- Pharo-dev - lists.pharo.org

Re: [Pharo-project] 'foo' asTime

by Mariano Abel Coca

As an off topic of 3 seconds... Have you considered to include HernÃ¡n Wilkinson and Maximiliano Taborda's packages Aconcagua and Chalten in Pharo Dev? Chalten is a great Time framework and much better than the ansi Smalltalk 80's Time class, even with the addition of Duration. And it represents the Minute and Second concepts along with Month, Day, DayOfMonth, MonthOfYear, etc. :) With Chalten you could simply write: December second, 2005. If you want to try them, just evaluate: Gofer new squeaksource: 'Chalten'; package: 'Chalten'; load. Cheers, Mariano. On Mon, Aug 23, 2010 at 5:45 AM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr > wrote: > 3 seconds is indeed better. > Right now we are talking about string over extension. > > Stef > > On Aug 23, 2010, at 9:00 AM, Brent Pinkney wrote: > > >> I would really like to have the approach proposed by K. Beck on > conversion > >> methods. Conversion methods for API compatible objects. Of course it > will > >> lead to less compact program (URL readFrom: 'http://' instead of > 'http://' > >> url) but avoids a lot of code in String (of course class extensions > limit > >> the plague). > > > > Hi, please reconsider this. There is just no way (Duration seconds: 3) is > > better than '3 seconds'. > > > > This smacks of working for the computer instead of the other way around. > > > > Brent > > > > _______________________________________________ > > 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 >

Aug. 23, 2010

Re: [Pharo-project] [Metacello] Re: Loading platform-specific code

by Dale Henrichs

Yes it's getting to be about that time:) Dale Mariano Martinez Peck wrote: > > > On Mon, Aug 23, 2010 at 6:45 PM, Dale Henrichs <dhenrich(a)vmware.com > <mailto:dhenrich@vmware.com>> wrote: > > Let's talk about OS dependency at ESUG... > > > we should write our to-talk list hahahah > > > Dale > > Mariano Martinez Peck wrote: > > > > On Tue, Aug 10, 2010 at 3:03 PM, Sean P. DeNigris > <sean(a)clipperadams.com <mailto:sean@clipperadams.com> > <mailto:sean@clipperadams.com <mailto:sean@clipperadams.com>>> > wrote: > > > > Mariano Martinez Peck wrote: > > > > ...did you know that Metacello let you > > handle platform specific code? > > > I'm assuming you mean [Squeak|Pharo...] platform - we're talking > [Mac|Windows...] > > > > ahhh sorry....and what about conditional loading? and using > spec:for or something like that? > > once I suggested Dale to add OS dependency to Metacello...I > don't rememeber what happened. > > > > Mariano Martinez Peck wrote: > > > > Thus, you can define this in the > ConfigurationOFfExternalWebBrowser, and > > not > > depend on a class side initialize. > > > Yes, it all started when I decided to move the load out of the > class>>initialization and into the Config, and I just wanted > to poll > to see > if my thinking was correct. > > > Norbert Hartl wrote: > > > > I think I don't fully understand what you mean with > "lazily load" and > > "required for other platforms". But if it is just a choice > when > to load > > something then I would do while having a internet > connection ;) > > > Yes, that is the choice, and that answers the question. Thanks. > > Sean > -- > View this message in context: > > http://forum.world.st/Loading-platform-specific-code-tp2318610p2319751.html > Sent from the Pharo Smalltalk mailing list archive at Nabble.com. > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > <mailto:Pharo-project@lists.gforge.inria.fr> > <mailto:Pharo-project@lists.gforge.inria.fr > <mailto:Pharo-project@lists.gforge.inria.fr>> > > > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > >

Aug. 23, 2010

[Pharo-project] ExternalAddress and block copy; ByteArray and external calls?

by Schwab,Wilhelm K

Are there any well-worn ways to move data into and out of external memory? The best I have been able to uncover are recommendations to either do it one byte at a time (yuk!) or to via bitblt. Those posts come from several years ago, and include mention that ExternalAddress does not clean up after itself. Looking in Pharo now, I see #finalize that calls free; have things move along to avoid memory leaks? There are no conditions in #finalize or #free, which raises questions about whether it is safe to send #free to quickly free memory that will no longer be used? In Dolphin, I reflexively free ODBC related handles because they quickly pile up and cause problems despite the finalization "water mark." I am currently working with a data acquisition board and will be needing memory for clearly defined periods. I also need to get data out of the buffers, which sent me looking for an official way to do so. To what extent, if any, can String and ByteArray be used as buffers for external calls? Dolphin made that robust, but there were hints that the gc was slightly tamed to allow it. The little reading I have done in squeak-dev archives suggests that squeak/pharo is not so forgiving. Is that true? Is there a way to request a byte array allocated in a fixed (or whatever the jargon is) heap so that external calls will be safe? Bill

Aug. 23, 2010

Re: [Pharo-project] [Pharo-users] 3600 seconds seconds = 0 ???

by Mariano Martinez Peck

are you sure you wanted pharo-users or pharo-dev? On Mon, Aug 23, 2010 at 6:04 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr > wrote: > 3600 seconds seconds = 0 > which means that this is a number of seconds in the hour:min:seconds > decomposition > > Now it looks to me that the method comment is wrong > > seconds > "Answer the number of seconds the receiver represents." > > ^ (seconds rem: SecondsInMinute) > > > 3600 seconds asSeconds = 3600 is what is working > > may be (Duration seconds: 3600) would be better > > To me there is something conceptually bogus with such definition of minutes > and seconds. > (Duration seconds: 3600) minutes = 0 > (Duration seconds: 3600) seconds = 0 > > > > Stef > > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users >

Aug. 23, 2010

Re: [Pharo-project] [Metacello] Re: Loading platform-specific code

by Mariano Martinez Peck

On Mon, Aug 23, 2010 at 6:45 PM, Dale Henrichs <dhenrich(a)vmware.com> wrote: > Let's talk about OS dependency at ESUG... > > we should write our to-talk list hahahah > Dale > > Mariano Martinez Peck wrote: > >> >> >> On Tue, Aug 10, 2010 at 3:03 PM, Sean P. DeNigris <sean(a)clipperadams.com<mailto: >> sean(a)clipperadams.com>> wrote: >> >> >> >> Mariano Martinez Peck wrote: >> > >> > ...did you know that Metacello let you >> > handle platform specific code? >> > >> I'm assuming you mean [Squeak|Pharo...] platform - we're talking >> [Mac|Windows...] >> >> >> >> ahhh sorry....and what about conditional loading? and using spec:for or >> something like that? >> >> once I suggested Dale to add OS dependency to Metacello...I don't >> rememeber what happened. >> >> >> >> Mariano Martinez Peck wrote: >> > >> > Thus, you can define this in the >> ConfigurationOFfExternalWebBrowser, and >> > not >> > depend on a class side initialize. >> > >> Yes, it all started when I decided to move the load out of the >> class>>initialization and into the Config, and I just wanted to poll >> to see >> if my thinking was correct. >> >> >> Norbert Hartl wrote: >> > >> > I think I don't fully understand what you mean with "lazily load" >> and >> > "required for other platforms". But if it is just a choice when >> to load >> > something then I would do while having a internet connection ;) >> > >> Yes, that is the choice, and that answers the question. Thanks. >> >> Sean >> -- >> View this message in context: >> >> http://forum.world.st/Loading-platform-specific-code-tp2318610p2319751.html >> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> <mailto:Pharo-project@lists.gforge.inria.fr> >> >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> >

Aug. 23, 2010

Re: [Pharo-project] How can I load metacello without OB

by Dale Henrichs

In the beginning, Metacello had an abstract class for configurations. but it became a chicken and egg problem (you had to get Metacello loaded into the image before you could work with a configuration). The current scheme was invented to get around that particular problem. Adding a method to Object only works if Squeak adds the method. If changes are to be made to the base. If changes are going to be made to the base then an abstract class should be added and be done with. I guess it is worth mentioning that the ConfigurationOf classes are really just smart versions of XML ... In other words all of the information in a configuration can be represented in a text file ... using Smalltalk means that every image has a built in configuration file editor and a built in configuration file compiler. Using Smalltalk means that there Monticello repositories that can be used to share and distribute configurations... Are we reaching the point where our effort should go into creating tools for building/managing/storing configuration files? Which would require that some bootstrapping code be loaded into each of the Smalltalks for Metacello support... I mention this because I think that there will be pain involved in changing the "format" of Metacello configurations. Right now, Metacello is basically self-contained and non-intrusive at the cost of a somewhat ugly configuration creation process... Should we consider the fact that if we are going to experience pain, maybe it should be the "last pain." In other words, lets make a change that is aimed at the "final format" for Metacello configurations rather than an intermediate format that will be used for awhile and then changed again... Dale Mariano Martinez Peck wrote: > > > On Mon, Aug 23, 2010 at 6:32 PM, Dale Henrichs <dhenrich(a)vmware.com > <mailto:dhenrich@vmware.com>> wrote: > > Is Gofer preloaded in Squeak? > > The #ensureMetacello is intended to work on all platforms. For > earlier versions of Pharo, gofer had to be loaded before Metacello > could be loaded and if Gofer isn't preloaded in Squeak, then we > still have that problem. > > I guess I should say that _I_ have that problem, because I am > interested in configurations that can be used with any Smalltalk > that supports Monticello:) > > > I want to say a couple of things: > > 0) Objective? Don't need to write #ensureMetacello in each conf as it is > 99% times the same and we don't want to update each conf to the latest > ensureMetacello that is in the wiki > > 1) I don't know if Gofer works or is preloaded in Squeak. But what I > really think is that if Squeak wants to "officially" supports Metacello, > then they should be ready to integrate things. Ok, maybe not a complete > package like Gofer, but at least some messages or something that would > be of help. > > 2) What about providing a solution that works for different cases. For > example, the following come to my mind (of course, hacky, but you can > make it better for sure): > > - we can creat Object>>ensureMetacello in squeak, phaor, gemstone. > Okokoko, puting those methods in Object sucks..but if you have a better > idea :) > > Then, in Pharo we can implement it this way: > > Object >> ensureMetacello > Gofer ensureMetacelloCore > > > Gofer >> ensureMetacelloCore > > "Bootstrap Gofer (if necessary), load latest mcz file for > ConfigurationOfMetacello (using old > Gofer API), then load the latest version of Metacello itself." > > Smalltalk at: #MetacelloProject ifAbsent: [ | error | > "list of repositories to try, in case primary repository > is not accessible" > (Array > with: > 'http://www.squeaksource.com/MetacelloRepository' > with: > 'http://seaside.gemstone.com/ss/metacello') do: [:repositoryUrl | > ([ > Smalltalk at: #Gofer ifAbsent: [ > "Current version of Gofer from > which to bootstrap - as of 1.0-beta.21" > self bootstrapPackage: > 'Gofer-Core-lr.115' from: repositoryUrl ]. > Smalltalk at: #Gofer ifPresent: > [:goferClass | | gofer | > gofer := goferClass new > url: repositoryUrl; > yourself. > [ gofer addPackage: > 'ConfigurationOfMetacello' ] on: Warning do: [:ex | ex resume ]. > gofer load ]] > on: Error > do: [ :ex | > error := ex. > Transcript > cr; > > show: 'failed ensureMetacello: '; > > show: ex description printString; > > show: '...retrying'. > "try again" > ex > return: nil ]) ~~ nil > ifTrue: [ > "load 'default' group of > Metacello" > (Smalltalk at: > #ConfigurationOfMetacello) perform: #load. > ^self ]]. > "shouldn't get here unless the load failed ... throw an > error" > self error: 'retry with alternate repository failed: ', > error description printString ] > > > here we can even remove the ifFalse if Gofer is not present...since in > Pharo it is present. > > So...in Pharo, if you have specified ensureMetacello in your Conf...then > perfect. That one will be use. If you don't want to implement it and let > it as default (like 99% of the cases), self ensureMetacello would be > object, which will delegate to Gofer, who will load the Metacello Core. > > In Squeak, it can be as it is now. They can use they own ensureMetacello > (the difference here is that they HAVE to implement such method as they > don't have it in they own image). > Maybe they can even implement ensureMetacello in object using Installer > or whatever they have in the core to load monticello packages. > > If #ensureMetacello has to be updated we only need to change > Gofer>>ensureMetacello and not all the confs. > > Those who need to load a different Metacello version, different groups, > or from different repo, they just implement #ensureMetacello in their > OWN conf and that's all. For example, I can do this in > ConfigurationOfPharo as I need not only core, but UI and tests. > > Now that I am thinking it...my "idea" is outside Metacello. The problem > is that if someone writes a conf, without #ensureMetacello, it will load > fine in Pharo but will fail in Squeak. > > what do you think?

Aug. 23, 2010

Re: [Pharo-project] How can I load metacello without OB

by Dale Henrichs

StÃ©phane Ducasse wrote: > may be can fork metacello? > > :) Actually you'd be forking the configurations, not Metacello Dale

Aug. 23, 2010

Re: [Pharo-project] Working on Pharo 1.2

by Marcus Denker

On Aug 23, 2010, at 7:45 PM, Adrian Lienhard wrote: > OK, I created: > > http://www.pharo-project.org/pharo-download/unstable-core > > Marcus, for updating: log into Cmsbox, select the page 'unstable-core' under 'download' and edit the code part on this page. > Thanks, that's nice! Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Aug. 23, 2010

Re: [Pharo-project] How can I load metacello without OB

by Mariano Martinez Peck

On Mon, Aug 23, 2010 at 6:32 PM, Dale Henrichs <dhenrich(a)vmware.com> wrote: > Is Gofer preloaded in Squeak? > > The #ensureMetacello is intended to work on all platforms. For earlier > versions of Pharo, gofer had to be loaded before Metacello could be loaded > and if Gofer isn't preloaded in Squeak, then we still have that problem. > > I guess I should say that _I_ have that problem, because I am interested in > configurations that can be used with any Smalltalk that supports > Monticello:) > > I want to say a couple of things: 0) Objective? Don't need to write #ensureMetacello in each conf as it is 99% times the same and we don't want to update each conf to the latest ensureMetacello that is in the wiki 1) I don't know if Gofer works or is preloaded in Squeak. But what I really think is that if Squeak wants to "officially" supports Metacello, then they should be ready to integrate things. Ok, maybe not a complete package like Gofer, but at least some messages or something that would be of help. 2) What about providing a solution that works for different cases. For example, the following come to my mind (of course, hacky, but you can make it better for sure): - we can creat Object>>ensureMetacello in squeak, phaor, gemstone. Okokoko, puting those methods in Object sucks..but if you have a better idea :) Then, in Pharo we can implement it this way: Object >> ensureMetacello Gofer ensureMetacelloCore Gofer >> ensureMetacelloCore "Bootstrap Gofer (if necessary), load latest mcz file for ConfigurationOfMetacello (using old Gofer API), then load the latest version of Metacello itself." Smalltalk at: #MetacelloProject ifAbsent: [ | error | "list of repositories to try, in case primary repository is not accessible" (Array with: ' http://www.squeaksource.com/MetacelloRepository' with: 'http://seaside.gemstone.com/ss/metacello') do: [:repositoryUrl | ([ Smalltalk at: #Gofer ifAbsent: [ "Current version of Gofer from which to bootstrap - as of 1.0-beta.21" self bootstrapPackage: 'Gofer-Core-lr.115' from: repositoryUrl ]. Smalltalk at: #Gofer ifPresent: [:goferClass | | gofer | gofer := goferClass new url: repositoryUrl; yourself. [ gofer addPackage: 'ConfigurationOfMetacello' ] on: Warning do: [:ex | ex resume ]. gofer load ]] on: Error do: [ :ex | error := ex. Transcript cr; show: 'failed ensureMetacello: '; show: ex description printString; show: '...retrying'. "try again" ex return: nil ]) ~~ nil ifTrue: [ "load 'default' group of Metacello" (Smalltalk at: #ConfigurationOfMetacello) perform: #load. ^self ]]. "shouldn't get here unless the load failed ... throw an error" self error: 'retry with alternate repository failed: ', error description printString ] here we can even remove the ifFalse if Gofer is not present...since in Pharo it is present. So...in Pharo, if you have specified ensureMetacello in your Conf...then perfect. That one will be use. If you don't want to implement it and let it as default (like 99% of the cases), self ensureMetacello would be object, which will delegate to Gofer, who will load the Metacello Core. In Squeak, it can be as it is now. They can use they own ensureMetacello (the difference here is that they HAVE to implement such method as they don't have it in they own image). Maybe they can even implement ensureMetacello in object using Installer or whatever they have in the core to load monticello packages. If #ensureMetacello has to be updated we only need to change Gofer>>ensureMetacello and not all the confs. Those who need to load a different Metacello version, different groups, or from different repo, they just implement #ensureMetacello in their OWN conf and that's all. For example, I can do this in ConfigurationOfPharo as I need not only core, but UI and tests. Now that I am thinking it...my "idea" is outside Metacello. The problem is that if someone writes a conf, without #ensureMetacello, it will load fine in Pharo but will fail in Squeak. what do you think? cheers Mariano Dale > > > Tudor Girba wrote: > >> Indeed, now we could add this method in Gofer. >> >> Cheers, >> Doru >> >> >> On 20 Aug 2010, at 16:01, Mariano Martinez Peck wrote: >> >> >>> On Fri, Aug 20, 2010 at 3:54 PM, StÃ©phane Ducasse < >>> stephane.ducasse(a)inria.fr >>> >>>> wrote: >>>> >>>> The problem is in ConfigurationOfRPackage as in #ensureMetacello >>>> >>> is like this: >>> >>>> ensureMetacello >>>> Smalltalk at: #MetacelloProject ifAbsent: [ >>>> Gofer new >>>> disableRepositoryErrors; >>>> url: 'http://seaside.gemstone.com/ss/metacello'; >>>> package: 'ConfigurationOfMetacello'; >>>> load. >>>> (Smalltalk at: #ConfigurationOfMetacello) project >>>> >>> latestVersion load: #('default' 'UI') ] >>> >>> I hate the fact that we should copy all the code around.... >>> what a crap. >>> >>> This was discussed several, severaaaaal times. As I remember the problem >>> was that otherwise you class depends on a base class that should be >>> included in the image you are working on. >>> But I guess we don't have problem to add such base class in >>> PharoCore...if we have Gofer.. >>> >>> >>> >>>> If you see, it is loading #('default' 'UI') of metacello..... UI - OB. >>>> >>>> change that to #('default') and wala >>>> >>>> BTW, I would use the latest recommended #ensureMetacello: >>>> >>>> http://code.google.com/p/metacello/wiki/EnsureMetacello >>>> >>>> cheers >>>> >>>> mariano >>>> >>>> >>>> And stuck because OB does not load in 1.2 so I lost one hour...... >>>> >>> boring. >>> >>>> Stef >>>> >>>> >>>> _______________________________________________ >>>> 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 >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>> >> >> -- >> www.tudorgirba.com >> >> "Live like you mean it." >> >> >> _______________________________________________ >> 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 >

Aug. 23, 2010

Re: [Pharo-project] Working on Pharo 1.2

by Adrian Lienhard

OK, I created: http://www.pharo-project.org/pharo-download/unstable-core Marcus, for updating: log into Cmsbox, select the page 'unstable-core' under 'download' and edit the code part on this page. Cheers, Adrian On Aug 23, 2010, at 08:31 , Marcus Denker wrote: > > On Aug 22, 2010, at 9:54 PM, Adrian Lienhard wrote: > >> Yes, this is possible (its done with a code part in Cmsbox that triggers the redirect, so updating is similar to editing some other content). But we should then regularly update the links so they don't get outdated. Is it worth? >> > > good question... right now people can't easily find the last pre-built unstable. I would update the link as soon as I put a new pre-built online. > > (of course this is just an intermediate step... the goal is to auto-build everything. Which we will have at some point... but if we learning anything > the last 10 years is that because one could (and plans to) do it better, one should nevertheless do the dumb intermediate solution nevertheless) > > Marcus > > -- > Marcus Denker -- http://www.marcusdenker.de > INRIA Lille -- Nord Europe. Team RMoD. > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Aug. 23, 2010