September 2013 - Pharo-users

Re: [Pharo-users] Benchmark

by Camillo Bruni

Short Answer: ------------- problemSize is used to calibrate your benchmark, so usually you adapt this value for your machine. And yes you are right, the result is not divided by the problemSize. Long Answer: ------------ The typical use case is like this: MathBench >> benchLoopSinus 1 to: self problemSize to: [ :i | i sin ] Now you have two parameters to modifiy: 1. the number of samples you take (aka, how many times you measure the time of #benchLoopSinus) 2. the problem size (aka, how many times you run #sin inside #benchLoopSinus) You increase (1) to get a more stable result: MathBench run: 1 "for debugging". MathBench run: 100 "will take a long time, but results are more accurate" You modify (2) to change the duration of your benchmark, in my silly example above, the method is quite small and the benchmark would finish too quickly. Rule of thumb is to get the run time of your benchmark (here #benchLoopSinus) in the range of 10ms and more. This way you don't have to worry about the timer resolution used to capture the duration of your method. Does this answer your question? On 2013-09-14, at 08:06, Natalia Tymchuk <natalia.tymchuk(a)unikernel.net> wrote: > Hello. > I have a question about problemSize in benchmarking. I got the small times from running benchmarks and that's why I try to use problemSize like in the video http://vimeo.com/68494202. But the benchmark for which I rewrote code and used problemSize has given me very big result. Maybe it looks like there is no division on that problemSize. Am I wrong? What can I do? > Best regards, > Natalia > > 12.09.13 15:20, StÃ©phane Ducasse Ð½Ð°Ð¿Ð¸ÑÐ°Ð²(Ð»Ð°): >> On Sep 11, 2013, at 9:22 PM, Natalia Tymchuk <natalia.moskovchuk(a)unikernel.net> wrote: >> >>> Hello Stephane. >>> Yes, I'm interested in that and it will be an honor for me)). >> Excellent. >> If you have a github account let me know and I will add you as a contributor >> Have a look at >> >> https://github.com/SquareBracketAssociates/PharoForTheEnterprise-english/tr… >> >>> But I never wrote the books and my English is not at its best level. However I think that after a couple reviews my English will be better . >> No stress :) >> My english sucks too and we should not care for now. >> >>> Best regards, >>> Natalia >> > >

Sept. 14, 2013

Re: [Pharo-users] Labeling

by Natalia Tymchuk

Thanks 13.09.13 12:06, Damien Cassou Ð½Ð°Ð¿Ð¸ÑÐ°Ð²(Ð»Ð°): > On Wed, Sep 11, 2013 at 9:04 PM, Natalia Tymchuk > <natalia.moskovchuk(a)unikernel.net> wrote: >> I want to draw a bar diagram for my benchmarks in Graph-ET. Does anybody >> know how to add the labels to x-axis? > > you may want to ask on the moose mailing list or on stackoverflow (tag #pharo) >

Sept. 14, 2013

Re: [Pharo-users] Benchmark

by Natalia Tymchuk

Hello. I have a question about problemSize in benchmarking. I got the small times from running benchmarks and that's why I try to use problemSize like in the video http://vimeo.com/68494202. But the benchmark for which I rewrote code and used problemSize has given me very big result. Maybe it looks like there is no division on that problemSize. Am I wrong? What can I do? Best regards, Natalia 12.09.13 15:20, StÃ©phane Ducasse Ð½Ð°Ð¿Ð¸ÑÐ°Ð²(Ð»Ð°): > On Sep 11, 2013, at 9:22 PM, Natalia Tymchuk <natalia.moskovchuk(a)unikernel.net> wrote: > >> Hello Stephane. >> Yes, I'm interested in that and it will be an honor for me)). > Excellent. > If you have a github account let me know and I will add you as a contributor > Have a look at > > https://github.com/SquareBracketAssociates/PharoForTheEnterprise-english/tr… > >> But I never wrote the books and my English is not at its best level. However I think that after a couple reviews my English will be better . > No stress :) > My english sucks too and we should not care for now. > >> Best regards, >> Natalia >

Sept. 14, 2013

Re: [Pharo-users] FileTree Monticello Repository

by StÃ©phane Ducasse

Thierry I think that what ben want to say is: - provide a default "but I don't want to enforce a specific way of organising the git repositoryâ¦ " Stef PS: I would love to be able to use git for managing Pharo so keep pushing. Now we cannot allocate more resources to this task but last year I request to inria some resources for this task (and it was rejected). > > > Le 13/09/2013 17:04, btc(a)openinworld.com a Ã©crit : >> Goubier Thierry wrote: >>> >>> >>> Le 13/09/2013 04:25, Esteban A. Maringolo a Ã©crit : >>>> Is anybody using FileTree repositories? >>> >>> Which repositories? Dale's github? >>> >>>> If so, I'd like to know the pros and cons of using it. >>> >>> More complex to use than Smalltalkhub/Squeaksource because it is not >>> yet integrated in the default image. Some features may not work, but >>> average MC work under Pharo2.0 works fine with all features of >>> Filetree (including gitfiletree). >>> >>> Easier to use than Smalltalkhub/Squeaksource for authentification. >>> >>>> I want to use to version everything with git, but I don't know how it >>>> keeps git (it is, FileTree files) in sync with MC versions. >>> >>> A filetree repository appears as a MC repository in the same way as >>> another type of MC repository. >>> >>> Filetree will require that you do all your version management in the >>> git repo by hand (i.e. save a version of your package in a filetree >>> repo, then switch to a console to commit it. If you forget to commit >>> and save a new version of a package on the filetree repo, you >>> overwrite the previous version and git will not record it. Filetree by >>> itself will only show you the state of the HEAD of your git repository). >>> >>> If you use gitfiletree (it sits over Filetree), then git commit will >>> be done for you when saving your package, and the repository browser >>> will show you all the history of the repository, exactly as if it was >>> a Smalltalkhub or Squeaksource repository. What will remain to do on >>> the command line are the git push and whatever management you do on >>> the repository (merging, branching, even if you can also merge via MC >>> instead). In short, gitfiletree calls git for you and parse the >>> results :) >>> >>> I'm still wondering about adding more git commands inside the MC GUI >>> to do more from inside Pharo (like automagically pushing), but I don't >>> want to enforce a specific way of organising the git repository... and >>> I'm just trying to get gitfiletree to pass its tests in the pharo3.0 >>> branch of Filetree at the moment. >>> >>> Does this answer your questions? >>> >>> Thierry >> > but I don't want to enforce a specific way of organising the git >> repository >> >> I see this extreme altruism :) around the community, and while it is >> very admirable, I think sometimes its misplaced :). The flip side is >> that some people (like me) aren't really familiar with git don't know >> how we might want git organised, or even what the options are. >> Unfortunately this stalls my motivation to try it out (sorry, weak, I >> know - but I admit it, and I think its common (or do I just think that >> to make myself feel better)). Having a concrete implementation of any >> additional functionality that might make git more transparent "out of >> the box" would be beneficial for its uptake. As well, without a >> concrete example, others who might want to implement something like that >> have to start more from scratch, whereas your concrete example might be >> 90% of what someone wants and then its only 10% that they need to >> customize. The reality of human nature is that its easier to criticize >> to pick out what you don't like, than to create something new. However >> the hidden benefit of this criticism is that it engages people in >> discussion and then you hook them into helping develop what they need. >> So anyway, my short response is, if "you" would benefit from that added >> functionality then "just do it". :) >> cheers -ben >> >> P.S. I admire what I read about progress with Filtetree and GitFileTree >> towards using git. For some segments outside the Pharo community that >> we want to draw in, this is going to be an important consideration. > > Hi Ben, > > I'm sorry that this doesn't convince you to use git with Pharo :) > > So far, I very happy with the functionality gitfiletree has; for a common user like me it has been working very nicely for quite some time now :):) It has the features I need: allow the introduction of pharo development in any pre-existing / mixed language repository; allow me to build in 5 sec a new repo for a package; integrate with github; better long history exploration than anything else under Pharo. > > At the same time, it was a bit of a mock-up, this gitfiletree: just adding some automation to a filetree over git repository, and also add some motivation for someone else to see what can be done and how, so as to eventually reimplement it as a FileSystem / or as a libcgit instead of calling git log and git archive :) > > It's now a bit of a bench for some future directions of FileTree, such as removing the metadata which is provoking conflicts when merging under git, or as a new replacement of changesets (but EPICEA is going in another interesting direction, so I'm not trying anything there). > > And don't worry for my altruism not adding functionalities :) I'm the guy who wrote a complete alternative IDE for pharo 2.0, just because I prefer it to Nautilus :O > > Thierry > -- > Thierry Goubier > CEA list > Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s > 91191 Gif sur Yvette Cedex > France > Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95 >

Sept. 14, 2013

Re: [Pharo-users] FileTree Monticello Repository

by GOUBIER Thierry

Yes, there is a way. If you use a Gofer import script[1] to move a package from say, a squeaksource repository to a git repository with gitfiletree, you will import and commit all the versions one by one inside the git repo, starting with the oldest. You will loose the timestamps, but not the history. Thierry [1] I should give you a link, but a google "Gofer migrate from Squeaksource to Smalltalkhub" should bring you to a few good blog pages with scripts example, something along the lines of Gofer fetch. Gofer push :) ________________________________________ De : Pharo-users [pharo-users-bounces(a)lists.pharo.org] de la part de Esteban A. Maringolo [emaringolo(a)gmail.com] Date d'envoi : vendredi 13 septembre 2013 17:52 Ã : Any question about pharo is welcome Objet : Re: [Pharo-users] FileTree Monticello Repository 2013/9/13 Paul DeBruicker <pdebruic(a)gmail.com> > Also, when beginning to use git, if I make the initial commit of the > package is all the prior editing and version history lost? It happened to me. Maybe there is a way of, in a blank image, load the MC package editions from a regular repo and commit them in the GitFileTree one. It is... to replay every commit in the new repo. Esteban A. Maringolo

Sept. 13, 2013

Re: [Pharo-users] FileTree Monticello Repository

by GOUBIER Thierry

When you change a method and save, git will see one file changed (the method) and put it in the commit (there will also be a few changes in the metadata associated with the method and this will be included in the commit, but this won't matter for gitfiletree). Everything else will stay as before, i.e. owned by the original author (if it is done on git). Additionally, the message in the git log will be the one you wrote for the package in MC when saving. When you start to use git, what you should do is move all your package versions onto the new repository, for example with a Gofer script. It will commit all the versions, starting from the first one, and so recreate the history (but without the merges and resetting timestamps and attributing ownership to you); I don't know enough of git to do a perfect history import, but it could be done. Thierry ________________________________________ De : Pharo-users [pharo-users-bounces(a)lists.pharo.org] de la part de Paul DeBruicker [pdebruic(a)gmail.com] Date d'envoi : vendredi 13 septembre 2013 17:47 Ã : pharo-users(a)lists.pharo.org Objet : Re: [Pharo-users] FileTree Monticello Repository On 09/13/2013 07:53 AM, Goubier Thierry wrote: > It also bypass most of the stored metadata in the package: methods > timestamps are the git commit timestamps, package history is the git > commit history of the repo (pruned to remove all commits not related to > the package, in the case of a repo containing multiple packages). So does this mean that If I make a bug fix in someone's package and then commit it, I become the author of every method of the package? Or can it tell from the commit that its just that one method that changed and all the other methods are from prior commits? Also, when beginning to use git, if I make the initial commit of the package is all the prior editing and version history lost? Thanks Paul

Sept. 13, 2013

Re: [Pharo-users] FileTree Monticello Repository

by Esteban A. Maringolo

I knew about github "paid" private repos. But I use https://bitbucket.org/ instead (it's free and I like it more). I didn't know about the private SmalltalkHub repos though. Thank you! Esteban A. Maringolo 2013/9/13 Paul DeBruicker <pdebruic(a)gmail.com>: > On 09/13/2013 08:50 AM, Esteban A. Maringolo wrote: >> >> But there are no "private repositories" in GitHub nor SmalltalkHub and >> not everybody does open source software nor have servers on their own >> premises. > > github private repos: https://github.com/plans > > smalltalkhub private repos: change the license to Private > > > > >

Sept. 13, 2013

Re: [Pharo-users] FileTree Monticello Repository

by Paul DeBruicker

On 09/13/2013 08:50 AM, Esteban A. Maringolo wrote: > > But there are no "private repositories" in GitHub nor SmalltalkHub and > not everybody does open source software nor have servers on their own > premises. github private repos: https://github.com/plans smalltalkhub private repos: change the license to Private

Sept. 13, 2013

Re: [Pharo-users] FileTree Monticello Repository

by Esteban A. Maringolo

2013/9/13 Paul DeBruicker <pdebruic(a)gmail.com> > Also, when beginning to use git, if I make the initial commit of the > package is all the prior editing and version history lost? It happened to me. Maybe there is a way of, in a blank image, load the MC package editions from a regular repo and commit them in the GitFileTree one. It is... to replay every commit in the new repo. Esteban A. Maringolo

Sept. 13, 2013

Re: [Pharo-users] FileTree Monticello Repository

by Esteban A. Maringolo

As a I side note: I know Dale uses FileTree repos to publish his projects to GitHub and build it with Travis CI. For many people is okay with using SmalltalkHub to publish their MC Packages, and that's fine and a real progress compared with SqueakSource. But there are no "private repositories" in GitHub nor SmalltalkHub and not everybody does open source software nor have servers on their own premises. I'm already using BitBucket for my non-smalltalk code, and and I already pushed my MC FileTree based packages to it. I like how it looks. Regards! Esteban A. Maringolo 2013/9/13 Esteban A. Maringolo <emaringolo(a)gmail.com> > Git it is not just a benefit for outsiders, it is for Smalltalkers too. > > In the company I work for (we use Dolphin), we moved away from its VCS > (STS, an ENVY like repo) a few months ago to file based packages (Dolphin > supports this format too to be used by "regular" VCS, in a one dir per > package, one file per class basis). > > Now we version everything with git. I'm talking about a huge system > (thousands of classes, and LOTS of SLOC). > > After moving to git we were able to leverage existing tools for continuous > integration tools with simple build scripts. > We already made two production releases (another one coming at the end of > the month), and the branching capabilities is really useful to apply > hotfixes to previous releases. We're all happy with its outcome. > We're building runtime (deploy) versions, and development images > continuously, in less than 5 minutes. > > I'm no git expert, and no MC "proficient user" either. But I'm trying to > replicate what we have in my personal projects. > > That's why FileTree caught my attention way back when Dale Henrichs > started coding it. > > I still have to get my head around the packaging system of Pharo, I don't > understand it, or actually I'm not sure if I do or not, because there is > PackageInfo, Monticello, RPackage. Metacello is a different beast I'm > getting to know because of its resemblance with ENVY's ConfigurationMaps. > > > So... bottom line, I'll give gitfiletree repos a try, and see how it goes. > > Regards! > > > > > > > > Esteban A. Maringolo > > > 2013/9/13 <btc(a)openinworld.com> > > Goubier Thierry wrote: >> >>> >>> >>> Le 13/09/2013 04:25, Esteban A. Maringolo a Ã©crit : >>> >>>> Is anybody using FileTree repositories? >>>> >>> >>> Which repositories? Dale's github? >>> >>> If so, I'd like to know the pros and cons of using it. >>>> >>> >>> More complex to use than Smalltalkhub/Squeaksource because it is not yet >>> integrated in the default image. Some features may not work, but average MC >>> work under Pharo2.0 works fine with all features of Filetree (including >>> gitfiletree). >>> >>> Easier to use than Smalltalkhub/Squeaksource for authentification. >>> >>> I want to use to version everything with git, but I don't know how it >>>> keeps git (it is, FileTree files) in sync with MC versions. >>>> >>> >>> A filetree repository appears as a MC repository in the same way as >>> another type of MC repository. >>> >>> Filetree will require that you do all your version management in the git >>> repo by hand (i.e. save a version of your package in a filetree repo, then >>> switch to a console to commit it. If you forget to commit and save a new >>> version of a package on the filetree repo, you overwrite the previous >>> version and git will not record it. Filetree by itself will only show you >>> the state of the HEAD of your git repository). >>> >>> If you use gitfiletree (it sits over Filetree), then git commit will be >>> done for you when saving your package, and the repository browser will show >>> you all the history of the repository, exactly as if it was a Smalltalkhub >>> or Squeaksource repository. What will remain to do on the command line are >>> the git push and whatever management you do on the repository (merging, >>> branching, even if you can also merge via MC instead). In short, >>> gitfiletree calls git for you and parse the results :) >>> >>> I'm still wondering about adding more git commands inside the MC GUI to >>> do more from inside Pharo (like automagically pushing), but I don't want to >>> enforce a specific way of organising the git repository... and I'm just >>> trying to get gitfiletree to pass its tests in the pharo3.0 branch of >>> Filetree at the moment. >>> >>> Does this answer your questions? >>> >>> Thierry >>> >> > but I don't want to enforce a specific way of organising the git >> repository >> >> I see this extreme altruism :) around the community, and while it is very >> admirable, I think sometimes its misplaced :). The flip side is that some >> people (like me) aren't really familiar with git don't know how we might >> want git organised, or even what the options are. Unfortunately this >> stalls my motivation to try it out (sorry, weak, I know - but I admit it, >> and I think its common (or do I just think that to make myself feel >> better)). Having a concrete implementation of any additional functionality >> that might make git more transparent "out of the box" would be beneficial >> for its uptake. As well, without a concrete example, others who might want >> to implement something like that have to start more from scratch, whereas >> your concrete example might be 90% of what someone wants and then its only >> 10% that they need to customize. The reality of human nature is that its >> easier to criticize to pick out what you don't like, than to create >> something new. However the hidden benefit of this criticism is that it >> engages people in discussion and then you hook them into helping develop >> what they need. >> So anyway, my short response is, if "you" would benefit from that added >> functionality then "just do it". :) >> cheers -ben >> >> P.S. I admire what I read about progress with Filtetree and GitFileTree >> towards using git. For some segments outside the Pharo community that we >> want to draw in, this is going to be an important consideration. >> >> >> >> >> >

Sept. 13, 2013