February 2009 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Announcements for Pharo?

by Tudor Girba

That would be cool, actually. I have used the implementation of Lukas and it works just fine. Lukas pointed out that the VW Announcements code is under Cincom copyright. The nice thing about his implementation is that he did not look at the code. He only implemented the idea as he read it from the blog :). Cheers, Doru On 26 Feb 2009, at 01:28, Michael Rueger wrote: > Hi all, > > should we add Announcements as part of the core image? > And if so, which of the existing implementations? > > Michael > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project -- www.tudorgirba.com "Presenting is storytelling."

Feb. 26, 2009

[Pharo-project] Announcements for Pharo?

by Michael Rueger

Hi all, should we add Announcements as part of the core image? And if so, which of the existing implementations? Michael

Feb. 26, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Alexandre Bergel

> Please, please, keep UI feel homogeneous ! > Why not HOME, END, CTRL + -> like any other text editor in the world ? > Even squeak is following that convention in TextMorphs ! > Also let SHIFT+CONTROL+-> select over words and typed characters > overwrite > selection. > You don't have to imitate a poor man's shell! You could have several ways to get at the first character. Actually, not only "could", but "should". Why not "home", "end" ? because there is no such keys on a mac. > Let history cycle when at beginning/end of list (with some form of > advertising). > And let restrict history crawling among lines matching the first > letters if you > typed some (like Matlab does). As most shell do as well. Would be great to have, indeed. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Alexandre Bergel

> I would say that a command like "ls" is not a method. It's an object > (instance of a sublass of "Script" or "Command"). > > You want, for example, inherit things like command line parsing, > standard options > (e.g. --help) and things liket that. A command like ls is far too > complex to be > just one method. could also be a trait. But I understand what you're saying. Entering 'ls' will actually be translated into a 'ls executeInContext: currentContext' or something. currentContext will then be aware of the current path. Looks like to be an exciting project! Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Nicolas Cellier

Marcus Denker <denker@...> writes: > > > On 25.02.2009, at 22:27, Matthew Fulmer wrote: > > >> > >> Each command will then be a message sent to the object referenced by > >> the variable 'shell'. > >> > >> entering 'ls' will then invoke 'shell.ls'... > > I would say that a command like "ls" is not a method. It's an object > (instance of a sublass of "Script" or "Command"). > > You want, for example, inherit things like command line parsing, > standard options > (e.g. --help) and things liket that. A command like ls is far too > complex to be > just one method. > > Marcus > > -- > Marcus Denker -- denker@... > http://www.marcusdenker.de > I would say ls is a message, and Shell ls answer with a LSCommand which is kindOf: Command. A Command use lazy evaluation: it does not evaluate until a end of line or pipe message triggers evaluation (i mean send the message #value). Before being evaluated, Command behaviour can be modified by sending all sort of message, like -l --help etc... Of course, apart options which are messages sent to the command (eventually with arguments), a list of arguments can be passed to the Command itself (Command valueWithArguments: aList). Just have to invent the syntax now... Nicolas

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Nicolas Cellier

Michael van der Gulik <mikevdg@...> writes: > > > On Thu, Feb 26, 2009 at 9:26 AM, Alexandre Bergel <alexandre(a)bergel.eu> wrote: > > please > > try it out and let me know what you would expect from a real Smalltalk > > scripting environment. > Something that would really push me in using Coral is an intelligent > prompt. A list of the things the prompt should do: > - crawling into the history of typed commands using up and down key > arrow > - tab completion (classes and method name) > - colored output > - Ctl-D to quit > - Keybinds: > * Ctlr-A beginning of the line > * Ctrl-E end of the line > * Alt- -> / Alt- <- to jump words > > Please, please, keep UI feel homogeneous ! Why not HOME, END, CTRL + -> like any other text editor in the world ? Even squeak is following that convention in TextMorphs ! Also let SHIFT+CONTROL+-> select over words and typed characters overwrite selection. You don't have to imitate a poor man's shell! Let history cycle when at beginning/end of list (with some form of advertising). And let restrict history crawling among lines matching the first letters if you typed some (like Matlab does). Nicolas

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Marcus Denker

On 25.02.2009, at 22:27, Matthew Fulmer wrote: >> >> Each command will then be a message sent to the object referenced by >> the variable 'shell'. >> >> entering 'ls' will then invoke 'shell.ls'... I would say that a command like "ls" is not a method. It's an object (instance of a sublass of "Script" or "Command"). You want, for example, inherit things like command line parsing, standard options (e.g. --help) and things liket that. A command like ls is far too complex to be just one method. Marcus -- Marcus Denker -- denker(a)acm.org http://www.marcusdenker.de

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Matthew Fulmer

On Wed, Feb 25, 2009 at 10:14:30PM +0100, Alexandre Bergel wrote: > My dream is to completely get rid of bash and other awful shells. > > I remember to have seen a talk on a OO shell, where each command like > 'cat', 'ls', 'cd' where actually message sent to an instance of a > class Shell. The receiver would then be implicit. > Imagine a prompt like: > shell. > > Each command will then be a message sent to the object referenced by > the variable 'shell'. > > entering 'ls' will then invoke 'shell.ls'... At least in theory, quoth already does this http://netjam.org/quoth/ -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

Feb. 25, 2009

Re: [Pharo-project] Pharo scripting -- reborn as "Coral"

by Alexandre Bergel

My dream is to completely get rid of bash and other awful shells. I remember to have seen a talk on a OO shell, where each command like 'cat', 'ls', 'cd' where actually message sent to an instance of a class Shell. The receiver would then be implicit. Imagine a prompt like: shell. Each command will then be a message sent to the object referenced by the variable 'shell'. entering 'ls' will then invoke 'shell.ls'... Cheers, Alexandre On 25 Feb 2009, at 13:55, Oscar Nierstrasz wrote: > > Hi Folks, > > I spent some time yesterday together with Stef to take a closer look > at the Pharo scripting project. > > We have repackaged and renamed the project as "Coral". It is available > on SqueakSource at: > > http://www.squeaksource.com/Coral > > You can now load it simply into a fresh Pharo-dev image by evaluating: > > ScriptLoader new installer ss project: 'Coral'; install: 'LoadCoral'. > > This will load everything that is required and pop up a few Workspace > windows with further instructions. (You will need John McIntosh's > latest VM to get this to work -- all is described in the image.) > > There is also a window with TO-DO items. I am interested in starting > a student project to really make this fly, so I would be grateful for > your suggestions and desiderata. If you are a potential user, please > try it out and let me know what you would expect from a real Smalltalk > scripting environment. > > Thanks again to Mathieu and GwenaÃ«l for getting this to work! > > Oscar > --- > Prof. Dr. O. Nierstrasz -- Oscar.Nierstrasz(a)iam.unibe.ch > Software Composition Group -- http://www.iam.unibe.ch/~scg > University of Bern -- Tel/Fax +41 31 631.4618/3355 > vcard: http://www.iam.unibe.ch/~oscar/oscarNierstrasz.vcf > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Feb. 25, 2009

Re: [Pharo-project] MIMEdatabase fix?

by Michael Rueger

Michael Rueger wrote: > OK, here we go... > in the Pharo inbox there is a new version of Network-MIME and > KomHTTPServer. The Kom package should also work on Squeak. ok, one more thing ;-) Newer version of MIME package plus the attached change for Seaside makes at least the tests pass. Michael 'From Pharo0.1 of 16 May 2008 [Latest update: #10243] on 25 February 2009 at 1:09:34 pm'! !SeasidePlatformSupport class methodsFor: 'factory' stamp: 'michael.rueger 2/25/2009 13:05'! mimeDocumentOfType: type from: aFilename "WACachedDocument clearCache. (WACachedDocument fileName: 'SqueakDebug.log') asMIMEDocument. (WACachedDocument fileName: 'SqueakDebug.log') asMIMEDocument." | content mimeType | mimeType := type ifNil: [ self mimeDocumentClass guessContentTypeFromExtension: (FileDirectory extensionFor: aFilename) ]. content := (StandardFileStream readOnlyFileNamed: aFilename) contentsOfEntireFile. ^self mimeDocumentClass contentType: mimeType content: content! !

Feb. 25, 2009