- Pharo-dev - lists.pharo.org

Re: [Pharo-project] Remove #class as a Compiler optimization
by Marcus Denker April 5, 2011

April 5, 2011

On Mar 25, 2011, at 5:26 PM, Mariano Martinez Peck wrote: > Hi. I have to say that I hate that the compiler associates a special bytecode for #class and that the VM don't even send the message. I want to be able to overwrite #class in a proxy for example, or to debug it. > > I did some benchmarks disabling such optimization by doing: > > (ParseNode classVarNamed: 'StdSelectors') removeKey: #class ifAbsent: []. > Compiler recompileAll. > > And the difference is NOTHING in my tests (maybe I am doing something wrong). > > With Standard VM and #class optimization > > [SystemNavigation default allObjectsDo: [:each | each class]] timeToRun -> aprox 50 miliseconds > What is the constant over head of doing just the iteration? [SystemNavigation default allObjectsDo: [:each | ]] timeToRun > > What do you think? I am for... for the JIT is is a monomorphic send to a primitive, so the IC will be very effective and should take care that the lookup is actually almost never done. (and the lookup is the only thing the #class bytecode saves, as the target method is a primitive). Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

5 8

[Pharo-project] Hudson is not friendly to show us the used scripts ?
by Mariano Martinez Peck April 5, 2011

April 5, 2011

Maybe it is me that I am not finding it in the web interface. I am really interested in being able to see the scripts that Hudson uses for its own usage. For example, how does Hudson run all the tests from code (not using the TestRunner) ? and how does he write that in the correct output that Hudson wants? It would be nice if all used Hudson scrips are available somehow. Thanks -- Mariano http://marianopeck.wordpress.com

3 2

Re: [Pharo-project] Draft Pharo 1.2.1 Release Announcement
by Marcus Denker April 5, 2011

April 5, 2011

On Apr 5, 2011, at 9:48 AM, Mariano Martinez Peck wrote: > Thanks Sven. I would like to remark the ammount of fixes issues. > Is there an easy why to know this? > The bug tracker can be queried for issues of a certain tag: 57 for Full: http://code.google.com/p/pharo/issues/list?can=1&q=milestone%3D1.2-DevImage and 693 for Core: http://code.google.com/p/pharo/issues/list?can=1&q=milestone%3D1.2 For all versions, we had 3960 issue tracker submissions, out of which 473 are open. Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

2 1

Re: [Pharo-project] Hudson is not friendly to show us the used scripts ?
by Marcus Denker April 5, 2011

April 5, 2011

On Apr 5, 2011, at 10:26 AM, Mariano Martinez Peck wrote: > Maybe it is me that I am not finding it in the web interface. I am really interested in being able to see the scripts that Hudson uses for its own usage. > > For example, how does Hudson run all the tests from code (not using the TestRunner) ? and how does he write that in the correct output that Hudson wants? > > It would be nice if all used Hudson scrips are available somehow. > There are two components: 1) The config of Hudson what to do for a Project. That you can see if you are admin. 2) The scripts used. These are the ones of Lukas: https://github.com/renggli/builder To edit these, one needs a shell-account on the build slave. Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

2 1

[Pharo-project] ci.pharo-project.org redirect should be changed?
by Mariano Martinez Peck April 5, 2011

April 5, 2011

Hi... currently ci.pharo-project.org redirects to https://pharo-ic.lille.inria.fr/hudson/view/Pharo/ but I think it should be https://pharo-ic.lille.inria.fr/ since everything inside there is Pharo. Pharo is not just the Pharo image. what do you think? -- Mariano http://marianopeck.wordpress.com

3 3

Re: [Pharo-project] Draft Pharo 1.2.1 Release Announcement
by Sven Van Caekenberghe April 5, 2011

April 5, 2011

On 05 Apr 2011, at 09:19, laurent laffont wrote: > On Tue, Apr 5, 2011 at 9:05 AM, Sven Van Caekenberghe <sven(a)beta9.be> wrote: > > On 05 Apr 2011, at 07:09, laurent laffont wrote: > > > Remarks: > > - group changed by categories: Morphic/Polymorph, Tools, Under the hood, ... > > - Mocketry is not included in Pharo 1.2 > > Thanks for the feedback, I applied your suggestions. > > For more publicity we can have a word on the new success stories: Inceptive ERP, Issys tracking and DrGeo in http://www.pharo-project.org/about/success-stories > > Laurent. > > > More comments ? > > Sven OK, I added a paragraph 'Success Stories', have a look: http://homepage.mac.com/svc/Pharo/pharo-1.2.1-release.html Are the download links correct ?? Sven

1 0

[Pharo-project] IPAddressCountry & Iso3166CountryLongitudeLatitude
by Sven Van Caekenberghe April 4, 2011

April 4, 2011

I want to share a couple of classes. Did you ever want to add this trick to your website where you do something based on the country your visitors are from ? Now you can do so in Smalltalk ! T3EasyUtils getIpAddressCountry atAddress: '81.83.7.35' #BE T3EasyUtils getIso3166CountryLongitudeLatitude at: #BE 4.0(a)50.8333 #('www.beta9.be' 'www.inria.fr' 'nic.st' 'www.world.st' 'www.pharo-project.org') collect: [ :each | T3EasyUtils getIpAddressCountry atAddress: (NetNameResolver addressForName: each) ] #(#BE #FR #SE #US #CH) (T3EasyUtils is holding unique, loaded instances of IPAddressCountry & Iso3166CountryLongitudeLatitude). As this is part of one of our commercial applications, you can even access it over the internet as a JSON REST call: $ curl http://easy.t3-platform.net/rest/geo-ip?address=81.83.7.35 {"latitude":50.8333,"address":"81.83.7.35","country":"BE","longitude":4.0} (Please note that this is not a public service as such, you can try it, but you are not supposed to call this from your own applications.) BTW, T3 Easy is our new entry level track & trace product, an HTML5 web app (using client side HTML+Javascript+CSS) on top of a REST server written in Pharo Smalltalk. It consists of 4 stateless Pharo Cog VM's running behind an Apache load balancer. The server uses Zinc HTTP Components, both for its server part, as well as for clients talking to other systems behind the scenes. Even the Open Street Map tile serving, which is pushing out many megabytes, is currently going through this server. Here is a live demo URL in case your are curious: http://easy.t3-platform.net?t=livedemo (You need a fast web browser like Google Chrome, Safari, Opera 10 or Firefox 4.) The code for IPAddressCountry & Iso3166CountryLongitudeLatitude is on SqueakSource: http://www.squeaksource.com/ADayAtTheBeach.html These are the class comments: --- IPAddressCountry maps IP addresses to their country code. I hold a collection of IPAddressRangeCountry objects that I do a binary search on. I am constructed based on a CSV data file using my class' #readFrom: which expects lines like 3651886848,3651887103,BE or start,stop,code You can download a version of this file from http://homepage.mac.com/svc/data/GeoIPCountry.csv.gz ((IPAddressCountry readFrom: 'GeoIPCountry.csv') atAddress: '81.83.7.35') = #BE Addresses can also be specified as #(81 83 7 35) for #atAddress: or as 1364395811 for #at: You should cache one instance of me. Based on data from http://www.maxmind.com/app/geolitecountry --- Iso3166CountryLongitudeLatitude maps country codes to their average or central longtiude/latitude. My keys are two letter Iso3166 Symbols representing a country, such as #BE My values are Points holding coordinates as longtiude@latitude I am constructed based on a CSV data file using my class' #readFrom: which expects lines like BE,50.8333,4.0000 or code,latitude,longitude You can download a version of this file from http://homepage.mac.com/svc/data/CountryLatLong.csv.gz ((Iso3166CountryLongitudeLatitude readFrom: 'CountryLatLong.csv') at: #BE) = 4.0(a)50.8333 You should cache one instance of me. --- Sven

5 7

April 4, 2011

Hi guys We found a problem with ben and RPackage. Here is the scenario and may be you can help us finding the right solutions. Nautilus renames a class the class got renamed -> an announcement is raised - nautilus gets notified, it asks rpackageOrganizer (which also listens to the same event) -> rpackageOrg breaks because the class if was referting to does not exist anymore. Problem 1 - the second announcement was never sent, because the first one broke the second was blocked. >> we should make sure that if an announcement leads to an error, other annoucements on the same emit should pass Problem 2 If rappckageOrganizer would be reached before Nautilus we would not get any problem (beside the problem 1). - we spent one hour playing with alternatives. We thought that if RPackage would points to classes a rename would not break the system. While this is true for class, the same problem will occur for method rename since compiled method are not shared and a new compiled method is compiled. - If RPackage was as deep inside the system as superclass/subclass, the invariant would be directly maintained by the system and we would not have to rely on announcement. But now we should find a solution. >> may be we should be able to say that an announcement has a priority. Typically we would like to say that when RPackageOrganizer register interests it should be served first. we could have a systemLevel priority that could be specified on registration. what do you think? Stef

7 24

[Pharo-project] [COTDC] 41 - ExclusiveWeakMessageSend
by laurent laffont April 4, 2011

April 4, 2011

Today: ExclusiveWeakMessageSend Comment Of The Day Contest - One Day One Comment Rules: #1: Each day a not commented class is elected. Each day the best comment will be integrated with name of the author(s). #2: If you cannot comment it, deprecate it. Results: http://code.google.com/p/pharo/wiki/CommentOfTheDayContest Laurent

1 1

[Pharo-project] On Pharo 1.2
by Torsten Bergmann April 4, 2011

April 4, 2011

Sorry for changing topic ... but with all these [ANN] replies we cant see if there are really new announcements. When 1.2 Hudson was red nobody seemed to care and now that it is green/blue again discussion starts. Good or bad? At least I wished that happened earlier and a red Hudson shows that we have to do our homework. Pharo 1.2 really took time and I think we can expect further releases to take even longer since the more we clean up the harder it would be to keep compatibility with all the existing (old) packages. We all know a core alone would be useless - we have to decide which packages are maintainable by the community and worth to move forward with Pharo-dev. We have two goals that sometimes collide: -> build a small kernel/core -> build a better Smalltalk (feature rich and with good tools) If we work on cleaned up loadable/unloadable and well tested packages it would be less problematic since we can work in both images and load/unload as needed. That should be the main goal for 1.3. and beyond. Hudson is our watchdog. Nonetheless ... with Pharo 1.1 the open source ST world already changed and we have first companies/people using Pharo. Pharo is an accepted Smalltalk version. Since Pharo 1.2 is a much better and cleaner version I hope this will continue/raise and Pharo will find more new friends to help. :) Bye T. -- NEU: FreePhone - kostenlos mobil telefonieren und surfen! Jetzt informieren: http://www.gmx.net/de/go/freephone

1 0