November 2013 - Pharo-users

Re: [Pharo-users] Pharo VM + Travis Validation

by StÃ©phane Ducasse

Thanks Camillo! Now pay attention to your PhD, please. Stef On Nov 9, 2013, at 10:25 PM, Camillo Bruni <camillobruni(a)gmail.com> wrote: > I finished the new setup for the Pharo VM github repository [1] > - renamed codegen-scripts -> scripts > - separate setup scripts for linux (more to come later) > - added scripts/build.sh which builds the VM automatically under linux > - generating version info > - dowloading fresh 2.0 image > - installing VMMaker from scratch from the repository > - added scripts/test.sh which downloads a test image and runs it with the freshly built VM > - complete travis validation [2] > > Missing Goodies: > - OSX travis configuration > - OSX setup scripts (based on homebrew + XCode installation checks) > > I would like to follow a pull-request based contribution, now with the > travis validation in place this is very comfortable! This way it should be > much easier for non-core people to contribute something simple. > > Note that with travis you get free validation for your own fork / branch [3]! > > [1] https://github.com/pharo-project/pharo-vm/ > [2] https://travis-ci.org/pharo-project/pharo-vm/ > [3] https://travis-ci.org/camillobruni/pharo-vm/

Nov. 10, 2013

Re: [Pharo-users] Current 3.0 download is broken

by StÃ©phane Ducasse

Camillo I read the bug entry and your explanation. Now could not we find a solution that is not based on postscript? Why? because I'm thinking that soon (or later) we will bootstrap from sources and I would love to have as much as behavior inside the initialize so that we do not forget a little message and get the system shaky. Stef >> On 09 Nov 2013, at 15:37, Johan Fabry <jfabry(a)dcc.uchile.cl> wrote: >>> For double completeness ;-) The one-click download Pharo3.0-mac.zip has the same problem. The included source file is in the wrong directory. It should be moved out of Contents/Resources and moved into Contents/MacOS. Then the one-click also runs. >> >> That seems to be a regression from how it once was, so it should be fixed ASAP. > > It was fixed and actually extended to take more source locations into account. > But somehow FileLocator is not properly initialized and thus the additional locations > are not properly resolved. > >

Nov. 10, 2013

Re: [Pharo-users] Repository of petit parser

by Tudor Girba

Hi, I documented Snapshotcello here: http://www.tudorgirba.com/blog/snapshotcello-take-a-snapshot-when-you-re-re… http://www.tudorgirba.com/blog/snappier-snapshotcello In short, the reason for Snapshotcello is to deal with deeply nested configurations, like we have in Moose, in a reliable way. So, the goal is to be able to develop on the latest package version at all time, and when ready to take a reloadable snapshot. For now, this implies flattening everything. The actual version strings are not written in the version method because of the 256 literals limitation. So, we are writing all version strings in an array, and creating a Metacello version out of them. In any case, if you want to create a proper Metacello version, all you have to do is: - load the snapshot version - then create a version based on the corresponding baseline (usually the last baseline) - publish the configuration Doru On Sun, Nov 10, 2013 at 10:26 AM, Norbert Hartl <norbert(a)hartl.name> wrote: > > > Am 09.11.2013 um 22:04 schrieb Tudor Girba <tudor(a)tudorgirba.com>: > > There is a very easy way. Create a configuration version that explicitly > specify the package versions without Snapshotcello. > > There is still version 1.6 that is made that way. I tried using it with > Versionner but that doesn't work. I load 1.6 but the next time I create a > new version it assumes I have 1.8-snapshot loaded which I cannot load > afterwards because I refer only to the core but the group isn't there. So > even if I would add a manual version I assume it will break the next time > you guys add one. > Btw. is the metacello mangeling part in Snapshotocello und Versionner not > something where forces should be joined? Versionner generates versions that > have the individual package versions written in the method. > And I have no glue why snapshotocello is not writing the versions in the > version method. My gut feeling is that metacello is complex enough and > putting another layer of dynamism in the mix that tries to be clever is > probably not the best choice. > > Norbert > > Doru > > > On Sat, Nov 9, 2013 at 3:38 PM, Norbert Hartl <norbert(a)hartl.name> wrote: > >> >> >> Am 09.11.2013 um 12:21 schrieb StÃ©phane Ducasse < >> stephane.ducasse(a)inria.fr>: >> >> Indeed. >> I see >> and I did not got the time to fix the issue I saw when applying on >> private projects. >> >> Agreed. But, the ConfigurationOfPetitParser is up to date. It has groups >> to distinguish between between core and gui. >> >> The only problem that Norbert raised is that when using Snapshotcello, >> the groups are lost due to flattening of the entire configuration. >> >> Which makes it unusable for me. Or at least I didn't figure out a way how >> to circumvent this behaviour. >> >> Norbert >> >> Doru >> >> >> On Sat, Nov 9, 2013 at 2:12 AM, StÃ©phane Ducasse < >> stephane.ducasse(a)inria.fr> wrote: >> >>> Guys >>> >>> even if one day we may integrate petitparser (so far I do not see why - >>> but do not comment). We need an up to date configuration >>> because we want everything with configuration because we want to >>> bootstrap 4.0. >>> >>> Stef >>> >> >> >> >> -- >> www.tudorgirba.com >> >> "Every thing has its own flow" >> >> >> > > > -- > www.tudorgirba.com > > "Every thing has its own flow" > > -- www.tudorgirba.com "Every thing has its own flow"

Nov. 10, 2013

Re: [Pharo-users] Repository of petit parser

by Norbert Hartl

> Am 09.11.2013 um 22:04 schrieb Tudor Girba <tudor(a)tudorgirba.com>: > > There is a very easy way. Create a configuration version that explicitly specify the package versions without Snapshotcello. > There is still version 1.6 that is made that way. I tried using it with Versionner but that doesn't work. I load 1.6 but the next time I create a new version it assumes I have 1.8-snapshot loaded which I cannot load afterwards because I refer only to the core but the group isn't there. So even if I would add a manual version I assume it will break the next time you guys add one. Btw. is the metacello mangeling part in Snapshotocello und Versionner not something where forces should be joined? Versionner generates versions that have the individual package versions written in the method. And I have no glue why snapshotocello is not writing the versions in the version method. My gut feeling is that metacello is complex enough and putting another layer of dynamism in the mix that tries to be clever is probably not the best choice. Norbert > Doru > > >> On Sat, Nov 9, 2013 at 3:38 PM, Norbert Hartl <norbert(a)hartl.name> wrote: >> >> >>> Am 09.11.2013 um 12:21 schrieb StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: >>> >> >>> Indeed. >>> I see >>> and I did not got the time to fix the issue I saw when applying on private projects. >>> >>>> Agreed. But, the ConfigurationOfPetitParser is up to date. It has groups to distinguish between between core and gui. >>>> >>>> The only problem that Norbert raised is that when using Snapshotcello, the groups are lost due to flattening of the entire configuration. >>>> >> Which makes it unusable for me. Or at least I didn't figure out a way how to circumvent this behaviour. >> >> Norbert >> >>>> Doru >>>> >>>> >>>>> On Sat, Nov 9, 2013 at 2:12 AM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >>>>> Guys >>>>> >>>>> even if one day we may integrate petitparser (so far I do not see why - but do not comment). We need an up to date configuration >>>>> because we want everything with configuration because we want to bootstrap 4.0. >>>>> >>>>> Stef >>>> >>>> >>>> >>>> -- >>>> www.tudorgirba.com >>>> >>>> "Every thing has its own flow" >>> > > > > -- > www.tudorgirba.com > > "Every thing has its own flow"

Nov. 10, 2013

Re: [Pharo-users] Adjusting the size of a window using Spec

by Benjamin

This was indeed my goal :) Having kind of solid documentation for 3.0 release. Spec was in 2.0 because we already had some tools already re implemented in Spec but also that people can know it exists before it becomes âmainstream" Ben On 10 Nov 2013, at 04:45, btc(a)openinworld.com wrote: > Benjamin wrote: >> >> Shame on you dude :P >> >> It points out the huge lack of Spec documentation, so in fact, shame on me :P >> >> Ben >> > > I'm not sure how much Pharo 2.0 focused on the move to Spec for UI (I've ended up jumping from 1.4 straight to 3.0alpha) but regarding the professional presentation/reception of Pharo I feel some solid Spec documentation will be critical to coincide with the 3.0 release (or very shortly thereafter). Even though there has been a great amount of other work under-the-hood, Spec is the front-line system that end-user-developers will be working with. I'm sure that already was your aim. This is just a supporting note for that action. > > cheers -ben > >> On 09 Nov 2013, at 20:34, Sean P. DeNigris <sean(a)clipperadams.com> wrote: >> >> >>> Benjamin Van Ryseghem-2 wrote >>> >>>> You just need to override the method #initialExtent >>>> >>> Cool! I didn't know that one. I've been writing ugly #open methods which >>> call #extent:... >>> >>> >>> >>> ----- >>> Cheers, >>> Sean >>> -- >>> View this message in context: http://forum.world.st/Adjusting-the-size-of-a-window-using-Spec-tp4719478p4… >>> Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. >>> >>> >> >> >> >

Nov. 10, 2013

Re: [Pharo-users] Adjusting the size of a window using Spec

by btc＠openinworld.com

Nov. 10, 2013

Re: [Pharo-users] Spec - Nested Layouts

by Benjamin

Cool :) As soon as the last piece is integrated I will send a mail to announce it with some explanations, examples, and how to migrate :) Ben On 09 Nov 2013, at 13:41, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > > On Nov 7, 2013, at 10:07 PM, Benjamin <Benjamin.VanRyseghem.Pharo(a)gmail.com> wrote: > > integrated >> - 11920 (the valueHolder one) >> - 12054 (TreeNode and TreeColumn) >> - There is one also about SpecMenu, I do not remember the number > > > https://pharo.fogbugz.com/f/cases/12051/Add-MenuModels > are on work needed but this is unclear to me. I will check and probably integrate it and please let us know. > > https://pharo.fogbugz.com/f/cases/12050 got integrated > > Stef >

Nov. 9, 2013

Re: [Pharo-users] Spec - Tutorial Series

by Benjamin

On 09 Nov 2013, at 09:53, btc(a)openinworld.com wrote: > Bahman Movaqar wrote: >> Hi all, >> >> A new episode is availabe: >> Spec - Part II: The Layout (http://www.bahmanm.com/blogs/spec-part-2-layout) >> >> I'd appreciate any comments/hints. TIA, >> >> > Some minor nit-picks... > > > "please do so as we are going to iterate over what was done there." > I don't understand what you mean by 'iterate over'? Do you me 'build on'? > > > "let's first settle for a target design. For the rest of this episode, we will try to achieve the following design. " > I get to the bottom and the last snapshot it doesn't look the same as the one at the top. At the top you should mention that your target design is achieved over several posts (if that is the case). > > > "Note lines 9, 10, 17 and 18." > "Note new lines 9, 10, 17 and 18." > > * At the top you should specify what version of Pharo the tutorial relates to. Information on the web hangs around for a very long time and someone could get trapped later trying your tutorial on a later version of Pharo. > > * Regarding #defaultSpec, I happen to like what is used in PharoLauncher > specLayout > <spec: #default> > ^ SpecLayout composed â¦etc The pragma should be preferred :) Actually the check is done like this: - take the first method found with defaultSpec (if multiple, te first found is used) - if none, and there is only one method with the pragma spec, use it - otherwise look for defaultSpec (this should be deprecated soon) - DNU the defaultSpec is a relic of the first version of Spec :) > > Overall it a nice next step introducing just a bit more functionality. It reads very well. I look forward to the next one. > > cheers -ben > > P.S. Some questions for Spec gurus... > > * looking at #defaultSpec from this episode, if the pattern is always... > mainColumn newRow: [ :aRow | aRow doStuff; otherStuff... ] > what is gained over a pattern like... > mainColumn newRow doStuff ; otherStuff. As I understood your question, you answered it :) I use block here so one can easily have a pointer to the row to do some magic here :) It also represents the columns/rows hierarchy > * is #defaultSpec always going to return a SpecLayout? In which case #defaultSpec feels like is should be #defaultSpecLayout. Even if it returns a SpecLayout most of the time, it could as well return an array with a spec structure. Actually, when you return a SpecLayout, it is converted into an array before wing interpreted :) (try to inspect one of your favorite SpecLayout and look at myLayout asArray ^^) > cheers -ben >

Nov. 9, 2013

Re: [Pharo-users] Adjusting the size of a window using Spec

by Benjamin

Shame on you dude :P It points out the huge lack of Spec documentation, so in fact, shame on me :P Ben On 09 Nov 2013, at 20:34, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > Benjamin Van Ryseghem-2 wrote >> You just need to override the method #initialExtent >>> > > Cool! I didn't know that one. I've been writing ugly #open methods which > call #extent:... > > > > ----- > Cheers, > Sean > -- > View this message in context: http://forum.world.st/Adjusting-the-size-of-a-window-using-Spec-tp4719478p4… > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. >

Nov. 9, 2013

Re: [Pharo-users] Spec setting color (was Re: Spec - Tutorial Series)

by Benjamin

I think itâs currently missing. Not that much to implement though, but still not here :P Ben On 09 Nov 2013, at 18:26, btc(a)openinworld.com wrote: > Bahman Movaqar wrote: >> Hi all, >> >> A new episode is availabe: >> Spec - Part II: The Layout (http://www.bahmanm.com/blogs/spec-part-2-layout) >> >> I'd appreciate any comments/hints. TIA, >> >> > Question to Spec gevelopers... > > Referring to Bahman's example, how do you set the color of: > > * the 'Greet Me' button - I tried adding "buttonGreet color: Color red." to both #initializeWidgets and #initializePresenter, > with no error, but also no effect. > > * dialog background - I don't see anything like #color under ComposableModel. > > cheers -ben >

Nov. 9, 2013