 02:02, piÅ¡e Andy Burnett: > We would love to develop more of our ideas in Pharo, but we are > completely allergic to doing sys admin. We don't want to deal with > anything below the Pharo application level. What I would ideally like to > see if something like Heroku, or similar, but for Pharo. However, just > wishing won't make it happen, so I am wondering if there is anyone out > there who is thinking about this, and whether there are enough of us to > constitute are market. > > Our requirements are pretty simple. We need a way of uploading an image, > and some standard for of persistence. At this point, we would be > perfectly happy to make compromises by agreeing that it had to be e.g. > Mongo or whatever (although ideally it would be gemstone). We just want > to make some progress on this. > > Personally, I think lots of bits are coming together. We now have > various persistence options, websockets (both in Zinc and Aida-web), > OAuth ( in both places again), and lots of other exciting elements. We > just want a really simple way to take advantage of them. > > Does anyone else feel this need? Certainly! Smalltalk as a dynamic language with environment which enables applications to grow organically (saying simpler: to be updated while running) has actually a big potential in the cloud. To provide web services via REST APIs to the so called front-end clients like mobile apps or client-side web apps. And yes, we have now adequate support for REST API services, including OpenID for authentication and OAuth for authorization. Maybe OpenID/OAuth server would be also nice.. You are probably aware of work by Gemstone guys, specially by James Foster, to put Smalltalk web app on the CloudFoundry [1]. This can be the start and their experiences adapting Gemstone to work in existing cloud infrastructure is valuable to tho the same with Pharo. I'd also go adapting Pharo to some existing cloud infrastructure like Amazon, Heroku, CloudFoudry, ... We seems to don't have enough resources to build and maintain our own ones, as experience with our hosting attempts so far shows. It would be also long-term better to join the cloud wave instead building our own replicas. Because a whole IT world goes that way. Best regards Janko [1] Preparing Smalltalk for Cloud Foundry http://programminggems.wordpress.com/2012/02/14/preparing-smalltalk-for-clo… -- Janko MivÅ¡ek Aida/Web Smalltalk Web Application Server http://www.aidaweb.si

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Esteban Lorenzano

he, sorry for insist, but cloudfoundry already does that... and is opensource too :) at ESUG I seen James present it to deploy pharo images... so is viable :) cheers, Esteban On Jan 26, 2013, at 1:52 PM, Andy Burnett <andy.burnett(a)knowinnovation.com> wrote: > <I am merging my responses to Bernat and Stephane> > > I agree that there are probably too few webdevs to make this viable - > yet :-). However, we may be able to build on other people's work. > > For example, I have been quite impressed by OpenShift. I wondered if > someone could create a Pharo (or gemstone) cartridge. That would then > give us access to the rest if their system. It might then allow anyone > to set up their own system, just by using that cartridge. > > Obviously, someone would need to keep the cartridge up to date etc, > and maybe a first step would be to adapt the work on Harbour to make > image uploading easier. > > At this point, as I am sure you can tell, I am way beyond my > expertise, so I am just guessing. However, it does seem as though > people - such as redhat - are doing a lot if the work for us, and it > could be smart to build on their structures. > > Has anyone tried creating a cartridge? Or would someone like to try? I > am happy to sponsor part of this, if that helps. > > Cheers > Andy >

Jan. 26, 2013

Re: [Pharo-users] Phaas - anyone interested in setting up Pharo as a Service?

by Andy Burnett

<I am merging my responses to Bernat and Stephane> I agree that there are probably too few webdevs to make this viable - yet :-). However, we may be able to build on other people's work. For example, I have been quite impressed by OpenShift. I wondered if someone could create a Pharo (or gemstone) cartridge. That would then give us access to the rest if their system. It might then allow anyone to set up their own system, just by using that cartridge. Obviously, someone would need to keep the cartridge up to date etc, and maybe a first step would be to adapt the work on Harbour to make image uploading easier. At this point, as I am sure you can tell, I am way beyond my expertise, so I am just guessing. However, it does seem as though people - such as redhat - are doing a lot if the work for us, and it could be smart to build on their structures. Has anyone tried creating a cartridge? Or would someone like to try? I am happy to sponsor part of this, if that helps. Cheers Andy

Jan. 26, 2013

tuesday we will do a coding sprint here at Santiago

by StÃ©phane Ducasse

Hi guys We will do a coding sprint here at Santiago @ DCC (Pleaid building) http://pleiad.dcc.uchile.cl We should have a list of issue to fix :) I will check the issues in the bug tracker but any help, wishes from this sprint is welcome. Stef

Jan. 26, 2013