- Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Filetree and Git and Github

by Ben Coman

Thierry Goubier wrote: > Le 20/09/2014 14:24, phil(a)highoctane.be a Ã©crit : >> Now history of methods is indeed super nice. > > It is :) > > Thierry > Is that just from the command line? I guess with libgit integration we might get this with in-Image "versions" ? cheers -ben

Sept. 20, 2014

Pragmas - compile time versus run time

by Ben Coman

In consideration of Issue 11880, I wanted to better understand the concept of pragmas and how they are used in Pharo. "Pragmas" seem to be used in other languages to provide "compile time" directives, but Pharo seems search for pragmas at runtime, for example to do things like build menus. Mostly I'd consider menus fairly static (with any dynamism statically defined). In Issue 11880 adding a few caches brings menu generation from 80ms down to 5ms - but then there is the complexity of resetting them when the code that defines them changes. So I wonder... can/do pragmas cause actions at compile time, like setting a class variable to nil when a method is compiled? That seems a good way to reset a cache whose builder uses that method, but maybe there are problems I don't see? https://pharo.fogbugz.com/default.asp?11880 cheers -ben

Sept. 20, 2014

Re: [Pharo-dev] Filetree and Git and Github

by Thierry Goubier

Le 20/09/2014 14:24, phil(a)highoctane.be a Ã©crit : > > One thing for me is that I do have other stuff that Pharo code. Me as well. > So filetree keeps all together in commits/pushes. You synchronize the commits over the Pharo packages and the 'other' code and resources? > Now history of methods is indeed super nice. It is :) Thierry

Sept. 20, 2014

AbstractNautilusUI "error, should not happen" from LayoutEmptyScope(LayoutAbstractScope)>>rebase:to:

by Ben Coman

In build 40239, if I try to add an instance variable (e.g. 'xx') to AbstractNautilusUI, I get "Error: should not happen" from LayoutEmptyScope(LayoutAbstractScope)>>rebase:to: A tried a few other classes and they are okay. I've tracked this down to occur since build 40147... 13697 Yet another cleanup of MessageBrowser API https://pharo.fogbugz.com/f/cases/13697 13698 Small cleanup of Model and the UI package https://pharo.fogbugz.com/f/cases/13698 13696 Nautilus Refactoring 1: Classlist widget https://pharo.fogbugz.com/f/cases/13696 Is this a known problem? I can't identify a ticket on the tracker. cheers -ben

Sept. 20, 2014

Re: [Pharo-dev] Visualising Network Latency

by Tudor Girba

Nice showcase for agile visualization :) Doru On Fri, Sep 19, 2014 at 3:23 PM, Leonel Merino <merino(a)iam.unibe.ch> wrote: > Hi All, > > I have been working on a visualisation of latency on a network. I found an > interesting dataset with latency information of 140 cities around the world. > In the visualisation I show different metrics regarding latency such as > from each city which other has the best/worst latency; cities that > geographically are close but have high latency; far cities that have low > latency; and average latency for each city. > > When visualising the worst latency for each city I found that, for this > network (at least), Valencia and in a minor extent Nairobi represent the > worst endpoint for connections (in terms of latency). Also in the > visualisation of far cities with low latency, Newmarket in Canada and > Toledo in Spain concentrate the best endpoint for connections coming from > Europe and North-America respectively. > > > Here a couple of screenshots: > > > > > > > > > Here there is a screencast: > http://www.youtube.com/watch?v=gDQnE-5yeWI > > > > > If you want to load the code: > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= > Gofer new smalltalkhubUser: 'merino' > project: 'NetworkLatency'; > package: 'ConfigurationOfNetworkLatency'; > load. > (ConfigurationOfNetworkLatency project version: '0.1') load. > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-= > > To execute the visualisation: > NLVisualise new open > > > > > I would be glad to have your comments and suggestions. > > Best regards, > > Leonel Merino > Research Assistant > Software Composition Group > Institute of Computer Science and Applied Mathematics > University of Bern > > -- www.tudorgirba.com "Every thing has its own flow"

Sept. 20, 2014

Re: [Pharo-dev] Visualising Network Latency

by Alexandre Bergel

@Markus: I got the same error Alexandre > Le 19-09-2014 Ã 9:06, Markus Fritsche <mfritsche(a)reauktion.de> a Ã©crit : > > >> On 19.09.2014 15:23, Leonel Merino wrote: >> If you want to load the code: >> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= >> Gofer new smalltalkhubUser: 'merino' >> project: 'NetworkLatency'; >> package: 'ConfigurationOfNetworkLatency'; >> load. >> (ConfigurationOfNetworkLatency project version: '0.1') load. >> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-= >> >> To execute the visualisation: >> NLVisualise new open > What are the prerequisites? I tried to load it in Pharo-3.0 (with latest > updates) and it was missing Trachel. So I updated Roassal2 to the latest > version and now I get a MNU: > MessageNotUnderstood: RTMapBuilder>>mapPing: > > Kind regards, > Markus >

Sept. 20, 2014

Re: [Pharo-dev] Visualising Network Latency

by Alexandre Bergel

Let me know. I will be happy to try again Alexandre > Le 19-09-2014 Ã 8:10, Leonel Merino <merino(a)iam.unibe.ch> a Ã©crit : > > Thanks Alex ! > Sure, Iâll do it. > > âLeonel > >> On 19 Sep 2014, at 16:30, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: >> >> This is excellent Leonel! >> Iâve tried, but I could not make the visualization work. Apparently you are using a rather old version of Roassal2. I suggest you to update it to the last version of Roassal2. >> >> Cheers, >> Alexandre >> >> >>> On Sep 19, 2014, at 6:48 AM, Leonel Merino <merino(a)iam.unibe.ch> wrote: >>> >>> Hi Max, >>> >>> Thanks ! >>> >>> The input is a CSV file which contains a grid with city names and latency. >>> Surely you can zoom-in at city level. However, the map only contains information of the cities that come in the dataset (is not Bern). >>> >>> Best regards, >>> Leonel >>> >>>> On 19 Sep 2014, at 15:30, Max Leske <maxleske(a)gmail.com> wrote: >>>> >>>> Very cool. >>>> >>>> - Do you parse data from a file? If so, what format does that file need to be in? >>>> - Is the map zoomable to city level (to see connections between Bern and ZÃ¼rich for instance)? >>>> >>>> Cheers, >>>> Max >>>> >>>>> On 19.09.2014, at 15:23, Leonel Merino <merino(a)iam.unibe.ch> wrote: >>>>> >>>>> Hi All, >>>>> >>>>> I have been working on a visualisation of latency on a network. I found an interesting dataset with latency information of 140 cities around the world. >>>>> In the visualisation I show different metrics regarding latency such as from each city which other has the best/worst latency; cities that geographically are close but have high latency; far cities that have low latency; and average latency for each city. >>>>> >>>>> When visualising the worst latency for each city I found that, for this network (at least), Valencia and in a minor extent Nairobi represent the worst endpoint for connections (in terms of latency). Also in the visualisation of far cities with low latency, Newmarket in Canada and Toledo in Spain concentrate the best endpoint for connections coming from Europe and North-America respectively. >>>>> >>>>> >>>>> Here a couple of screenshots: >>>>> >>>>> <PastedGraphic-4.pdf> >>>>> >>>>> <PastedGraphic-5.pdf> >>>>> >>>>> <PastedGraphic-6.pdf> >>>>> >>>>> >>>>> Here there is a screencast: >>>>> http://www.youtube.com/watch?v=gDQnE-5yeWI >>>>> >>>>> >>>>> >>>>> >>>>> If you want to load the code: >>>>> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= >>>>> Gofer new smalltalkhubUser: 'merino' >>>>> project: 'NetworkLatency'; >>>>> package: 'ConfigurationOfNetworkLatency'; >>>>> load. >>>>> (ConfigurationOfNetworkLatency project version: '0.1') load. >>>>> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-= >>>>> >>>>> To execute the visualisation: >>>>> NLVisualise new open >>>>> >>>>> >>>>> >>>>> >>>>> I would be glad to have your comments and suggestions. >>>>> >>>>> Best regards, >>>>> >>>>> Leonel Merino >>>>> Research Assistant >>>>> Software Composition Group >>>>> Institute of Computer Science and Applied Mathematics >>>>> University of Bern >> >> -- >> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >> Alexandre Bergel http://www.bergel.eu >> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > >

Sept. 20, 2014

Re: [Pharo-dev] Implementing lazy loading for expandable collections

by Alexandre Bergel

Hi Max! Thanks! This is something I wanted to push to Pharo for a long time. A simpler approach, would be to initialize the internal array to #() when initialized per default. This will provide pretty much the same benefits, and will not break the existing tools. Does this make sense to you? Alexandre > Le 19-09-2014 Ã 2:35, Max Leske <maxleske(a)gmail.com> a Ã©crit : > > Iâve written an implementation of lazily initialized expandable collections (for OrderedCollection and subclasses only for now), inspired by Alexandreâs talk at ESUG (http://www.youtube.com/watch?v=x0YJ2dsZdKg&list=UUO-vBhaKVZf0al-ISMMPvRw) The implementation is pretty much straight forward but there are a couple of culprits I want to point out in case anybody else wants to do this: > > - my implementation requires an extra instance variable in OrderedCollection to store the requested size. This extra instance variable will break Monticello (and possibly other tools) because Monticello uses OrderedCollections to load code and the particular way it uses them makes it impossible to change the number of instance variables on OrderedCollection. Also, all .mcz files written after the change will not be loadable by images without the new instance variable (same reason). > - the above means that you have to modify the code in the image manually and save a new âbase image" > - read only messages should obviously not initialize the array. The âfirstIndexâ and âlastIndexâ variables are quite handy for that. I initialize those variables to 1 and 0 respectively which makes most things work already (e.g. #size, #isEmpty etc.). > - when trying to implement the same for HashedCollection I couldnât. The image didnât just stop working, there was no output at all on the console, not even when using a debug VM. The problem seems to be MethodDictionary, at least thatâs the only subclass of Dictionary for which adding an instance variable doesnât work. > > Iâll share some memory monitoring as soon as I have something significant (only just rolled it out). > > > Big thanks to Alex for his talk and the cool work he and his students did! > > Cheers, > Max

Sept. 20, 2014

Re: [Pharo-dev] Filetree and Git and Github

by phil＠highoctane.be

Le 20 sept. 2014 13:47, "Thierry Goubier" <thierry.goubier(a)gmail.com> a Ã©crit : > > Le 20/09/2014 12:19, Martin McClure a Ã©crit : > >> On 09/19/2014 02:32 AM, kilon alios wrote: >> >>> I have been using not gitfiletree but filetree alone for the past week >>> for my project Ephestos, because I am ok commiting to git via terminal/ >>> >>> https://github.com/kilon/phEphestos >>> >>> I have to congratulate people behind filetree it has been a smooth ride >>> so far.I had no problems with it at all and I am using git in terminal >>> for doing all the git commits and push etc. >> >> >> I've also been using filetree and manually using git (and github) for >> the past two years. It works really well for me. > > > I have done that for a year too, with Pharo 1.4. Then I wrote GitFileTree :) I've used that forever since, on 2.0, 3.0 and 4.0. > > >>> filetree also has been very succesful at some major refactoring I have >>> done to Ephestos, I had to change the name of an entire class and I had >>> to change methods to other classes . Filetree did an excellent job not >>> just with the export to st file but actually git was able to see which >>> files were deleted and which were renamed. >>> >>> That meant I did not have to do any git rm . >>> >>> However I do need to git add every time I add a new method or filetree >>> creates a new file. Usually that is not a problem because I can see >>> what files are not tracked by git with git status and copy paste their >>> paths to git add. >> >> >> After I save in Monticello to my Filetree local filesystem, I just do >> >> git add -A >> >> then I'm ready to commit. The -A (different from -a) picks up all >> changed, added, and removed files automatically. > > > I had a few issues with doing git commands by hand (forgot do it a few times, had conflicts with removed method reapparing with squeak) so this is why I automatized it. > > Then I got a few benefits, like having access to any version of a package in a git repository without checkouts, and even access to all the versions of a method (even up to one or two years old). And a reminder in the GUI that I need to "push". > > >>> So I will say overall my commits have been a process of saving in >>> monticello , git add, git commit and git push which exactly the same >>> process I have used for when I was and still coding in python. >> >> >> That's my workflow as well. I also do work towards a particular bugfix >> or feature out on a Git branch, then merge that branch into the master >> whenever I'm done with that particular issue. Basically, normal Git >> workflow, which makes tracking things very easy. > > > If you often merge, consider using the GitFileTree-MergeDriver; installation is a bit complex, but it may really help if you're merging by reducing conflicts a lot. > > I do need to finish writing that part in the PharoForTheEnterprise book ;) > One thing for me is that I do have other stuff that Pharo code. So filetree keeps all together in commits/pushes. Now history of methods is indeed super nice. Phil > Thierry > > >>> So thank you for this tool , I got fed up with the slow downs of >>> Smalltalkhub and all its problems and it feels great to be back at using >>> github :) >>> >>> I will keep testing for a couple a months and then completely move to >>> github . The only thing I noticed is that sometimes monticello slows >>> down but that may well be a problem with Pharo 4. >> >> >> I haven't noticed this, but I'm still on Pharo 3 for most of my project >> work. >> >> Regards, >> >> -Martin >> >> >> >> > >

Sept. 20, 2014

Re: [Pharo-dev] Filetree and Git and Github

by Thierry Goubier

Le 20/09/2014 12:19, Martin McClure a Ã©crit : > On 09/19/2014 02:32 AM, kilon alios wrote: > >> I have been using not gitfiletree but filetree alone for the past week >> for my project Ephestos, because I am ok commiting to git via terminal/ >> >> https://github.com/kilon/phEphestos >> >> I have to congratulate people behind filetree it has been a smooth ride >> so far.I had no problems with it at all and I am using git in terminal >> for doing all the git commits and push etc. > > I've also been using filetree and manually using git (and github) for > the past two years. It works really well for me. I have done that for a year too, with Pharo 1.4. Then I wrote GitFileTree :) I've used that forever since, on 2.0, 3.0 and 4.0. >> filetree also has been very succesful at some major refactoring I have >> done to Ephestos, I had to change the name of an entire class and I had >> to change methods to other classes . Filetree did an excellent job not >> just with the export to st file but actually git was able to see which >> files were deleted and which were renamed. >> >> That meant I did not have to do any git rm . >> >> However I do need to git add every time I add a new method or filetree >> creates a new file. Usually that is not a problem because I can see >> what files are not tracked by git with git status and copy paste their >> paths to git add. > > After I save in Monticello to my Filetree local filesystem, I just do > > git add -A > > then I'm ready to commit. The -A (different from -a) picks up all > changed, added, and removed files automatically. I had a few issues with doing git commands by hand (forgot do it a few times, had conflicts with removed method reapparing with squeak) so this is why I automatized it. Then I got a few benefits, like having access to any version of a package in a git repository without checkouts, and even access to all the versions of a method (even up to one or two years old). And a reminder in the GUI that I need to "push". >> So I will say overall my commits have been a process of saving in >> monticello , git add, git commit and git push which exactly the same >> process I have used for when I was and still coding in python. > > That's my workflow as well. I also do work towards a particular bugfix > or feature out on a Git branch, then merge that branch into the master > whenever I'm done with that particular issue. Basically, normal Git > workflow, which makes tracking things very easy. If you often merge, consider using the GitFileTree-MergeDriver; installation is a bit complex, but it may really help if you're merging by reducing conflicts a lot. I do need to finish writing that part in the PharoForTheEnterprise book ;) Thierry >> So thank you for this tool , I got fed up with the slow downs of >> Smalltalkhub and all its problems and it feels great to be back at using >> github :) >> >> I will keep testing for a couple a months and then completely move to >> github . The only thing I noticed is that sometimes monticello slows >> down but that may well be a problem with Pharo 4. > > I haven't noticed this, but I'm still on Pharo 3 for most of my project > work. > > Regards, > > -Martin > > > >

Sept. 20, 2014