March 2015 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] [Pharo4] Image size now 23.2MB

by Eliot Miranda

On Thu, Mar 5, 2015 at 4:45 AM, phil(a)highoctane.be <phil(a)highoctane.be> wrote: > On Thu, Mar 5, 2015 at 12:05 PM, Sven Van Caekenberghe <sven(a)stfx.eu> > wrote: > > Phil, > > > >> On 05 Mar 2015, at 11:44, phil(a)highoctane.be wrote: > >> > >> Images are starting slow, that's my experience. Here, a Pharo3 on > >> Windows with quite a bunch of things in it takes six seconds to load > >> on first launch and three after that. > >> > >> And that's on a i7 4770K clocked 3.85GHz with SSD drives. > >> > >> It takes less time to revive a VMWare VM :-( > > > > Maybe it is Windows ? Or certain extra stuff that you have in your image > that does work at startup ? Because you should see sub second startup times. > > > > (This is on the smallest Digital Ocean instance, 1 virtual CPU, 512 Mb > RAM, SSD) > > > > root@stfx:~/experiments/pharo4# curl get.pharo.org/40+vm | bash > > % Total % Received % Xferd Average Speed Time Time Time > Current > > Dload Upload Total Spent Left > Speed > > 100 2885 100 2885 0 0 17974 0 --:--:-- --:--:-- --:--:-- > 18031 > > Downloading the latest 40 Image: > > http://files.pharo.org/image/40/latest.zip > > Pharo.image > > Downloading the latest pharoVM: > > http://files.pharo.org/vm/pharo/linux/stable.zip > > pharo-vm/pharo > > Downloading PharoV10.sources: > > http://files.pharo.org/sources//PharoV10.sources.zip > > Downloading PharoV20.sources: > > http://files.pharo.org/sources//PharoV20.sources.zip > > Downloading PharoV30.sources: > > http://files.pharo.org/sources//PharoV30.sources.zip > > Creating starter scripts pharo and pharo-ui > > > > root@stfx:~/experiments/pharo4# ./pharo Pharo.image printVersion > > [version] 4.0 #40535 > > > > root@stfx:~/experiments/pharo4# ls -lah Pharo.image > > -rw-rw-r-- 1 root root 22M Mar 4 14:51 Pharo.image > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval '100 > factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.386s > > user 0m0.184s > > sys 0m0.028s > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval '100 > factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.422s > > user 0m0.192s > > sys 0m0.032s > > > > root@stfx:~/experiments/pharo4# time ./pharo Pharo.image eval '100 > factorial' > > > 93326215443944152681699238856266700490715968264381621468592963895217599993229915608941463976156518286253697920827223758251185210916864000000000000000000000000 > > > > real 0m0.863s > > user 0m0.404s > > sys 0m0.064s > > > > It is the latest (smaller) image, but that should not make much > difference. > > On my CentOS Linux, headless tests are fast as well. > > Granted, my Windows image was super large. 300Megs or something. > I cleaned the Monticello things with flush cache, which removed 41 megs. > > I can't test the CLI version on Windows, as I do get a window opening. > > There are a ton of windows and morphs open in there and I've run a > couple of experiments. > > Now, the Linux image was large too. 200Megs or so. > And the window showed immediately, then it took a second and a half or > so to show the images back. > There is a black background before. > Maybe the system is processing its startUp list, unhibernating forms, > I don't know what's going on but it takes that 1.5 secs. > > The image gets large because of Monticello caches, and TWM (Tiling > Window Manager) keeping a handle on browsers etc over time. So, I > cleaned the caches, closed TWM, gc'ed a time or two and then got the > image back to 90 megs. > > Then the image opens in like 0.5 sec. It is hard to measure as what I > want is the time between issuing the pharo-ui XYZ.image and the fact > that I can actually see the UI. > > How would one do that? > > Sorry for the rant but generally speaking, the Pharo UI seems slower > than the Squeak 4.5 ui. > Or the EToys UI for that matter. > I am using those for little educational programs and they feel faster. > I'd like to know why. Both systems have morphic. Is it because of > aliased fonts? Squeak 4.5 also has that. > > Is there any way to monitor what's going on during image startup? > I modifier the AndreasSystemProfiler in the Squeak code base at Cadence to allow profiling across snapshot; you see both save and startup profiles. That's not showing e.g. disc activity from VM reads, but it does help track down tike spent in the SMalltalk start-up. IIRC this is in te Saueak code base. But if you're interested I can dig it out. The changes are minor; to do with correctly offsetting delays across snapshot. > Phil > > > > > > > > > Sven > > > > > > -- > > Sven Van Caekenberghe > > Proudly supporting Pharo > > http://pharo.org > > http://association.pharo.org > > http://consortium.pharo.org > > > > > > > > > > -- best, Eliot

March 5, 2015

Re: [Pharo-dev] SUnit testing questions - class side methods and methods defined in private protocols

by Max Leske

> On 05 Mar 2015, at 22:20, sheridan(a)sheridan-mahoney.com wrote: > > Hi all, > > in SUnit do we typically test methods found on the class side that are in "instance creation" and "class initialization" protocols? > > Thereâs a lot of personal preference involved there :) If your creation / init methods contain some non-trivial logic (e.g. more than setting values) then I suggest you write tests. Otherwise, using the objects in other tests automatically covers creation and initialization (at least for the trivial cases). > Also, do we test methods found on the instance side but in the protocol of private? > > Yes, absolutely, if they are not simply âprivate getters / settersâ. Black box testing just doesnât cut it in my experience. Private methods can often contain important code that needs to be tested but simply is not inteded for the public to be used. Cheers, Max > Many thanks, > > Sheridan >

March 5, 2015

New Cog VMs available

by Eliot Miranda

Hi All, recent instability seems to have been cured. Please find new VMs at http://www.mirandabanda.org/files/Cog/VM/VM.r3266. For Squeakers, these are close to release candidates for the upcoming 5.0 release. I have one last modification to do with free space handling on start-up in Spur (the current VMs over-allocate free space), and maybe a minor speed-up of the LargeIntegersPlugin; other than that I intend bug fixes only. CogVM binaries as per VMMaker.oscog-eem.1085/r3266 Avoid duplication in fetching active process and scheduler in process-switch machinery. Take advantage of endianness for simpler large integer initialization. Simplify primitiveMakePoint. Include primitiveMethodXray to introspect on Cog state of methods. Refactor positive32BitIntegerFor: and signed32BitIntegerFor: in the realization that these reduce to essentially integerObjectOf: in 64-bit Spur. The idea is to inline if in 64-bit Spur but not if in the 32-bit VMs. Add notOption: processing to allow excluding noInlineSigned32BitIntegerFor: et al. Add hasSixtyFourBitImmediates to no longer assume that wordSize = 8 implies 64-bit immediates. Use positiveMachineIntegerFor: in place of positive32BitIntegerFor: in some plugins. Simplify primitiveDisplayString given isArray:. Simplify two B3DAcceleratorPlugin prims given topRemappableOop and isArray: Slang: Changes to type inferrence/propagation and inlining to better support 64-bit Spur. See change log of VMMaker.oscog in 1067-1085 range for exhaustive details. -- best, Eliot

March 5, 2015

SUnit testing questions - class side methods and methods defined in private protocols

by sheridan＠sheridan-mahoney.com

Hi all, in SUnit do we typically test methods found on the class side that are in "instance creation" and "class initialization" protocols? Also, do we test methods found on the instance side but in the protocol of private? Many thanks, Sheridan

March 5, 2015

Pharo-Launcher - settings change are missed

by Maximiliano Taborda

Hi people, I think I have a problem with pharo-launcher. Today I installed pharo for ubuntu, from ppa, and I want to change some of the settings. But, when I change the settings (and press enter) and then close the launcher and open it again, the settings appear again with defaults values (the changes I made are missed). What I'm doing wrong? Thanks! Regards, Maxi

March 5, 2015

[Pharo4][Slots] Small progress of the day: Example 1st class Globals

by Marcus Denker

Hi, The Slots project contains not only Slots, but in addition âFirst classâ Globals, too, Pharo4 now contains an example: ExampleClassVariable, this class defines the ivar (slot) âstateâ. read ^state write: anObject state := anObject we can create a class defining a class var like that: Object subclass: #TT slots: { } classVariables: { #ClassVar => ExampleClassVariable } category: âPlayGround' Like with slots, the variable reads and write generate code to call the reflective read and write methods: 21 <20> pushConstant: ##ClassVar 22 <D1> send: read 23 <7C> returnTop and the write: 21 <10> pushTemp: 0 22 <69> popIntoTemp: 1 23 <20> pushConstant: ##ClassVar 24 <11> pushTemp: 1 25 <E1> send: write: 26 <87> pop 27 <78> returnSelf Full list of changes: - cleanup: moving all Reflectivity related code from Slot package to Reflectivity package - fix code generation to not use Object>>#writeToGlobal: but instead the idea with the temp as seen above - add example for Class Variable Marcus

March 5, 2015

Re: [Pharo-dev] Research Groups

by Yuriy Tymchuk

Just a small remark: REVEALâs official webpage is: http://reveal.inf.usi.ch <http://reveal.inf.usi.ch/> Thank you. Uko > On 05 Mar 2015, at 11:10, stepharo <stepharo(a)free.fr> wrote: > > give me what I should write exactly :) > > Le 5/3/15 15:49, Serge Stinckwich a Ã©crit : >> Great ! >> >> UMMISCO affiliations are IRD and UPMC. >> Just to have a coherent list, maybe you should put the country in parenthesis. >> >> On Thu, Mar 5, 2015 at 3:42 PM, stepharo <stepharo(a)free.fr> wrote: >>> Hello guys >>> >>> I started to collect all the research groups using Pharo. If you are not >>> part of the group or if your affiliation / web page >>> is not correct let me know >>> >>> http://pharo.org/ResearchGroups >>> >>> Stef >>> >> >> > >

March 5, 2015

Re: [Pharo-dev] Research Groups

by stepharo

give me what I should write exactly :) Le 5/3/15 15:49, Serge Stinckwich a Ã©crit : > Great ! > > UMMISCO affiliations are IRD and UPMC. > Just to have a coherent list, maybe you should put the country in parenthesis. > > On Thu, Mar 5, 2015 at 3:42 PM, stepharo <stepharo(a)free.fr> wrote: >> Hello guys >> >> I started to collect all the research groups using Pharo. If you are not >> part of the group or if your affiliation / web page >> is not correct let me know >> >> http://pharo.org/ResearchGroups >> >> Stef >> > >

March 5, 2015

CFP IWST 2015 - Reminder - Deadline April 10th

by jannik laval

[Please accept our apologies if you receive multiple copies of this call] [Please send to interested colleagues / mailing-lists] ************************************************************************************************************ CALL FOR PAPERS IWST15 â International Workshop on Smalltalk Technologies Brescia, Italy; July 15, 2015 http://www.esug.org/wiki/pier/Conferences/2015/International-Workshop-IWST_… ************************************************************************************************************ ------------------- Goals and scopes ------------------- The goals of the workshop is to create a forum around advances or experience in Smalltalk and to trigger discussions and exchanges of ideas. The topics of your paper can be on all aspect of Smalltalk, theoretical as well as practical. Participants are invited to submit research articles or industrial papers. This year we want to open two different tracks: one research track and one industrial track with less scientific constraints. We expect papers of three kinds: Short position papers describing emerging ideas Long research papers with deeper description of experiments and of research results. Industrial papers with presentation of real and innovative Smalltalk applications; this kind of paper should enlighten why Smalltalk is really appropriate for your application. We will not enforce any length restriction. -------------------- Important Dates -------------------- *Submission deadline: April 10th, 2015* Notification deadline: May 20th, 2015 Workshop : July 15th, 2015 All accepted papers will be published in ACM DL ------------------- Topics ------------------- We welcome contributions on all aspects, theoretical as well as practical, of Smalltalk related topics such as: -Aspect-oriented programming, -Design patterns, -Experience reports, -Frameworks, -Implementation, new dialects or languages implemented in Smalltalk, -Interaction with other languages, -Meta-programming and Meta-modeling, -Tools ------------------- Best Paper Award ------------------- To encourage the submission of high-quality papers, the IWST organizing committee is very proud to announce a Best Paper Award for this edition of IWST. We thank the Lam Research Corporation for its financial contribution which makes it possible for prizes for the three best papers: 1000 USD for first place, 600 USD for second place and 400 USD for third place. The ranking will be decided by the program committee during the review process. The awards will be given during the ESUG conference social event. The Best Paper Award will take place only with a minimum of six submissions. Notice also that to be eligible, a paper must be presented at the workshop by one of the author and that the presenting author must be registered at the ESUG conference. ------------------- Publication ------------------- Both submissions and final papers must be prepared using the ACM SIGPLAN 10 point format. Templates for Word and LaTeX are available at http://www.acm.org/sigs/sigplan/authorInformation.htm. This site also contains links to useful informations on how to write effective submissions. ------------------- Submission ------------------- All submissions must be sent via easychair: https://easychair.org/conferences/?conf=iwst2015 ------------------- Program chairs ------------------- Anne Etien (UniversitÃ© de Lille 1, France) Jannik Laval (Ecole des Mines de Douai, France) -- ~~Jannik Laval~~ Ãcole des Mines de Douai Enseignant-chercheur http://www.jannik-laval.eu http://www.phratch.com http://www.approchealpes.info http://car.mines-douai.fr/

March 5, 2015

[pharo-project/pharo-core]

by GitHub

Branch: refs/tags/40537 Home: https://github.com/pharo-project/pharo-core

March 5, 2015