May 2015 - Pharo-users - lists.pharo.org

Re: [Pharo-users] Pharo 4 Playground open file menu

by Esteban Lorenzano

> On 02 May 2015, at 23:28, phil(a)highoctane.be wrote: > > When some sysadmin has to edit them on servers, you want them in .st files. > > No class. No IDE. Not too much Smalltalk. > but then - if not smalltalk, the scripts should not be in the imageâ¦ even in workspaces - if the sysadmin has to edit them, he can always do something like: #! /bin/bash pharo MyImage.image eval â my multilined more or less smalltalk scriptâ - you can always see and edit your scripts by doing: 'play-cache' asFileReference inspect (instead âplay-cacheâ you can put: âmy-script-folderâ, whatever) and you will have a complete inspector that allows you to see and edit your scripts (who are in the file system, where a sysadmin can find them, and not in an obscure workspace). also I bet you would take no more than 5â to add functionality to gtinspector (it is designed to be moldable, after all) to add new scripts, and no matter what other functionality you needâ¦ and the result will be a lot more âpharoishâ than storing it in a workspace. Esteban > Just the DSL on an as needed to know basis to configure things. > > That's better that XML/YAML/JSON... > > So, that's the case. > > Startup scripts same story. > > Phil > > Le 2 mai 2015 17:56, "Esteban Lorenzano" <estebanlm(a)gmail.com <mailto:estebanlm@gmail.com>> a Ã©crit : > wellâ¦ IMO those scripts also should be in a method. > Probably under a class named: MyCoolProjectRunScripts or something like thatâ¦ but in a class. > If they are in a class you can: > - save them with your project > - version them > - if you add <script> pragma, you can even execute them by clicking on it (Pharo 4). > > soâ¦ even if you might have a case where you want the save/loadâ¦ you actually have (what I consider) a better option. > > Esteban > >> On 02 May 2015, at 15:17, Esteban A. Maringolo <emaringolo(a)gmail.com <mailto:emaringolo@gmail.com>> wrote: >> >> Ditto here. >> >> I have cron tasks that fire a smalltak script, the startup script itself or a small import script that doesn't belong to a class. All those are my cases for the workspace. >> >> El may 2, 2015 4:38 AM, "phil(a)highoctane.be <mailto:phil@highoctane.be>" <phil(a)highoctane.be <mailto:phil@highoctane.be>> escribiÃ³: >> playground cache is actually not nice when scripts are to be part of a project with a name etc. And I have a ton of files in it. I can't remember which is which. >> >> I have scripts to do lots of cli things and I like the save as of the workspace. >> >> But I have done extra key bindings for getting the ws or the playground. >> >> Phil >> Le 2 mai 2015 06:49, "Sergio Fedi" <sergio.fedi(a)gmail.com <mailto:sergio.fedi@gmail.com>> a Ã©crit : >> Oh! That link gives a GREAT explanation! Thanks. >> >> On the subject of how to show it better, I'm not a graphic designer (although I'm working with one) >> so I'll ask him for some insights on the matter. >> >> >> For now, he pointed out some issues like lack of consistency on some interfaces >> and some other details. >> >

May 3, 2015

Re: [Pharo-users] Can we have NeoCSV, and NeoJSON included in the base image?

by Stephan Eggermont

On 02/05/15 23:28, Andy Burnett wrote: > Stephan Eggermont said: > What is stopping you from adding a build 'Andy's Base Image' to the > Pharo-Contribution CI? > > Ignorance :-) Always the same lame excuse :) > I had no idea such a thing was possible. It sounds like the ideal > solution. Is the process documented somewhere? Create yourself an account at ci.inria.fr, ask to be added to pharo-contribution, create a new build by copying one that looks somewhat similar. Once you are a member, you can browse all the different build configurations. You probably want to copy from a build that generates both Pharo4 and Pharo5 images (and older ones for the things you haven't migrated yet). A build just needs a configuration. Stephan

May 3, 2015

Re: [Pharo-users] accessing themed colors from morph

by Esteban Lorenzano

> On 02 May 2015, at 22:15, Nicolai Hess <nicolaihess(a)web.de> wrote: > > > > 2015-05-01 23:54 GMT+02:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com <mailto:i.uhnak@gmail.com>>: > Thank's for the pointers. It seems to me that they access it via 'self theme'; in any case my problem is, for the moment, solved. > > Yes but, 'self theme' is 'the theme engine that defines all theme settings' and > 'self theme backgroundColor' is > 'what this theme defines as general backgroundColor' > It does not mean that self (the morph) is actually using this theme setting for anything. > > For a panel morph, that normally uses 'Color transparent' for its color, would answer to 'self theme backgroundColor' with the same answer as > any other morph. It is the 'theme backgroundColor' not the morphs background color. > > > Thanks, > Peter > > On Fri, May 1, 2015 at 11:11 PM, Nicolai Hess <nicolaihess(a)web.de <mailto:nicolaihess@web.de>> wrote: > > > 2015-05-01 22:22 GMT+02:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com <mailto:i.uhnak@gmail.com>>: > Hi, > > how is set the background color of a morph? > > A morph has the fillStyle/color attribute, but this does not mean it uses this color for its drawing method. > (for example ActiveHand color -> Color blue, and the hand/cursor is not blue). > > > I could access directly "Smalltalk ui theme backgroundColor" from within the Morph, > > this returns always the same color, it is actually the desktop backgroundcolor. no, it is not. it is the general background color (that should be applied *by default* to all morphs that recognises the background property), as Nicolai pointed. Nowâ¦ all that is really messy and bad structured (and worst documented)â¦ so do not expect it works all the time. Esteban > > but I don't see anywhere that something is using themed colors this way. > So I assume there is some other way; what am I missing? > > Note that I don't want to use the color for background of the morph but I need to make a decision based on the color in the morph. > > It depends on the morph, or the type of widget this morph represents. > A button has a theme dependent color, that may depend on the type and state (enabled/disabled) of the button. > A list morph may have different colors for even and odd index, selected, enabled, disabled ... > A label model font/color may depend on its use (window title/menu title/ button label ...). > > Look at the xxxFillStyleFor: messages in UITheme > > > > Thanks, > Peter > > >

May 3, 2015

Re: [Pharo-users] Use antlr ".g4" grammars-v4 files to generate parser in pharo?

by Thierry Goubier

Le 03/05/2015 10:42, Julien Delplanque a Ã©crit : > Hi everyone, > > I saw on this repository https://github.com/antlr/grammars-v4 that there > are a lot of grammar defined. I wonder if someone already managed to > generate a parser in pharo (using any framework you want) from a ".g4" > file? Not me, even if I have spent time pouring over ANTLR-based grammars when implementing compiler front-ends. ANTLR-based grammars are LL* (right recursive), so they don't fit very well with LALR/GLR (SmaCC), they tend to look a bit different from the programming language official doc or standard (i.e left-recursive), they are in some cases difficult to reuse because resolution may be done in ad-hoc fashion (adding arbitrary code to resolve ambiguities), and whether the grammar is really complete for a given language is hard to determine (working grammars for real-life complete uses are expensive to build, so they tend to be kept as company Intellectual Property). This is my experience. > I also saw there is the grammar of ".g4" files > (https://github.com/antlr/grammars-v4/tree/master/antlr4) in the > repository so I guess it would be possible to create a parser that > generate parsers from ".g4" files. This is correct. Thierry > > Regards, > > Julien > > >

May 3, 2015

Use antlr ".g4" grammars-v4 files to generate parser in pharo?

by Julien Delplanque

Hi everyone, I saw on this repository https://github.com/antlr/grammars-v4 that there are a lot of grammar defined. I wonder if someone already managed to generate a parser in pharo (using any framework you want) from a ".g4" file? I also saw there is the grammar of ".g4" files (https://github.com/antlr/grammars-v4/tree/master/antlr4) in the repository so I guess it would be possible to create a parser that generate parsers from ".g4" files. Regards, Julien

May 3, 2015

Re: [Pharo-users] Pharo 4 Playground open file menu

by itlists＠schrievkrom.de

Am 02.05.2015 um 23:28 schrieb phil(a)highoctane.be: > When some sysadmin has to edit them on servers, you want them in .st files. > > No class. No IDE. Not too much Smalltalk. > > Just the DSL on an as needed to know basis to configure things. > > That's better that XML/YAML/JSON... > > So, that's the case. > > Startup scripts same story. > More or less the same use case as I have .... Marten -- Marten Feldtmann

May 3, 2015

Building ObjVlisp language chapter

by Mark Rizun

Hi guys! Currently, I'm following the chapter: "Building ObjVlisp a Minimal, Uniform and Reflective Object-Oriented Language Kernel" (http://sdmeta.gforge.inria.fr/Teaching/ObjVLisp/) While doing this exercise I realized how cool it is :) Anywhays, I encountered a little problem which I solved, so it's more like a comment/feedback: In section 1.7 Message Passing and Dynamic Lookup, there are mentioned several methods, that are not actually implemented. Moreover, the readers task is to implement method #lookup: selector. But the skeleton(mcz package in the link above) does not contain this method. In the end of this section, it is said, that you have to run #testNilWhenErrorInLookup and #testRaisesErrorSendWhenErrorInLookup test methods. But, they does not exist too:) However, in test class there is a category called "tests-MessagePassingsAndDynamicsLookup" and it uses method #lookup: selector for: anObjObject. This method is empty, and when I implemented it, all tests from message passing category become green. Eventually my comment about this is: I'm guessing that either chapter or code(mcz file in the link I mentioned) is not up to date. Does anyone know if the rest of the chapter/code is not up to date? Or maybe I made a mistake and just couldn't find what I had to? P.S. I really recommend to follow this chapter/exercise :) Best, Mark -- View this message in context: http://forum.world.st/Building-ObjVlisp-language-chapter-tp4823907.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

May 3, 2015

Re: [Pharo-users] Can we have NeoCSV, and NeoJSON included in the base image?

by Sergio Fedi

> > (...) That could be integrated in the open window when starting a fresh > image. Either by GUI or by evaluating an expression that does the load. If > you have ideas for that just go ahead and provide a configuration of the > tools you find important and give it a name that is meaningful to others. > That seems perfect.

May 2, 2015

Re: [Pharo-users] Can we have NeoCSV, and NeoJSON included in the base image?

by Norbert Hartl

> Am 02.05.2015 um 17:23 schrieb Andy Burnett <andy.burnett(a)knowinnovation.com>: > > I would find it really useful to have some more classes included in the base image. The Neo system would be particularly useful, but I would also find an XML reader valuable. > > What is the process for deciding which classes get included by default? It is a discussion that comes up frequently. But the purpose of using the image is rather a subjective thing. I personally don't want to have things in the image I don't use because I need to remove them before going into production. We have metacello configurations which make it easy to load parts of the stack people like to use. Looking at the installation procedure of a debian/ubuntu image you are asked what is the purpose of this installation be it a mail server, web server, etc. Maybe the best way to tackle this is to have a definition of typical stacks and an easy to use selection of a proper metacello configuration that loads everything someone does not know where to get it. That could be integrated in the open window when starting a fresh image. Either by GUI or by evaluating an expression that does the load. If you have ideas for that just go ahead and provide a configuration of the tools you find important and give it a name that is meaningful to others. Norbert

May 2, 2015

Re: [Pharo-users] Pharo 4 Playground open file menu

by Tudor Girba

As explained before by Andrei and I, if you rename the tab, you will get a named file in play-stash and you can find this file afterwards by name via Spotter. You can also choose a custom location for play-stash in the settings browser. Cheers, Doru On Sat, May 2, 2015 at 11:30 PM, phil(a)highoctane.be <phil(a)highoctane.be> wrote: > never tought of renaming the tab. > I was somewhat aware of the feature. > > But I have the cache in my .gitignore > > Phil > Le 2 mai 2015 22:10, "Tudor Girba" <tudor(a)tudorgirba.com> a Ã©crit : > >> Hi, >> >> Did you try renaming the playground page as explained before? If yes, >> besides the fact that in order to open it you have to search it by the >> name, is there something else that is not working for you? >> >> Cheers, >> Doru >> >> >> On Sat, May 2, 2015 at 3:17 PM, Esteban A. Maringolo < >> emaringolo(a)gmail.com> wrote: >> >>> Ditto here. >>> >>> I have cron tasks that fire a smalltak script, the startup script itself >>> or a small import script that doesn't belong to a class. All those are my >>> cases for the workspace. >>> El may 2, 2015 4:38 AM, "phil(a)highoctane.be" <phil(a)highoctane.be> >>> escribiÃ³: >>> >>>> playground cache is actually not nice when scripts are to be part of a >>>> project with a name etc. And I have a ton of files in it. I can't remember >>>> which is which. >>>> >>>> I have scripts to do lots of cli things and I like the save as of the >>>> workspace. >>>> >>>> But I have done extra key bindings for getting the ws or the playground. >>>> >>>> Phil >>>> Le 2 mai 2015 06:49, "Sergio Fedi" <sergio.fedi(a)gmail.com> a Ã©crit : >>>> >>>>> Oh! That link gives a GREAT explanation! Thanks. >>>>> >>>>> On the subject of how to show it better, I'm not a graphic designer >>>>> (although I'm working with one) >>>>> so I'll ask him for some insights on the matter. >>>>> >>>>> >>>>> For now, he pointed out some issues like lack of consistency on some >>>>> interfaces >>>>> and some other details. >>>>> >>>>> >> >> >> -- >> www.tudorgirba.com >> >> "Every thing has its own flow" >> > -- www.tudorgirba.com "Every thing has its own flow"

May 2, 2015