From renggli at gmail.com Fri Mar 3 09:30:06 2006 From: renggli at gmail.com (Lukas Renggli) Date: Fri, 3 Mar 2006 09:30:06 +0100 Subject: Porting Magritte over to VisualWorks.. In-Reply-To: <44079BBA.5090403@ca-flower.com> References: <24C40DFA333DC44882F9FB0115F33D8F499FC4@ARGON.insario.local> <67628d690602091412m3cd20dd9s656d6e4d7cd4de72@mail.gmail.com> <44079BBA.5090403@ca-flower.com> Message-ID: <67628d690603030030p741dd950j86c87f7cd3c5df21@mail.gmail.com> > > This is a subclass of the Squeak to VisualWorks code exporter > > (implemented by Avi) adapted to properly export all the Magritte and > > Pier packages and to get the namespaces right. Running the exporter > > makes it possible to load all the code into VisualWorks, but I never > > had the time to make it actually run in there ... so that is what > > requires the next steps ... > > > Lukas, > > Just wondering if I wanted to go down this path, where can I find this > code exporter tool -- is > it on the Squeak side or the Visual Works side? I'd like to give it a > shot even though I'm somewhat > green behind the ears! First of all you need PackageInfo-Exporters from SqueakMap. Then you need the package "Magritte-VisualWorks" from . Evaluating "MADistribution fileOutForVisualWorks" should produce you a couple of files that can be loaded into VisualWorks, however I never actually tried to do it ;-) Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From frank.urbach at schmees.com Mon Mar 6 11:06:53 2006 From: frank.urbach at schmees.com (Frank Urbach) Date: Mon, 6 Mar 2006 10:06:53 +0000 Subject: Problem Editing-Field of Contents in Pier Message-ID: After one update of pier I have only the MATextInputFieldComponent in the section PRContentsWidget. There was the MATextAreaComponent before and I didn't find the place changing the widget from MATextInputFieldComponent to MATextAreaComponent. Thanks in advance, Frank -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20060306/db1c21e0/attachment.html From chad at clearwaterassociates.ca Tue Mar 7 03:06:41 2006 From: chad at clearwaterassociates.ca (Chad Nantais) Date: Mon, 6 Mar 2006 21:06:41 -0500 Subject: Magritte Tutorial - MAPersonEditor Message-ID: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> I've called a MAPersonEditor from my WATask, but pressing "Save" in the editor does not seem to answer to the task. Do I need to override a save message or something? Thanks, Chad From chad at clearwaterassociates.ca Tue Mar 7 04:13:25 2006 From: chad at clearwaterassociates.ca (Chad Nantais) Date: Mon, 6 Mar 2006 22:13:25 -0500 Subject: Magritte Tutorial - MAPersonEditor In-Reply-To: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> References: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> Message-ID: <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> Nevermind that last message. I was overcomplicating the problem. All I had to do was change my change one method in my task. (from this) getPerson ^ self call: (MAPersonEditor new) (to this) getPerson ^self call: (MAPersonModel new asComponent addValidatedForm; yourself). A very simple, time-saving solution. Plus, didn't need to use the MAPersonEditor. I can see Magritte being a big help when building a questionnaire or survey application which involves many input forms. Cheers, Chad On 3/6/06, Chad Nantais wrote: > I've called a MAPersonEditor from my WATask, but pressing "Save" in > the editor does not seem to answer to the task. Do I need to override > a save message or something? > > Thanks, > > Chad > From renggli at iam.unibe.ch Tue Mar 7 07:06:12 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Tue, 7 Mar 2006 07:06:12 +0100 Subject: Magritte Tutorial - MAPersonEditor In-Reply-To: <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> References: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> Message-ID: <4BCC0467-949C-435D-861A-B2DC807C9BE2@iam.unibe.ch> Hi Chad, welcome to the list and thanks for answering your question ;-) > getPerson > ^self call: (MAPersonModel new asComponent > addValidatedForm; > yourself). Indeed you have to use Seaside decorations to add a validation- handler and a form handler. This allows to customize those two aspects, say if you want multiple Magritte components within one form, different button labels, different button actions, different error handling, etc. > I can see Magritte being a big help when building a questionnaire or > survey application which involves many input forms. Certainly, but there are many more thing you can do: building reports, reflecting about model objects, letting users build their own forms from custom descriptions, ... Letting users customize descriptions can be done like ... model := MAScaffolder description: MAPersonModel description. model asComponent -> an editor onto the model-object model description asComponent -> an editor onto the description (meta-model) Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From damien.cassou at laposte.net Tue Mar 7 07:06:37 2006 From: damien.cassou at laposte.net (Damien Cassou) Date: Tue, 07 Mar 2006 07:06:37 +0100 Subject: Magritte Tutorial - MAPersonEditor In-Reply-To: <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> References: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> Message-ID: <440D22ED.2060408@laposte.net> Chad Nantais a ?crit : > Nevermind that last message. I was overcomplicating the problem. All > I had to do was change my change one method in my task. > > (from this) > getPerson > ^ self call: > (MAPersonEditor new) > > (to this) > getPerson > ^self call: (MAPersonModel new asComponent > addValidatedForm; > yourself). For those who has not understand why the first won't work: In fact, MAPersonEditor is just a demo to see how easily one can describe a model and have the component updated. MAPersonEditor should _not_ be used in another application as it never returns the modified models. -- Damien Cassou From damien.cassou at laposte.net Tue Mar 7 07:08:34 2006 From: damien.cassou at laposte.net (Damien Cassou) Date: Tue, 07 Mar 2006 07:08:34 +0100 Subject: Magritte Tutorial - MAPersonEditor In-Reply-To: <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> References: <3de931600603061806r51e8672boeb8d5c9b3d7013f3@mail.gmail.com> <3de931600603061913i2f39bf80rc918490fb220df63@mail.gmail.com> Message-ID: <440D2362.1040101@laposte.net> Chad Nantais a ?crit : > Nevermind that last message. I was overcomplicating the problem. All > I had to do was change my change one method in my task. > > (from this) > getPerson > ^ self call: > (MAPersonEditor new) > > (to this) > getPerson > ^self call: (MAPersonModel new asComponent > addValidatedForm; > yourself). For those who has not understand why the first won't work: In fact, MAPersonEditor is just a demo to see how easily one can describe a model and have the component updated. MAPersonEditor should _not_ be used in another application as it never returns the modified models. -- Damien Cassou From cbeler at enit.fr Tue Mar 7 20:44:31 2006 From: cbeler at enit.fr (=?ISO-8859-1?Q?C=E9drick_B=E9ler?=) Date: Tue, 07 Mar 2006 20:44:31 +0100 Subject: Pier as default seaside application Message-ID: <440DE29F.4060904@enit.fr> Hi I made an image with Pier as the default application (http://cdrick.seasidehosting.st) but when I want to access component I have a problem because it tries to access them via another path... cdrick.seasidehosting.st/mini-calendar instead of .../seaside/mini-calendar I change the default application after having made the image... nevertheless, it's also impossible to add new pages, component and so on... Is it because I made the image before changing the default path ? Also I noticed there isn't the remove command ... How to remove an existing structure ? See you C?drick *//* -------------------------------- L'ENIT vous invite a sa journee portes ouvertes le 17 mars 2006 de 13h30 a 19h30 Enit , 47 avenue d'Azereix 65000 Tarbes Bus N?1, arret ENI From renggli at iam.unibe.ch Tue Mar 7 21:43:18 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Tue, 7 Mar 2006 21:43:18 +0100 Subject: Pier as default seaside application In-Reply-To: <440DE29F.4060904@enit.fr> References: <440DE29F.4060904@enit.fr> Message-ID: <17B032E5-C09E-40B1-8FBA-886A8217C26F@iam.unibe.ch> Hi C?drick, > I made an image with Pier as the default application (http:// > cdrick.seasidehosting.st) but when I want to access component I > have a problem because it tries to access them via another path... > cdrick.seasidehosting.st/mini-calendar instead of .../seaside/ > mini-calendar Looks cool ;-) Yes, this is a known problem of Seaside, making it impossible to strip of the parts of the path not ment for Pier. The problem is, that the WAApplication doesn't know how it is nested into WADispatcher, and in your case this is not the same as the base-path. You seem not to have loaded the latest code of Pier. In there there is a ugly hack, fixing this particular problem by looking at the path and trying to strip-off /seaside/pier if this makes sense. Well, only a temporary solution, but it should solve your problem > Also I noticed there isn't the remove command ... How to remove an > existing structure ? I don't see any commands on your Pier instance? Did you install a Security plugin? Else you can always delete a page by sending an URL of the following form: "/mypage?command=PRRemoveCommand". This should show clearly how important it is to properly setup one of the available security frameworks for all public accessible pier sites ;-) Lukas -- Lukas Renggli http://www.lukas-renggli.ch From cbeler at enit.fr Wed Mar 8 10:14:56 2006 From: cbeler at enit.fr (=?ISO-8859-1?Q?C=E9drick_B=E9ler?=) Date: Wed, 08 Mar 2006 10:14:56 +0100 Subject: Pier as default seaside application In-Reply-To: <17B032E5-C09E-40B1-8FBA-886A8217C26F@iam.unibe.ch> References: <440DE29F.4060904@enit.fr> <17B032E5-C09E-40B1-8FBA-886A8217C26F@iam.unibe.ch> Message-ID: <440EA090.50409@enit.fr> >Looks cool ;-) > >Yes, this is a known problem of Seaside, making it impossible to >strip of the parts of the path not ment for Pier. The problem is, >that the WAApplication doesn't know how it is nested into >WADispatcher, and in your case this is not the same as the base-path. > >You seem not to have loaded the latest code of Pier. In there there >is a ugly hack, fixing this particular problem by looking at the path >and trying to strip-off /seaside/pier if this makes sense. Well, only >a temporary solution, but it should solve your problem > > uhmm I thought I had the latest one... what's funny is it work for WACalendar in the environment and not in the content panel. Anyway, I'll check or I'll provide the full web adress as I did for some to avoid to load all scripts and style sheets for the ajax demos. That makes me ask you if we have to load all of them in the main pier configuration if needed in child components ? > > >>Also I noticed there isn't the remove command ... How to remove an >>existing structure ? >> >> > >I don't see any commands on your Pier instance? Did you install a >Security plugin? Else you can always delete a page by sending an URL >of the following form: "/mypage?command=PRRemoveCommand". This should >show clearly how important it is to properly setup one of the >available security frameworks for all public accessible pier sites ;-) > > Oh yes I understand :) actually I didn't... I just removed the tree and command widget from the environment... I 'll see what I can do C?drick -------------------------------- L'ENIT vous invite a sa journee portes ouvertes le 17 mars 2006 de 13h30 a 19h30 Enit , 47 avenue d'Azereix 65000 Tarbes Bus N?1, arret ENI From thf at impara.de Wed Mar 8 15:55:40 2006 From: thf at impara.de (Thomas Froeb) Date: Wed, 08 Mar 2006 15:55:40 +0100 Subject: [Pier] Internal Link Problems Message-ID: <440EF06C.6050600@impara.de> Hi, since the structures have a "name" beside the "title" it's very difficult to create internal Links when you renamed (rather "retitled") a structure. You have to always remember the name you gave the structure first because if you want to create internal links you have to use now this name else the link is broken. I don't think thats a good Idea because no one can guess the hidden "name" of structures.... Regards, Thomas From dominiqued at versateladsl.be Thu Mar 9 01:01:43 2006 From: dominiqued at versateladsl.be (Dominique Dutoit) Date: Thu, 9 Mar 2006 01:01:43 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: <440EF06C.6050600@impara.de> References: <440EF06C.6050600@impara.de> Message-ID: Hi, I agree with you about the issue you are raising. In fact, it's the only reason why I am holding back my plan to upgrade to the latest version of Pier. We use a slightly customised Pier to manage our intranet and it was relatively easy to explain to the users that if they want to create a link to a structure, they just have to put the title of the target. The situation was not that smooth when we were using Plone and custom CMS. Even after a while, a large portion of our user base still had troubles to make the difference between the name of a structure and its title when it came to create links. The consequence was that the sites produced during that period contained a lot of endless pages and very few links. So far, we get a better balance between cross-referencing and page size, thanks to the sheer simplicity of the way Pier manages its links. I fear that we will lose this simplicity if we introduce back the same concepts that didn't performed that well during our previous attempts. Regards, Dominique Le 08-mars-06 ? 15:55, Thomas Froeb a ?crit : > Hi, > > since the structures have a "name" beside the "title" it's very > difficult to create internal Links when you renamed (rather > "retitled") > a structure. > You have to always remember the name you gave the structure first > because if you want to create internal links you have to use now this > name else the link is broken. I don't think thats a good Idea > because no > one can guess the hidden "name" of structures.... > > Regards, > Thomas From dominiqued at versateladsl.be Thu Mar 9 01:30:22 2006 From: dominiqued at versateladsl.be (Dominique Dutoit) Date: Thu, 9 Mar 2006 01:30:22 +0100 Subject: PRPage and empty contents Message-ID: <966CFB84-3C95-42F3-9AE6-92F507426984@versateladsl.be> If the text area for the contents is left empty, PRPage>>contents: receives a nil object and fails to parse it as a document. The reason is that the parser can only interpret string objects and not nil objects. PRDocumentParser parse: nil. The solution could be to assign an empty PRDocument if the value of aString is an nil object. PRPage>>contents: aString aString ifNil: [ self document: PRDocument new ] ifNotNil: [ self document: (self parserClass parse: aString) ]. From dominiqued at versateladsl.be Thu Mar 9 02:01:21 2006 From: dominiqued at versateladsl.be (Dominique Dutoit) Date: Thu, 9 Mar 2006 02:01:21 +0100 Subject: MADateDescription and nil objects Message-ID: <7886DFC8-AF6C-4BA2-8B50-4067898C2DC7@versateladsl.be> I have a couple of MADateDescription to manage the effective date and the expiration date of a structure and their values could be either a valid date or nil. The problem is that once a date object is assigned to these variables through the fields generated by MADateDescription, it is not possible to assign a nil value just by deleting the content of the fields. MADateDescription always tries to parse the field as a date. 1. Left the field empty, effectiveDate is nil. 2. Input a date in the field, effectiveDate is a valid date. 3. Delete the content of the field, MADateDescription complains about an invalid date format. Is there a way to avoid that parsing? From renggli at iam.unibe.ch Thu Mar 9 09:15:07 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 9 Mar 2006 09:15:07 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: References: <440EF06C.6050600@impara.de> Message-ID: Hi Dominique, Hi Thomas, > since the structures have a "name" beside the "title" it's very > difficult to create internal Links when you renamed (rather > "retitled") a structure. thanks for rising the issue with the 'name' and 'title'. To introduce a name was a difficult decision for me, and it is definitely not a final decision, I just wanted to give it a try. The reasons were the following: 1. People want to change the title of pages without that the URL changes, so that they don't need to update their bookmarks. 2. People want to use Pier to model existing (static) pages and want to model the URL to look the exactly the same. 3. People want use strange characters and very long titles, this makes the URL unreadable and difficult to parse and generate. 4. People want use the same title twice within the same context, this is impossible if we use the title as an unique key. > You have to always remember the name you gave the structure first > because if you want to create internal links you have to use now > this name else the link is broken. I don't think thats a good Idea > because no one can guess the hidden "name" of structures.... It is not so hidden after all, you can just copy the part of the URL into your link. > The situation was not that smooth when we were using Plone and custom > CMS. Even after a while, a large portion of our user base still had > troubles to make the difference between the name of a structure and > its title when it came to create links. The consequence was that the > sites produced during that period contained a lot of endless pages > and very few links. Ok, I see that this can be a problem, but I guess that there also a couple of solutions to this problem: a. Let the user browse a tree navigator to help them writing the links. b. Display the current name more clearly in the edit page. > So far, we get a better balance between cross-referencing and page > size, thanks to the sheer simplicity of the way Pier manages its > links. I fear that we will lose this simplicity if we introduce back > the same concepts that didn't performed that well during our previous > attempts. The current trial is definitely a loss of simplicity, I agree, however I don't know of another simple way to get 1. - 4. working only having a title. If there are any ideas how these problems could be cleanly solved otherwise, I would be happy to discuss and maybe (help to) implement them? Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From damien.cassou at laposte.net Thu Mar 9 09:23:06 2006 From: damien.cassou at laposte.net (Damien Cassou) Date: Thu, 09 Mar 2006 09:23:06 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: References: <440EF06C.6050600@impara.de> Message-ID: <440FE5EA.2050001@laposte.net> I often wanted to change a title of a page without changing links to it so the difference between title and name is useful I think. However, it should be clear to the user how to reference a page. Why not displaying in each page (in view mode) the name of the page ? Something like 'Link me with "/myName"' would be cool. -- Damien Cassou From renggli at iam.unibe.ch Thu Mar 9 09:25:41 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 9 Mar 2006 09:25:41 +0100 Subject: PRPage and empty contents In-Reply-To: <966CFB84-3C95-42F3-9AE6-92F507426984@versateladsl.be> References: <966CFB84-3C95-42F3-9AE6-92F507426984@versateladsl.be> Message-ID: <29333754-B9C2-4283-ABDA-C21949338DF7@iam.unibe.ch> > If the text area for the contents is left empty, PRPage>>contents: > receives a nil object and fails to parse it as a document. Thanks for reporting this problem. I will fix it. ;-) Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Thu Mar 9 09:26:43 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 9 Mar 2006 09:26:43 +0100 Subject: MADateDescription and nil objects In-Reply-To: <7886DFC8-AF6C-4BA2-8B50-4067898C2DC7@versateladsl.be> References: <7886DFC8-AF6C-4BA2-8B50-4067898C2DC7@versateladsl.be> Message-ID: > I have a couple of MADateDescription to manage the effective date and > the expiration date of a structure and their values could be either a > valid date or nil. How does the description-definition of your MADateDescription look like? I quickly tried on my machine and it works as expected. > 3. Delete the content of the field, MADateDescription complains about > an invalid date format. Sorry, I am unable to reproduce. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From thf at impara.de Thu Mar 9 11:41:58 2006 From: thf at impara.de (Thomas Froeb) Date: Thu, 09 Mar 2006 11:41:58 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: <440FE5EA.2050001@laposte.net> References: <440EF06C.6050600@impara.de> <440FE5EA.2050001@laposte.net> Message-ID: <44100676.5000909@impara.de> > I often wanted to change a title of a page without changing links to it > so the difference between title and name is useful I think. However, it > should be clear to the user how to reference a page. > > Why not displaying in each page (in view mode) the name of the page ? > Something like 'Link me with "/myName"' would be cool. > > > 1. People want to change the title of pages without that the URL > changes, so that they don't need to update their bookmarks. > > 2. People want to use Pier to model existing (static) pages and want > to model the URL to look the exactly the same. > > 3. People want use strange characters and very long titles, this > makes the URL unreadable and difficult to parse and generate. > > 4. People want use the same title twice within the same context, this > is impossible if we use the title as an unique key. I think it's a good idea to have a name beside the title. I also intoduced this in my Smallwiki 1 for my XML persistency to have constant directeries (references) in the filesystem and I use this in Pier too. I fixed the problem with the internal links for me by changing the PRPathLookup >> visitChildren: method so that it finds the child also by the title. It is not very clean but works for my purposes. Perhaps it is better to change the reference of a link - after editing a page - to the right path? Regards, Thomas From renggli at iam.unibe.ch Thu Mar 9 13:49:43 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 9 Mar 2006 13:49:43 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: <44100676.5000909@impara.de> References: <440EF06C.6050600@impara.de> <440FE5EA.2050001@laposte.net> <44100676.5000909@impara.de> Message-ID: <3E535A92-BE9F-4E5D-9CC4-682CAC8E01D7@iam.unibe.ch> > I fixed the problem with the internal links for me by changing the > PRPathLookup >> visitChildren: method so that it finds the child also > by the title. It is not very clean but works for my purposes. Yes, that sounds like a simple solution: when parsing a path it first tries to look it up by name, if it fails by title; paths are always printed by name. > Perhaps it is better to change the reference of a link - after > editing a > page - to the right path? I think this happens already. At least the paths should be automatically rewritten if you rename the target page. What do others think about this? Cheers, Lukas > > Regards, Thomas > > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Thu Mar 9 19:20:16 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 9 Mar 2006 19:20:16 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: <3E535A92-BE9F-4E5D-9CC4-682CAC8E01D7@iam.unibe.ch> References: <440EF06C.6050600@impara.de> <440FE5EA.2050001@laposte.net> <44100676.5000909@impara.de> <3E535A92-BE9F-4E5D-9CC4-682CAC8E01D7@iam.unibe.ch> Message-ID: <85EE36DC-D6DB-450F-BB61-83F774F5A765@iam.unibe.ch> >> I fixed the problem with the internal links for me by changing the >> PRPathLookup >> visitChildren: method so that it finds the child >> also >> by the title. It is not very clean but works for my purposes. > > Yes, that sounds like a simple solution: when parsing a path it first > tries to look it up by name, if it fails by title; paths are always > printed by name. Pier-All-lr.72 does exactly that. It does not fix the reference string automatically thou. Moreover it tries to be smarter when stripping of the seaside parts of the path. This works well for standard applications, if you make your application the default one (to be run when accessing /seaside) bookmarking won't work at all, due to a limitation in Seaside. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From dominiqued at versateladsl.be Thu Mar 9 23:06:02 2006 From: dominiqued at versateladsl.be (Dominique Dutoit) Date: Thu, 9 Mar 2006 23:06:02 +0100 Subject: [Pier] Internal Link Problems In-Reply-To: <3E535A92-BE9F-4E5D-9CC4-682CAC8E01D7@iam.unibe.ch> References: <440EF06C.6050600@impara.de> <440FE5EA.2050001@laposte.net> <44100676.5000909@impara.de> <3E535A92-BE9F-4E5D-9CC4-682CAC8E01D7@iam.unibe.ch> Message-ID: <39133925-74A4-43B5-AB3C-6CBADE503859@versateladsl.be> Hi Thomas, Hi Lukas, >> I fixed the problem with the internal links for me by changing the >> PRPathLookup >> visitChildren: method so that it finds the child >> also by the title. It is not very clean but works for my purposes. > > Yes, that sounds like a simple solution: when parsing a path it > first tries to look it up by name, if it fails by title; paths are > always printed by name. > ... > What do others think about this? It sounds like a good compromise for me. The interim solution (as implemented in Pier-All.72) already feels much simpler to deal with. Thanks! Dominique From renggli at iam.unibe.ch Fri Mar 10 14:01:37 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Fri, 10 Mar 2006 14:01:37 +0100 Subject: seaside hosting - decompression In-Reply-To: <4411743C.7040203@enit.fr> References: <440EB713.3000207@enit.fr> <440ECD9D.5060301@enit.fr> <67C166DA-6140-4249-85EB-FDC88E5EAEC7@iam.unibe.ch> <44100288.50303@enit.fr> <6D1D795D-607E-4585-B98F-D8DA29C803D2@iam.unibe.ch> <4411743C.7040203@enit.fr> Message-ID: <882C2A1C-DB9F-4B4F-A45D-BCF5735E9CF1@iam.unibe.ch> Hi C?drick, > I'm still doing tests on how to model my exprience attributes with > Magritte... I don't avance as I'd like but I get a better > understanding on how Magritte and Seaside works... I think I > mainly have to use existing descriptions while creating specific > views as you suggested to me whereas at start, I was trying to do > specific descriptions... > > ExpAttribute are a kind of proxy on a real value (traditional > attribute) to which I add a simple attribute (range) that reflects > an analyses (traduces the relevance of the attribute in the > occurence of the event). > > I hope on monday I'll be able to show you something better... > I've made some hacks in Magritte like the possibility to put > descriptions in a collection in the class side ( not sure useful > though :) but in that way I can see and change all of them in one > place). Also I added a revert (or cancel) buttons to > MAOneToOneComponent so as to cancel the memento and choose another > classe... the descriptions are already within a collection accessible from the class-side, even-tough you define them within separate methods. Send the message #description to your class and you get your collection. Well actually it is a bit more, you get an instance of MAPriorityContainer supporting the whole collection protocol. > I ve tried to adapt what Michel Bany did (http:// > cdrick.seasidehosting.st/seaside/LiveTests the scrollbar) to the > new renderer to have a MARangeComponent compatible with all browser > but without success... Does it seems possible to you as it uses > some functions of seaside Async ? anyway, I'll see that point > later... No, MARangeComponent only works in KHTML based browsers, such as Safari or Konqueror. It is based on the draft of the next XHTML standard and it gently degenerates to a plain input box if not supported. I don't use SeasideAsync, there are a few experimental controls using script.aculo.us. This library works nicely together with seasides canvas framework and would also provide a cross-browser compatible slider, that could be integrated if I (or somebody else) had the time. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From dominiqued at versateladsl.be Sat Mar 11 18:10:42 2006 From: dominiqued at versateladsl.be (Dominique Dutoit) Date: Sat, 11 Mar 2006 18:10:42 +0100 Subject: MADateDescription and nil objects In-Reply-To: References: <7886DFC8-AF6C-4BA2-8B50-4067898C2DC7@versateladsl.be> Message-ID: <45514941-7696-478C-88C3-374987E70D6D@versateladsl.be> Hi Lukas >> 3. Delete the content of the field, MADateDescription complains about >> an invalid date format. > > Sorry, I am unable to reproduce. Hum, I shouldn't trust my own memory. In fact, the problem has nothing to do with the date format but when Magritte is committing the changes to the original object. I can reproduce the problem with a form structure (Pier.72, Magritte. 142). 1) Create a form 2) Edit the meta of that form 3) Add a date element 4) Save the element 5) Save the meta 6) Edit the form 7) Assign a date 8) Save the form 9) Edit the form 10) Select the date and delete it (the field should be empty) 11) Save the form At that moment, I get this traceback: MessageNotUnderstood: UndefinedObject>>start UndefinedObject(Object)>>doesNotUnderstand: #start self nil temps aMessage start Date(Timespan)>>= self 2 March 2006 temps comparand nil inst vars start 2006-03-02T00:00:00+00:00 duration 1:00:00:00 [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key ifAbsent: [^ false]) = assoc v...]} self a Dictionary(a MADateDescription label: 'Date' comment: nil- >nil a MAStringDescription label: 'Title...etc... temps aDictionary a Dictionary(a MADateDescription label: 'Date' comment: nil->2 March 2006 a MAStringDescription labe...etc... assoc a MADateDescription label: 'Date' comment: nil->nil inst vars tally 2 array #(nil nil a MAStringDescription label: 'Title' comment: nil- >'Untitled' nil a MADateDescription labe...etc... Dictionary(Set)>>do: self a Dictionary(a MADateDescription label: 'Date' comment: nil- >nil a MAStringDescription label: 'Title...etc... temps aBlock [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key ifAbsent: [^ false]) = assoc v......etc... index 5 each a MADateDescription label: 'Date' comment: nil->nil indexLimiT 6 inst vars tally 2 array #(nil nil a MAStringDescription label: 'Title' comment: nil- >'Untitled' nil a MADateDescription labe...etc... Dictionary>>associationsDo: self a Dictionary(a MADateDescription label: 'Date' comment: nil- >nil a MAStringDescription label: 'Title...etc... temps aBlock [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key ifAbsent: [^ false]) = assoc v......etc... inst vars tally 2 array #(nil nil a MAStringDescription label: 'Title' comment: nil- >'Untitled' nil a MADateDescription labe...etc.. From renggli at iam.unibe.ch Sun Mar 12 01:02:19 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 12 Mar 2006 01:02:19 +0100 Subject: MADateDescription and nil objects In-Reply-To: <45514941-7696-478C-88C3-374987E70D6D@versateladsl.be> References: <7886DFC8-AF6C-4BA2-8B50-4067898C2DC7@versateladsl.be> <45514941-7696-478C-88C3-374987E70D6D@versateladsl.be> Message-ID: Hi, >>> 3. Delete the content of the field, MADateDescription complains >>> about >>> an invalid date format. >> >> Sorry, I am unable to reproduce. > > Hum, I shouldn't trust my own memory. In fact, the problem has > nothing to do with the date format but when Magritte is committing > the changes to the original object. > > I can reproduce the problem with a form structure (Pier.72, Magritte. > 142). > > 1) Create a form yes, now I can observe some problems. I have changed and simplified a lot in Pier in the past, but never looked at the PRForm "meta" things for quite a while, so it is likely that there are some problems in that area. I will try to fix these problems tomorrow. Cheers, Lukas > 2) Edit the meta of that form > 3) Add a date element > 4) Save the element > 5) Save the meta > > 6) Edit the form > 7) Assign a date > 8) Save the form > > 9) Edit the form > 10) Select the date and delete it (the field should be empty) > 11) Save the form > > At that moment, I get this traceback: > > MessageNotUnderstood: UndefinedObject>>start > UndefinedObject(Object)>>doesNotUnderstand: #start > self nil > temps > aMessage start > Date(Timespan)>>= > self 2 March 2006 > temps > comparand nil > inst vars > start 2006-03-02T00:00:00+00:00 > duration 1:00:00:00 > [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key ifAbsent: > [^ false]) = assoc v...]} > self a Dictionary(a MADateDescription label: 'Date' comment: nil- >> nil a MAStringDescription label: 'Title...etc... > temps > aDictionary a Dictionary(a MADateDescription label: 'Date' comment: > nil->2 March 2006 a MAStringDescription labe...etc... > assoc a MADateDescription label: 'Date' comment: nil->nil > inst vars > tally 2 > array #(nil nil a MAStringDescription label: 'Title' comment: nil- >> 'Untitled' nil a MADateDescription labe...etc... > Dictionary(Set)>>do: > self a Dictionary(a MADateDescription label: 'Date' comment: nil- >> nil a MAStringDescription label: 'Title...etc... > temps > aBlock [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key > ifAbsent: [^ false]) = assoc v......etc... > index 5 > each a MADateDescription label: 'Date' comment: nil->nil > indexLimiT 6 > inst vars > tally 2 > array #(nil nil a MAStringDescription label: 'Title' comment: nil- >> 'Untitled' nil a MADateDescription labe...etc... > Dictionary>>associationsDo: > self a Dictionary(a MADateDescription label: 'Date' comment: nil- >> nil a MAStringDescription label: 'Title...etc... > temps > aBlock [] in Dictionary>>= {[:assoc | (aDictionary at: assoc key > ifAbsent: [^ false]) = assoc v......etc... > inst vars > tally 2 > array #(nil nil a MAStringDescription label: 'Title' comment: nil- >> 'Untitled' nil a MADateDescription labe...etc.. > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch From cbeler at enit.fr Mon Mar 13 16:17:23 2006 From: cbeler at enit.fr (=?ISO-8859-1?Q?C=E9drick_B=E9ler?=) Date: Mon, 13 Mar 2006 16:17:23 +0100 Subject: magritte questions In-Reply-To: <882C2A1C-DB9F-4B4F-A45D-BCF5735E9CF1@iam.unibe.ch> References: <440EB713.3000207@enit.fr> <440ECD9D.5060301@enit.fr> <67C166DA-6140-4249-85EB-FDC88E5EAEC7@iam.unibe.ch> <44100288.50303@enit.fr> <6D1D795D-607E-4585-B98F-D8DA29C803D2@iam.unibe.ch> <4411743C.7040203@enit.fr> <882C2A1C-DB9F-4B4F-A45D-BCF5735E9CF1@iam.unibe.ch> Message-ID: <44158D03.8080203@enit.fr> Hi :) > the descriptions are already within a collection accessible from the > class-side, even-tough you define them within separate methods. Send > the message #description to your class and you get your collection. > Well actually it is a bit more, you get an instance of > MAPriorityContainer supporting the whole collection protocol. Actually I've been learning quite a bit how descriptions are generated as it's necessary for what I want to achieve. To sum up, I have an attribute which is a container (wrapper) that points to an object model (domain) and to an analysis element (a simple object that has a value) nb: for now I have two parallel hierarchies. One for ExpAttributes and one for the domain objects. ExpAttribute model:= aMAPerson analyse:= anAnalyse All my domain objects are meta described as the analyse one. So now I have an experience that is composed of ExpAttribute and for these I'd like to dynamically generate the description what is not so easy as the model element is not of the same type for all ExpAttribute (I only differentiate Attribute that are Leaf (equivalent to Magritte descriptions minus MARelationDescription...) and those that are composed (equivalent to MARelationDescription)). Whereas on your slides, "dynamic descriptions" are build from existing descriptions on class side, here I need to build them according to the object wrapped in the ExpAttribute>>model So that is why I was interested in description generation. Is it possible for instance to overwrite description like: Model>>description container := super description <- return a MAPriorityContainer container add: aNewDescription etc... doing like that, I had a problem because descriptions are added over and over after each new session because MACachedBuilder keeps a cache with the previous descriptions. I've overwritten in my ExpAttribute class description ^MABuilder default build: self It now works but not sure it's a good solution... What is the purpose of MACachedBuilder instead of MABuilder? avoiding creating descriptions over and over? Also I was thinking of doing that: Model>>description ^MASpecificBuilder new build: self So before going further, can you tell me if it is a good practice - to redefine description on the instance side ? - to use a specific builder ? Is there a better way to create dynamically descriptions according to the Magritte framework (when some descriptions are accessible in the class side and others in class side of other objects class side...)... Maybe I should do that with a MAScafolder ?? >> I ve tried to adapt what Michel Bany did (http:// >> cdrick.seasidehosting.st/seaside/LiveTests the scrollbar) to the >> new renderer to have a MARangeComponent compatible with all browser >> but without success... Does it seems possible to you as it uses >> some functions of seaside Async ? anyway, I'll see that point later... > > > No, MARangeComponent only works in KHTML based browsers, such as > Safari or Konqueror. It is based on the draft of the next XHTML > standard and it gently degenerates to a plain input box if not > supported. Yes I knew that... but what I tried was to adapt the scrollbar component as provided in the exemple (http:// cdrick.seasidehosting.st/seaside/LiveTests) to WARenderCanvas. It's a live div but Michel Bany (I think he made that) did it for WAHtmlRenderer. WAHtmlRenderer>>scrollbarWithHeight: heightNumber size: sizeNumber liveCallback: liveBlock | id uri | id := self ensureId. self attributeAt: 'style' put: 'position: relative; left: -3px; width: 19px; overflow: auto; height: ', heightNumber displayString 'px'. self div: [self attributeAt: 'style' put: 'width: 1px; height: ', sizeNumber displayString, 'px'. self div: '']. uri := self **urlForLiveRequest:** [:t :h | liveBlock value: t asInteger value: h]. self script: 'document.getElementById(' , id asString printString , ').onscroll = liveScroll(' , id asString printString , ', ' , (self class encode: **uri** asString) printString , ')' I just started to do it "like" you did with MARangeComponent (renderEditorOn: and updateScript) so as to adapt to WARenderCanvas but I didnt' succeed yet and as it is not that important, I gave up for now ;)... > > I don't use SeasideAsync, there are a few experimental controls using > script.aculo.us. This library works nicely together with seasides > canvas framework and would also provide a cross-browser compatible > slider, that could be integrated if I (or somebody else) had the time. ok didn't know that :) where can I find the javascript code for that ? see you C?drick -------------------------------- L'ENIT vous invite a sa journee portes ouvertes le 17 mars 2006 de 13h30 a 19h30 Enit , 47 avenue d'Azereix 65000 Tarbes Bus N?1, arret ENI From renggli at iam.unibe.ch Mon Mar 13 17:04:10 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 13 Mar 2006 17:04:10 +0100 Subject: magritte questions In-Reply-To: <44158D03.8080203@enit.fr> References: <440EB713.3000207@enit.fr> <440ECD9D.5060301@enit.fr> <67C166DA-6140-4249-85EB-FDC88E5EAEC7@iam.unibe.ch> <44100288.50303@enit.fr> <6D1D795D-607E-4585-B98F-D8DA29C803D2@iam.unibe.ch> <4411743C.7040203@enit.fr> <882C2A1C-DB9F-4B4F-A45D-BCF5735E9CF1@iam.unibe.ch> <44158D03.8080203@enit.fr> Message-ID: <4230B1F1-0C0F-4DAB-B23F-243A35DBC738@iam.unibe.ch> Hi C?drick, > To sum up, I have an attribute which is a container (wrapper) that > points to an object model (domain) and to an analysis element (a > simple object that has a value) > nb: for now I have two parallel hierarchies. One for ExpAttributes > and one for the domain objects. > > ExpAttribute > model:= aMAPerson > analyse:= anAnalyse > > All my domain objects are meta described as the analyse one. > > So now I have an experience that is composed of ExpAttribute and > for these I'd like to dynamically generate the description what is > not so easy as the model element is not of the same type for all > ExpAttribute (I only differentiate Attribute that are Leaf > (equivalent to Magritte descriptions minus > MARelationDescription...) and those that are composed (equivalent > to MARelationDescription)). > > Whereas on your slides, "dynamic descriptions" are build from > existing descriptions on class side, here I need to build them > according to the object wrapped in the ExpAttribute>>model Mhh ... it is a bit hard to understand how the objects play together, maybe some diagram could help to understand your object model? > So that is why I was interested in description generation. Is it > possible for instance to overwrite description like: > > Model>>description > container := super description <- return a MAPriorityContainer > container add: aNewDescription etc... Yes, it is common to do something like this. You can always override #description on the class- or on the instance-side of any model. > I had a problem because descriptions are added over and over after > each new session because MACachedBuilder keeps a cache with the > previous descriptions. Hehe, I thought I said that somewhere in the slides: you have to manually make a copy of your original description, else you modify the original description that is cached by Magritte. Model>>description container := super description copy. container add: aNewDescription > What is the purpose of MACachedBuilder instead of MABuilder? > avoiding creating descriptions over and over? Yes, the current way of creating descriptions is not really efficient. I have to manually collect and compose the descriptions from special methods, taking inheritance and overrides into play, etc. This will become simpler as soon as Pragmas will be standard in Squeak ;-) > Model>>description > ^MASpecificBuilder new build: self > > So before going further, can you tell me if it is a good practice > - to redefine description on the instance side ? > - to use a specific builder ? > > Is there a better way to create dynamically descriptions according > to the Magritte framework (when some descriptions are accessible in > the class side and others in class side of other objects class > side...)... Yes, you can do that. It is possible hook for Magritte, however I never really used it myself. I prefer to do static stuff in methods on the class-side and let Magritte cache it, and to do dynamic stuff on the instance-side and take a copy of what Magritte returns as a prototype. > Maybe I should do that with a MAScafolder ?? Yes, MAScaffolder is doing it the way I described. >>> I ve tried to adapt what Michel Bany did (http:// >>> cdrick.seasidehosting.st/seaside/LiveTests the scrollbar) to >>> the new renderer to have a MARangeComponent compatible with all >>> browser but without success... Does it seems possible to you >>> as it uses some functions of seaside Async ? anyway, I'll see >>> that point later... >> >> No, MARangeComponent only works in KHTML based browsers, such as >> Safari or Konqueror. It is based on the draft of the next XHTML >> standard and it gently degenerates to a plain input box if not >> supported. > > Yes I knew that... but what I tried was to adapt the scrollbar > component as provided in the exemple (http:// > cdrick.seasidehosting.st/seaside/LiveTests) to WARenderCanvas. It's > a live div but Michel Bany (I think he made that) did it for > WAHtmlRenderer. I never worked much with the Seaside life framework. I do only advertise script.aculo.us and they have a slider that works pretty well, it is not integrated into Seaside yet but that should be simple. >> I don't use SeasideAsync, there are a few experimental controls >> using script.aculo.us. This library works nicely together with >> seasides canvas framework and would also provide a cross-browser >> compatible slider, that could be integrated if I (or somebody >> else) had the time. > > ok didn't know that :) > where can I find the javascript code for that ? Demos: http://scriptaculous.seasidehosting.st Source: http://www.squeaksource.com/Seaside.html Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From thf at impara.de Mon Mar 20 17:26:12 2006 From: thf at impara.de (Thomas Froeb) Date: Mon, 20 Mar 2006 17:26:12 +0100 Subject: [Pier] Document Parser problem Message-ID: <441ED7A4.1070208@impara.de> Hi, If I trie to create a Page with "_ *aLink* text" as content the Document Parser runs into an error "Token not expected". I tried to convince the Document Parser to accept a link behind a horizontal rule but I don't get the right rules for the parser and scanner. I think the parser creates a paragraph and don't expects a horizontal rule there. If I add HorizontalRule to paragraph start and let it return an empty array it works but creates beside the horizontal rule a paragraph if i have linebreak after the "_". Any ideas how to solve it a better way? Regards, Thomas From renggli at iam.unibe.ch Mon Mar 20 20:18:54 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 20 Mar 2006 20:18:54 +0100 Subject: [Pier] Document Parser problem In-Reply-To: <441ED7A4.1070208@impara.de> References: <441ED7A4.1070208@impara.de> Message-ID: > If I trie to create a Page with "_ *aLink* text" as content the > Document > Parser runs into an error "Token not expected". That should not happen, I added a test-case to catch that bug. > I tried to convince the Document Parser to accept a link behind a > horizontal rule but I don't get the right rules for the parser and > scanner. I think the parser creates a paragraph and don't expects a > horizontal rule there. The bug is fixed in "Pier-Model-lr.52.mcz". Thanks for reporting. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From family.entwisle at btinternet.com Tue Mar 21 11:21:05 2006 From: family.entwisle at btinternet.com (Duncan Entwisle) Date: Tue, 21 Mar 2006 10:21:05 +0000 Subject: Changing admin password? Message-ID: Please could someone tell me the proper way to change the administrator's password in Pier? Last week, I could do it by using the Pier Browser, exploring the root, and sending a password: message to the adminUser instance. I've now created a new image, with the latest Pier packages from Squeak Map. Even though the password appears to change in the instance, the old passwords (user=admin, password=pier) remain in use. I realise I'm probably doing something stupid :-) Thanks for any help, Duncan. P.S. With the latest Squeak Map Pier package the Pier Browser doesn't appear to be added to the Tools flap like last time. From renggli at iam.unibe.ch Tue Mar 21 18:17:06 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Tue, 21 Mar 2006 18:17:06 +0100 Subject: Changing admin password? In-Reply-To: References: Message-ID: > I've now created a new image, with the latest Pier packages from > Squeak > Map. Even though the password appears to change in the instance, the > old passwords (user=admin, password=pier) remain in use. This is strange, it works in my image when exactly following your steps. Are you sure that you modified the right kernel? Mmh, while trying to reproduce your bug I discovered that there is a nasty recursion when logging out. I will publish a new version on SqueakMap shortly that fixes this issue. I also found out that there is a problem with the latest version of Seaside and the default action in forms. Later tonight I will publish a new version of Magritte including Seaside, but first I have to check with Avi. > P.S. With the latest Squeak Map Pier package the Pier Browser doesn't > appear to be added to the Tools flap like last time. Yes, I removed that, because it did not work in Squeak 3.9. It is still in the World Menu > Open > Pier Browser. One of my goals for Magritte and Pier is to slowly move to 3.9, because it offers things (Pragmas) that will make our life easier and that I have been waiting for a long long time now. For now I will try to keep compatibility for 3.7, 3.8 and 3.9. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From rleon at insario.com Tue Mar 21 18:30:02 2006 From: rleon at insario.com (Ramon Leon) Date: Tue, 21 Mar 2006 10:30:02 -0700 Subject: Changing admin password? Message-ID: <24C40DFA333DC44882F9FB0115F33D8F49A293@ARGON.insario.local> > One of my goals for Magritte and Pier is to slowly move to > 3.9, because it offers things (Pragmas) that will make our > life easier and that I have been waiting for a long long time > now. For now I will try to keep compatibility for 3.7, 3.8 and 3.9. > > Lukas I would hope that 3.8 remains the primary target at the very least until after 3.9 is actually released. From rleon at insario.com Tue Mar 21 18:36:04 2006 From: rleon at insario.com (Ramon Leon) Date: Tue, 21 Mar 2006 10:36:04 -0700 Subject: Pier Security Message-ID: <24C40DFA333DC44882F9FB0115F33D8F49A294@ARGON.insario.local> Why does Pier-Security depend on Pier-Tests? I don't want to have to load the test package just to load security. From renggli at iam.unibe.ch Tue Mar 21 19:03:58 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Tue, 21 Mar 2006 19:03:58 +0100 Subject: Pier Security In-Reply-To: <24C40DFA333DC44882F9FB0115F33D8F49A294@ARGON.insario.local> References: <24C40DFA333DC44882F9FB0115F33D8F49A294@ARGON.insario.local> Message-ID: > Why does Pier-Security depend on Pier-Tests? I don't want to have to > load the test package just to load security. Yes, because there are tests included with the security package. It is a bit annoying to have a special test-package for small packages such as the security one. So I don't know a solution for now ... I guess that you can just proceed after the warning of Monticello and it should load everything except for the security tests. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Tue Mar 21 19:05:21 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Tue, 21 Mar 2006 19:05:21 +0100 Subject: Changing admin password? In-Reply-To: <24C40DFA333DC44882F9FB0115F33D8F49A293@ARGON.insario.local> References: <24C40DFA333DC44882F9FB0115F33D8F49A293@ARGON.insario.local> Message-ID: <36801D4D-BE2E-47BF-9C28-139C3F2B46DA@iam.unibe.ch> >> One of my goals for Magritte and Pier is to slowly move to >> 3.9, because it offers things (Pragmas) that will make our >> life easier and that I have been waiting for a long long time >> now. For now I will try to keep compatibility for 3.7, 3.8 and 3.9. >> >> Lukas > > I would hope that 3.8 remains the primary target at the very least > until > after 3.9 is actually released. Yes, certainly. However the next major versions (Magritte 1.1 and Pier 1.1) will be targeted at 3.9. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From family.entwisle at btinternet.com Wed Mar 22 20:54:30 2006 From: family.entwisle at btinternet.com (Duncan Entwisle) Date: Wed, 22 Mar 2006 19:54:30 +0000 Subject: Changing admin password? In-Reply-To: References: Message-ID: On 21 Mar 2006, at 17:17, Lukas Renggli wrote: > This is strange, it works in my image when exactly following your > steps. Are you sure that you modified the right kernel? Yes - I only had one. It does seem to work now (I re-recreated the image from scratch today). > Mmh, while trying to reproduce your bug I discovered that there is a > nasty recursion when logging out. I will publish a new version on > SqueakMap shortly that fixes this issue. Sorry - I did actually notice the log-out problem, but I thought it was something to do with my method of changing the password. Thanks for the quick turn-a-round, Duncan. From damien.cassou at laposte.net Thu Mar 23 08:57:43 2006 From: damien.cassou at laposte.net (Damien Cassou) Date: Thu, 23 Mar 2006 08:57:43 +0100 Subject: [Magritte] Pragmas Message-ID: <442254F7.4090106@laposte.net> Hi, how pragmas are planned to be used with magritte and pier ? Bye -- Damien Cassou From renggli at iam.unibe.ch Thu Mar 23 10:15:33 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 23 Mar 2006 10:15:33 +0100 Subject: [Magritte] Pragmas In-Reply-To: <442254F7.4090106@laposte.net> References: <442254F7.4090106@laposte.net> Message-ID: > how pragmas are planned to be used with magritte and pier ? The details have to be determined yet, it mainly helps to ensure extensibility and avoids to have naming conventions; probably there are other possibilites: * Magritte-Descriptions - Now: There is a naming convention that all descriptions have to be put on the class-side of the model and that their selectors should look like #description[A-Z][A-Za-z0-9]+. Not only this is extremely ugly but also forces other packages to override, if they want to extend another package. - Pragmas: It should be possible to define descriptions within different methods, so that different packages are able to customize the same description. Probably there are This could look like: EditPage class>>contentsDescription (core package) ^ MAMemoDescription selector: #contents EditPage class>>contentsDescriptionComponent: aDescription (seaside view) ^ aDescription componentClass: PRContentsEditor EditPage class>>contentsDescriptionPersistency: aDescription (persistency paackge) ^ aDescription persistencyStrategy: PROmniPersistent new * Pier-Parser - Now: The Pier-Wiki parser right now is quite limited, because extensions cannot modify or extend its behaviour without duplicating everything. - Pragmas: I would like so simplify it so that it triggers sub- parsers from line prefixes. These prefixes are defined as method- pragmas so that other packages are able to add their own extensions. Any other ideas? Lukas -- Lukas Renggli http://www.lukas-renggli.ch From frank.urbach at schmees.com Thu Mar 23 14:43:22 2006 From: frank.urbach at schmees.com (Frank Urbach) Date: Thu, 23 Mar 2006 13:43:22 +0000 Subject: Persistence of Pier Message-ID: Hi all, I'm searching the right way for saving my work in pier every 5 minutes automaticly. I've tryed to connect with the http://smallwiki.unibe.ch/smallwiki but no answer. Thanks in advance Cheers, Frank -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20060323/fe4aa8c6/attachment.html From renggli at iam.unibe.ch Thu Mar 23 15:15:32 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 23 Mar 2006 15:15:32 +0100 Subject: Persistence of Pier In-Reply-To: References: Message-ID: <331BDFF5-FB69-4029-893E-89804947B545@iam.unibe.ch> > I'm searching the right way for saving my work in pier every 5 > minutes automaticly. What about something like? [ (Delay forMilliseconds: 5 * 60 * 1000) wait. Smalltalk saveSession ] repeat > I've tryed to connect with the http://smallwiki.unibe.ch/smallwiki > but no answer. > Thanks in advance Thanks, there was again a problem with the machine. It seems to work again. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From btaylor at healthauditors.com Thu Mar 23 20:02:31 2006 From: btaylor at healthauditors.com (btaylor@healthauditors.com) Date: Thu, 23 Mar 2006 11:02:31 -0800 (PST) Subject: Moving the contents of Pier from one image to another Message-ID: <20060323110231.EUN99927@mp1.mc.surewest.net> Hi all, I have an instance of pier running in an image which seems to be hosed. The problem is that while Pier will still display pages, it will not allow them to be edited. This started sometime after I put in a back ground process to periodically save the image. So, I need to know how to migrate the contents of this wiki to another image. Any tips on this? BTW, I can still edit methods in the System browser, and execute workspace code. Thanks, Brett Taylor From renggli at iam.unibe.ch Thu Mar 23 20:29:31 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 23 Mar 2006 20:29:31 +0100 Subject: Moving the contents of Pier from one image to another In-Reply-To: <20060323110231.EUN99927@mp1.mc.surewest.net> References: <20060323110231.EUN99927@mp1.mc.surewest.net> Message-ID: <0DAD9CEF-ACE2-4A6A-8178-2CE5BD6F0344@iam.unibe.ch> > I have an instance of pier running in an image which seems to be > hosed. The problem is that while Pier will still display pages, it > will not allow them to be edited. This started sometime after I > put in a back ground process to periodically save the image. This is strange, did you check the process browser and tried to kill any suspicious processes? Did you reset the monitor in the kernel? PRKernel instances do: [ :e | e monitor: Monitor new ] > So, I need to know how to migrate the contents of this wiki to > another image. Any tips on this? BTW, I can still edit methods in > the System browser, and execute workspace code. ReferenceStream (simple and easy) http://minnow.cc.gatech.edu/squeak/2318 ImageSegments (fast, but complicated) http://minnow.cc.gatech.edu/squeak/1672 SIXX (cross platform) http://www.mars.dti.ne.jp/~umejava/smalltalk/sixx/index.html ... and a lot of other possiblities. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From stephane.ducasse at univ-savoie.fr Sun Mar 26 20:30:24 2006 From: stephane.ducasse at univ-savoie.fr (=?ISO-8859-1?Q?St=E9phane_Ducasse?=) Date: Sun, 26 Mar 2006 20:30:24 +0200 Subject: About magritte for XML serialization Message-ID: Hi lukas I wanted to know this was possible to specify how an field should be saved and reloaded. For example I wanted to save a date and reload it. I added MADateDescription>>parseString: aString ^ self parseString: aString protected: [ aString asDate ]. to MADateDescription :) but I have the impression that been able to specify at the description level the way the values are saved and filed in would make sense. I checked on my old version of MA, but I could not find how the MADescription itself was describing the condition because it would be the same process may be been able to specify a block or a selector for each of the action would be cool. Lukas? Stef From renggli at iam.unibe.ch Sun Mar 26 22:10:32 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 26 Mar 2006 22:10:32 +0200 Subject: About magritte for XML serialization In-Reply-To: References: Message-ID: <7E9E6767-C4EB-4D89-A43F-B4DD82BA5B1D@iam.unibe.ch> Hi Stef, > I wanted to know this was possible to specify how an field should > be saved and reloaded. For example I wanted to save a date and > reload it. sure, this is possible. > I added > > MADateDescription>>parseString: aString > ^ self parseString: aString protected: [ aString asDate ]. I don't have #parseString:protected: in my image, I guess you are using a very old version. In fact, in very early versions the parsing and printing of strings was fixed, but now it is done using a configurable visitor: have a look at the methods #toString: and #fromString: in MADescription. > but I have the impression that been able to specify at the > description level the way the values are saved and filed in would > make sense. > > I checked on my old version of MA, but I could not find how the > MADescription itself was describing the condition because it would > be the same process may be been able to specify a block or a > selector for each of the action would be cool. The visitor can be changed on a per-description base. It can also be chosen at the meta-description level, if you want to give this possibility to the end-users. The default visitors used are MAStringReader and MAStringWriter. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From ducasse at iam.unibe.ch Mon Mar 27 08:18:52 2006 From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=) Date: Mon, 27 Mar 2006 08:18:52 +0200 Subject: About magritte for XML serialization In-Reply-To: <7E9E6767-C4EB-4D89-A43F-B4DD82BA5B1D@iam.unibe.ch> References: <7E9E6767-C4EB-4D89-A43F-B4DD82BA5B1D@iam.unibe.ch> Message-ID: :) indeed Iwas a early adopter :) Stef On 26 mars 06, at 22:10, Lukas Renggli wrote: > Hi Stef, > >> I wanted to know this was possible to specify how an field should >> be saved and reloaded. For example I wanted to save a date and >> reload it. > > sure, this is possible. > >> I added >> >> MADateDescription>>parseString: aString >> ^ self parseString: aString protected: [ aString asDate ]. > > I don't have #parseString:protected: in my image, I guess you are > using a very old version. In fact, in very early versions the parsing > and printing of strings was fixed, but now it is done using a > configurable visitor: have a look at the methods #toString: and > #fromString: in MADescription. > >> but I have the impression that been able to specify at the >> description level the way the values are saved and filed in would >> make sense. >> >> I checked on my old version of MA, but I could not find how the >> MADescription itself was describing the condition because it would >> be the same process may be been able to specify a block or a >> selector for each of the action would be cool. > > The visitor can be changed on a per-description base. It can also be > chosen at the meta-description level, if you want to give this > possibility to the end-users. The default visitors used are > MAStringReader and MAStringWriter. > > Cheers, > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From philippe.marschall at gmail.com Mon Mar 27 15:35:31 2006 From: philippe.marschall at gmail.com (Philippe Marschall) Date: Mon, 27 Mar 2006 15:35:31 +0200 Subject: [Ann] StellDichEin VCard and ICal support for Pier Message-ID: <66666f210603270535ucba78d7w8f8117b61d1febe@mail.gmail.com> Hi StellDichEin is a Pier addon that allows you to create VCards and ICal events and todos in Pier like you would create structures (eg pages). Later these can be downloaded and imported (iCalendar, Outlook, ....). It not tested very roughly and very likely has several bugs. It can be found at http://mc.lukas-renggli.ch/pieraddons/ and depends on the following packages: - NetworkAddress (http://mc.lukas-renggli.ch/magritteaddons/) - ICal (http://squeaksource.com/ical/) - ICalMagritte (http://squeaksource.com/ical/) There is also a configuration map that should load all these automatically. There is also a very simple poll in the same repository. Cheers Philippe From ducasse at iam.unibe.ch Mon Mar 27 18:12:06 2006 From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=) Date: Mon, 27 Mar 2006 18:12:06 +0200 Subject: [Ann] StellDichEin VCard and ICal support for Pier In-Reply-To: <66666f210603270535ucba78d7w8f8117b61d1febe@mail.gmail.com> References: <66666f210603270535ucba78d7w8f8117b61d1febe@mail.gmail.com> Message-ID: <70BF789E-F9F1-4277-A43A-5A17B7FEE085@iam.unibe.ch> Coooool I want to play with that. Do you have a running example somewhere (aka seaside hosting)? Because this would be cool to show that to people. Is the ical package based on the work of david? Stef On 27 mars 06, at 15:35, Philippe Marschall wrote: > Hi > > StellDichEin is a Pier addon that allows you to create VCards and ICal > events and todos in Pier like you would create structures (eg pages). > Later these can be downloaded and imported (iCalendar, Outlook, ....). > > It not tested very roughly and very likely has several bugs. > > It can be found at http://mc.lukas-renggli.ch/pieraddons/ and depends > on the following packages: > > - NetworkAddress (http://mc.lukas-renggli.ch/magritteaddons/) > - ICal (http://squeaksource.com/ical/) > - ICalMagritte (http://squeaksource.com/ical/) > > There is also a configuration map that should load all these > automatically. > > There is also a very simple poll in the same repository. > > Cheers > Philippe > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From philippe.marschall at gmail.com Mon Mar 27 23:29:24 2006 From: philippe.marschall at gmail.com (Philippe Marschall) Date: Mon, 27 Mar 2006 23:29:24 +0200 Subject: [Ann] StellDichEin VCard and ICal support for Pier In-Reply-To: <70BF789E-F9F1-4277-A43A-5A17B7FEE085@iam.unibe.ch> References: <66666f210603270535ucba78d7w8f8117b61d1febe@mail.gmail.com> <70BF789E-F9F1-4277-A43A-5A17B7FEE085@iam.unibe.ch> Message-ID: <66666f210603271329i4cc3da74kc8dd05d3e8298710@mail.gmail.com> > Do you have a running example somewhere (aka seaside hosting)? http://spielverderber.seasidehosting.st/spielwiese > Is the ical package based on the work of david? Yes, I took it as a starting point. It contains no SW2 dependencies or render code anymore. Current known users are DabbleDB and an internal application at netstyle.ch. ICalMagritte adds magritte descriptions to events and todos. StellDichEin is just a wrapper around cards, todos and events. It doesn't really contain any code. It also doesn't include any rendering code and is not a calendaring or contanct management application. It just allows you to create and download VCards and ICal events and todos. (I saw something similar for dokuwik and had to build something better) Philippe From cbeler at enit.fr Tue Mar 28 11:38:30 2006 From: cbeler at enit.fr (=?ISO-8859-1?Q?C=E9drick_B=E9ler?=) Date: Tue, 28 Mar 2006 11:38:30 +0200 Subject: [Magritte] Pragmas In-Reply-To: References: <442254F7.4090106@laposte.net> Message-ID: <44290416.9050204@enit.fr> Hi :) I haven't been there for a week... and I see there are big changes in Magritte... I had an error while loading the last version Magritte-All-lr.148 (because of MACachedBuilder>>modified) I had the 142 version before... I also cant access the last test package version on monticello... so I'm thinking of preparing another image... Is it possible to remove all magritte related stuff and install a brand new one ( I have empty categories etc...). What is the "best way" to have a brand new magritte ? :) You are getting rif of MACachedBuilder ? there is MADescriptionBuilder, Pragma...Named ... Is there something important here to know or is it more about renaming and preparing pragma in magritte ? Thanks C?drick *//* From renggli at iam.unibe.ch Wed Mar 29 00:52:38 2006 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Wed, 29 Mar 2006 00:52:38 +0200 Subject: [Magritte] Pragmas In-Reply-To: <44290416.9050204@enit.fr> References: <442254F7.4090106@laposte.net> <44290416.9050204@enit.fr> Message-ID: > I haven't been there for a week... and I see there are big changes in > Magritte... No, no, not so big. > I had an error while loading the last version Magritte-All-lr.148 > (because of MACachedBuilder>>modified) I had the 142 version > before... I > also cant access the last test package version on monticello... so I'm > thinking of preparing another image... Is it possible to remove all > magritte related stuff and install a brand new one ( I have empty > categories etc...). What is the "best way" to have a brand new > magritte > ? :) Argh, this is due to a problem of Monticello being unable to load code from a newer version itself. If you experience problems loading code with Monticello retry with the latest version of Monticello. I guess it is the best to stick with your old code for now, there is nothing that changed much. > You are getting rif of MACachedBuilder ? there is > MADescriptionBuilder, Pragma...Named ... Is there something > important > here to know or is it more about renaming and preparing pragma in > magritte ? No, I just renamed the classes. - MADescriptionBuilder is the new common superclass of the builders. - MANamedBuilder is the new default builder that works about the same as the old MACachedBuilder. - MAPragmaBuilder is an experimental builder taking the descriptions from Pragmas, it should be loadable in old versions. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From chapbr+pier at sciencegeeks.org Fri Mar 31 22:51:28 2006 From: chapbr+pier at sciencegeeks.org (Brian Chapados) Date: Fri, 31 Mar 2006 12:51:28 -0800 (PST) Subject: "pretty" urls for links to internal pages? Message-ID: <59853.137.131.108.119.1143838288.squirrel@webmail.tuffmail.net> Hi, I'm new to smalltalk, seaside and pier. I want to use pier to host a small site, and if possible, I'd like to change the way internal links to other pages are rendered, so that the wiki text: *Introduction>introduction* produces the link: http://localhost:9090/seaside/pier/information/introduction instead of: http://localhost:9090/seaside/pier/information/introduction?view=PRDefaultView&command=PRViewCommand&_k=jFcQckKB&25&_s=lEBbpcqiBRMHlTFF Both links return the same URL when typed into the browser, so clearly view is the default command. I've tried searching around for information about this but I couldn't find anything obvious to me. While browsing the source code, I came across WAUrl adaptToContext: and WAAnchorTag goto:, which look potentially relevant, but I'm honestly not sure if I'm on the right track. I'd appreciate it if someone could point me in the right direction. Also, if this is a really bad idea for some reason, that would also be good to know. I'm using Sqeak-3.8, Seaside 2.63a3-lr.51 and Pier-all-lr.81 Thanks, Brian