From keith_hodges at yahoo.co.uk Sun Jan 6 22:06:17 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Sun, 06 Jan 2008 21:06:17 +0000 Subject: Easiest way to present Dates Message-ID: <478142C9.9040602@yahoo.co.uk> I have been looking for the easiest way to get a date to display with the desired format. By adding BlockContext-write: anObject description: aDescription ^ self value: anObject The following becomes possible: ^ MADateDescription new selectorAccessor: #orderDate; beWithinYears: 2; stringWriter: [ :i | i mmddyyyy ]; label: 'Order Date'; beReadonly; beSearchable; yourself Is there a better way? Is this worth adding to the mainstream? Keith From renggli at iam.unibe.ch Sun Jan 6 22:19:46 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 6 Jan 2008 22:19:46 +0100 Subject: Easiest way to present Dates In-Reply-To: <478142C9.9040602@yahoo.co.uk> References: <478142C9.9040602@yahoo.co.uk> Message-ID: <4C56C1CF-E389-49EC-BB2C-7B6EBCB5F221@iam.unibe.ch> > BlockContext-write: anObject description: aDescription > > ^ self value: anObject > > The following becomes possible: > > ^ MADateDescription new > selectorAccessor: #orderDate; > beWithinYears: 2; > stringWriter: [ :i | i mmddyyyy ]; > label: 'Order Date'; > beReadonly; > beSearchable; > yourself > > Is there a better way? > Is this worth adding to the mainstream? This is a nice trick. I am not a big fan of unnecessarily adding block- contexts to my model. As soon as you want to make your meta-model persistent or editable you run into troubles ... What I usually did for my applications is to subclass the standard string writer with my configuration for dates and times to generate the desired output. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080106/b2d92b8f/smime.bin From keith_hodges at yahoo.co.uk Sun Jan 6 22:52:06 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Sun, 06 Jan 2008 21:52:06 +0000 Subject: Easiest way to present Dates In-Reply-To: <4C56C1CF-E389-49EC-BB2C-7B6EBCB5F221@iam.unibe.ch> References: <478142C9.9040602@yahoo.co.uk> <4C56C1CF-E389-49EC-BB2C-7B6EBCB5F221@iam.unibe.ch> Message-ID: <47814D86.80209@yahoo.co.uk> > > This is a nice trick. I am not a big fan of unnecessarily adding > block-contexts to my model. As soon as you want to make your > meta-model persistent or editable you run into troubles ... > > What I usually did for my applications is to subclass the standard > string writer with my configuration for dates and times to generate > the desired output. > > Cheers, > Lukas Thanks Lukas, I guess I was just being lazy... I conceded in the end. Keith From keith_hodges at yahoo.co.uk Mon Jan 7 01:50:56 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Mon, 07 Jan 2008 00:50:56 +0000 Subject: Magritte-Scriptaculous now includes ShoreDatePicker Message-ID: <47817770.9090806@yahoo.co.uk> I have just added a ShoreComponents DatePicker to the Magritte-Scriptaculous library. This library already includes a scriptaculousified tree, and list composition components. Keith From keith_hodges at yahoo.co.uk Mon Jan 7 02:20:37 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Mon, 07 Jan 2008 01:20:37 +0000 Subject: Nifty Corners Message-ID: <47817E65.3000804@yahoo.co.uk> I just discovered the only rounded corners solution I have seen for IE... http://www.html.it/articoli/nifty/index.html I might give this a go for pier Keith From lochan94 at gmail.com Mon Jan 7 09:45:59 2008 From: lochan94 at gmail.com (Rajeev Lochan) Date: Mon, 7 Jan 2008 14:15:59 +0530 Subject: Nifty Corners In-Reply-To: <47817E65.3000804@yahoo.co.uk> References: <47817E65.3000804@yahoo.co.uk> Message-ID: <634d9f210801070045n22270f35s98c4a180e2676f81@mail.gmail.com> Hi Keith, I did give it a try a couple of days ago, but wasn't able to succeed. I was not able to get this code executed during onload window.onload=function(){ if(!NiftyCheck()) return; *Rounded("div#nifty","#377CB1","#9BD1FA");* } I suppose the explanation is given by Lukas in the following mail http://www.nabble.com/javascript-window.onLoad-not-called-td9655382.html#a9655382 An advanced version of the link you sent me is http://www.html.it/articoli/niftycube/index.html If you go to Example 13 in that, the Developer mentions the following "We've finally arrived at the last example. One of the frequent reported question is in fact how to make coexist Nifty Corners with other script, given the fact only one assignation to window.onload is possible." I was also unsuccessful in getting example 13 working. I am sure, I would have done some mistake somewhere, If you can get them working, it would be great to learn from you. By the way, I did find a working solution for Rounded Corners which uses 4 images and CSS and which works across all browsers including IE6. http://www.redmelon.net/tstme/4corners/ HTH, Rajeev On Jan 7, 2008 6:50 AM, Keith Hodges wrote: > I just discovered the only rounded corners solution I have seen for IE... > > http://www.html.it/articoli/nifty/index.html > > I might give this a go for pier > > Keith > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki > -- Rajeev Lochan Co-founder, AR-CAD.com http://www.ar-cad.com +91 9243468076 (Bangalore) 080 65355873 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080107/7f416446/attachment-0001.html From michal.perutka at gmail.com Wed Jan 9 12:11:27 2008 From: michal.perutka at gmail.com (Michal Perutka) Date: Wed, 9 Jan 2008 12:11:27 +0100 Subject: MADescription>>cssClass: Message-ID: Hallo, when I use #cssClass:, for example: descriptionInfo ^ (MAStringDescription auto: #info label: 'Info') beReadonly; cssClass: 'info'; yourself I don't see class="readonly info" in my page source code, only class="readonly". The problem seems to be in MAComponentRenderer>>classFor:, which ignores aDescription attributes. Right? Cheers, Michal -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080109/a6982989/attachment.html From ch.l.ngre at online.de Wed Jan 9 18:39:53 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Wed, 09 Jan 2008 18:39:53 +0100 Subject: Magritte examples in seaside/examples/examplebrowser Message-ID: <478506E9.9020109@online.de> Hello, trying to browse the Magritte examples in examplebrowser I found the following issues: One is with WAComponent class >> headerForExampleBrowser ^self name = self description ifTrue: [ self name ] ifFalse: [ self description , ' (' , self name , ')' ] The 'Comma' op is defined in MADescription and #, tries to send #asContainer to the following ByteStrings. An DNU Error is thrown. Another is: MAExampleEditor instancevariable 'description' is supposed to hold a Collection - at least it is used that way e.g.: MADescriptionEditor>> buildReport ^ (MAReport rows: self description description: MAElementDescription description) however it is initialized with: defaultDescription ^MAContainer new. I changed MAExampleEditor>> description ^ description ifNil: [ description := Array with: self defaultDescription ] (adding the Array constructor) and that works. I don't know if it breaks other things though... Thanks for any suggestions and hints. Cheers, Christoph From ch.l.ngre at online.de Wed Jan 9 18:54:07 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Wed, 09 Jan 2008 18:54:07 +0100 Subject: Magritte examples in seaside/examples/examplebrowser In-Reply-To: <478506E9.9020109@online.de> References: <478506E9.9020109@online.de> Message-ID: <47850A3F.2040502@online.de> Ch Lamprecht schrieb: > trying to browse the Magritte examples in examplebrowser I found the following > issues: > I changed > MAExampleEditor>> > description > ^ description ifNil: [ > description := Array with: self defaultDescription ] > > (adding the Array constructor) and that works. I don't know if it breaks other > things though... Oh yes, now I see: That breaks the edit/preview views - > Thanks for any suggestions and hints. > > Cheers, Christoph From keith_hodges at yahoo.co.uk Thu Jan 10 13:23:40 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Thu, 10 Jan 2008 12:23:40 +0000 Subject: [Ann] MAOneToManyDocumentsComponent Message-ID: <47860E4C.6080801@yahoo.co.uk> Name: Magritte-Scriptaculous-kph.17 Author: kph Time: 10 January 2008, 12:22:37 pm UUID: bb4a40fa-bf76-11dc-8b1a-000a95edb42a Ancestors: Magritte-Scriptaculous-kph.16 added MAOneToManyDocumentsComponent (it doesnt actually use scriptaculous) From renggli at iam.unibe.ch Thu Jan 10 13:28:52 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 13:28:52 +0100 Subject: MADescription>>cssClass: In-Reply-To: References: Message-ID: > I don't see class="readonly info" in my page source code, only > class="readonly". The problem seems to be in > MAComponentRenderer>>classFor:, which ignores aDescription > attributes. Right? #cssClass: is added to the form element, not to the enclosing div. In the case of the readonly view the attributes are not used. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/d71b7cdc/smime.bin From renggli at iam.unibe.ch Thu Jan 10 13:43:14 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 13:43:14 +0100 Subject: Magritte examples in seaside/examples/examplebrowser In-Reply-To: <478506E9.9020109@online.de> References: <478506E9.9020109@online.de> Message-ID: <73D56F78-6420-403A-9D4B-B83CD627C98B@iam.unibe.ch> > trying to browse the Magritte examples in examplebrowser I found the > following > issues: > One is with > WAComponent class >> > headerForExampleBrowser > ^self name = self description > ifTrue: [ self name ] > ifFalse: [ self description , ' (' , self name , ')' ] > > The 'Comma' op is defined in MADescription and #, tries to send > #asContainer to > the following ByteStrings. An DNU Error is thrown. I am unable to reproduce? How did you trigger this DNU? What is the stack trace? > MAExampleEditor instancevariable 'description' is supposed to hold a > Collection > - at least it is used that way e.g.: > > MADescriptionEditor>> > buildReport > ^ (MAReport rows: self description > description: MAElementDescription description) > > however it is initialized with: > > defaultDescription > ^MAContainer new. > > I changed > MAExampleEditor>> > description > ^ description ifNil: [ > description := Array with: self defaultDescription ] > > (adding the Array constructor) and that works. I don't know if it > breaks other > things though... > Thanks for any suggestions and hints. MAContainer is like a collection of descripiton. What breaks with the current code? Why do you need to change it? Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/a32ffbd3/smime-0001.bin From keith_hodges at yahoo.co.uk Thu Jan 10 13:47:12 2008 From: keith_hodges at yahoo.co.uk (Keith Hodges) Date: Thu, 10 Jan 2008 12:47:12 +0000 Subject: Magritte-Scriptaculous Message-ID: <478613D0.8020204@yahoo.co.uk> Dear all, The magritte-scriptaculous and magritte-realmemento packages allow generated form components to dynamically change as the values or selections made by the user change. For example, the list of options in a SelectListComponent can be supplied through an accessor, and the visibility of a field can be determined by the value obtained via an accessor. descriptionMyChoice ^ MASingleOptionDescription new accessor: (#choice asAccessor) options: #choiceOptions; The accessor framework has been extended to provide a clean and easy to use means for specifying "alternate accessors for other purposes". Other purposes supported include: #visible: #readonly: #cssClass: #tooltip: #options: #update: The model class must specify a MARealMemento for this to work. The real memento uses a real copy of the model object as a memento, so that the memento model retains behaviour, particularly the interactions between attributes. This allows an #isMyFieldVisible to depend upon the value of a another field which may change. The form is re-rendered using scriptaculous updates, if the components request it. I have to say that what started out as a bit of an experiment has been surprisingly successful, and I think it represents a significant step up in the abilities of magritte. I am about to go live in the next couple of days on my first commercial project which uses this stuff. best regards Keith From ch.l.ngre at online.de Thu Jan 10 14:07:04 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Thu, 10 Jan 2008 14:07:04 +0100 Subject: Magritte examples in seaside/examples/examplebrowser In-Reply-To: <73D56F78-6420-403A-9D4B-B83CD627C98B@iam.unibe.ch> References: <478506E9.9020109@online.de> <73D56F78-6420-403A-9D4B-B83CD627C98B@iam.unibe.ch> Message-ID: <47861878.5060607@online.de> Lukas Renggli schrieb: >> trying to browse the Magritte examples in examplebrowser I found the >> following >> issues: >> One is with >> WAComponent class >> >> headerForExampleBrowser >> ^self name = self description >> ifTrue: [ self name ] >> ifFalse: [ self description , ' (' , self name , ')' ] >> >> The 'Comma' op is defined in MADescription and #, tries to send >> #asContainer to >> the following ByteStrings. An DNU Error is thrown. > > > I am unable to reproduce? How did you trigger this DNU? What is the > stack trace? Starting with Magritte-Seaside 248 (MAReport demo was dropped in the meantime :251 ??) opening http://localhost:8080/seaside/examples/examplebrowser First needs fixes in MADescriptionEditor>>buildReport to open without an error MADescriptionEditor>>buildReport ^ (MAReport rows: self description children "added children here" description: MAElementDescription description) more fixes changing 'report refresh' in add/edit/remove/Description MADescriptionEditor example works in latest version of Magritte-Seaside (251) ! Then choosing MAReport from the dropdown list gives me: Stack Trace 1. thisContext ByteString(Object)>>doesNotUnderstand: #asContainer self ' (' aMessage asContainer 2. thisContext MAPriorityContainer(MADescription)>>, self a MAPriorityContainer label: 'Report' comment: nil aDescription ' (' 3. thisContext MAReport class(WAComponent class)>>headerForExampleBrowser self MAReport 4. thisContext WAExampleBrowser>>renderComponentOn: self a WAExampleBrowser html a WARenderCanvas 5. thisContext WAExampleBrowser>>renderContentOn: self a WAExampleBrowser html a WARenderCanvas > THanks a lot for your tutorial! I'm squeak newbie and learned a lot! Christoph > ------------------------------------------------------------------------ > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From renggli at iam.unibe.ch Thu Jan 10 14:17:52 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 14:17:52 +0100 Subject: Magritte examples in seaside/examples/examplebrowser In-Reply-To: <47861878.5060607@online.de> References: <478506E9.9020109@online.de> <73D56F78-6420-403A-9D4B-B83CD627C98B@iam.unibe.ch> <47861878.5060607@online.de> Message-ID: > (MAReport demo was dropped in the meantime :251 ??) > opening http://localhost:8080/seaside/examples/examplebrowser Please use with Magritte-Model-lr.301 and Magritte-Seaside-lr.251. Is the dropped needed for the tutorial? Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/27500ba3/smime.bin From ch.l.ngre at online.de Thu Jan 10 15:28:37 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Thu, 10 Jan 2008 15:28:37 +0100 Subject: Magritte examples in seaside/examples/examplebrowser In-Reply-To: References: <478506E9.9020109@online.de> <73D56F78-6420-403A-9D4B-B83CD627C98B@iam.unibe.ch> <47861878.5060607@online.de> Message-ID: <47862B95.3020402@online.de> Lukas Renggli schrieb: >> (MAReport demo was dropped in the meantime :251 ??) >> opening http://localhost:8080/seaside/examples/examplebrowser > > > Please use with Magritte-Model-lr.301 and Magritte-Seaside-lr.251. > > Is the dropped needed for the tutorial? > > Lukas > Hi, no, it is not. I already switched to 251 and thought I posted that yesterday, but I must have made a mistake and the post didn't get through to the list. Thank you, Christoph From michal.perutka at gmail.com Thu Jan 10 15:48:33 2008 From: michal.perutka at gmail.com (Michal Perutka) Date: Thu, 10 Jan 2008 15:48:33 +0100 Subject: MADescription>>cssClass: In-Reply-To: References: Message-ID: It was an example only. Generally I have been searching how to add css classes to my decsriptions. Is #cssClass: not intended for it ? So in which way I should do it, please ? (for labels or controls in MATableRenderer...) Michal 2008/1/10, Lukas Renggli : > > I don't see class="readonly info" in my page source code, only > > class="readonly". The problem seems to be in > > MAComponentRenderer>>classFor:, which ignores aDescription > > attributes. Right? > > #cssClass: is added to the form element, not to the enclosing div. In > the case of the readonly view the attributes are not used. > > Cheers, > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki > > From renggli at iam.unibe.ch Thu Jan 10 16:27:12 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 16:27:12 +0100 Subject: MADescription>>cssClass: In-Reply-To: References: Message-ID: <4B55BFD6-48FA-413F-A89E-E07FD28EF5D7@iam.unibe.ch> > It was an example only. Generally I have been searching how to add css > classes to my decsriptions. Is #cssClass: not intended for it ? So in > which way I should do it, please ? (for labels or controls in > MATableRenderer...) Ok, that was a dark area in Magritte. This is fixed in: Name: Magritte-Seaside-lr.252 Author: lr Time: 10 January 2008, 4:26:27 pm UUID: 844438d1-df1f-491d-95bc-ddbee1ae6da1 Ancestors: Magritte-Seaside-lr.251 - remove #attributes from description (too fuzzy and too generic), but added #cssClasses that will be used by the updated renderer - fixed several issues with the table renderer that generated invalid html Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/016cbee1/smime-0001.bin From merlyn at stonehenge.com Thu Jan 10 22:48:11 2008 From: merlyn at stonehenge.com (Randal L. Schwartz) Date: Thu, 10 Jan 2008 13:48:11 -0800 Subject: DNU in my code from latest Magritte Message-ID: <86abnd8i90.fsf@blue.stonehenge.com> result := myModel asComponent. "returns an MAContainerComponent" result description attributes addClass: 'gcbpagemodel'. Apparently, MADescription no longer understands #attributes. what do I call #addClass: on now? -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training! From merlyn at stonehenge.com Thu Jan 10 22:53:59 2008 From: merlyn at stonehenge.com (Randal L. Schwartz) Date: Thu, 10 Jan 2008 13:53:59 -0800 Subject: DNU in my code from latest Magritte In-Reply-To: <86abnd8i90.fsf@blue.stonehenge.com> (Randal L. Schwartz's message of "Thu, 10 Jan 2008 13:48:11 -0800") References: <86abnd8i90.fsf@blue.stonehenge.com> Message-ID: <8663y18hzc.fsf@blue.stonehenge.com> >>>>> "Randal" == Randal L Schwartz writes: Randal> result := myModel asComponent. "returns an MAContainerComponent" Randal> result description attributes addClass: 'gcbpagemodel'. Randal> Apparently, MADescription no longer understands #attributes. Randal> what do I call #addClass: on now? And after commenting that out, I get a DNU here: MACssRenderer>>renderContainer: renderContainer: aDescription html definitionList attributes: aDescription attributes; "HERE" class: (self classFor: aDescription); with: [super renderContainer: aDescription] That's not my code. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training! From renggli at iam.unibe.ch Thu Jan 10 22:54:59 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 22:54:59 +0100 Subject: DNU in my code from latest Magritte In-Reply-To: <86abnd8i90.fsf@blue.stonehenge.com> References: <86abnd8i90.fsf@blue.stonehenge.com> Message-ID: > result := myModel asComponent. "returns an MAContainerComponent" > result description attributes addClass: 'gcbpagemodel'. > > Apparently, MADescription no longer understands #attributes. > > what do I call #addClass: on now? As before, you just use #cssClass: directly on the description. #attributes didn't really belong into the description anyway. I think like this it is much cleaner. Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/faa5e032/smime.bin From renggli at iam.unibe.ch Thu Jan 10 23:04:23 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 10 Jan 2008 23:04:23 +0100 Subject: DNU in my code from latest Magritte In-Reply-To: <8663y18hzc.fsf@blue.stonehenge.com> References: <86abnd8i90.fsf@blue.stonehenge.com> <8663y18hzc.fsf@blue.stonehenge.com> Message-ID: >>>>>> "Randal" == Randal L Schwartz writes: > > Randal> result := myModel asComponent. "returns an > MAContainerComponent" > Randal> result description attributes addClass: 'gcbpagemodel'. > > Randal> Apparently, MADescription no longer understands #attributes. > > Randal> what do I call #addClass: on now? > > And after commenting that out, I get a DNU here: > > MACssRenderer>>renderContainer: > > renderContainer: aDescription > html definitionList > attributes: aDescription attributes; "HERE" > class: (self classFor: aDescription); > with: [super renderContainer: aDescription] > > That's not my code. I know, I am responsible for all bugs you might encounter. Name: Magritte-All-lr.258 Author: lr Time: 10 January 2008, 11:03:26 pm UUID: a708253b-3616-4fb4-9107-d17b2d6b37fd Ancestors: Magritte-All-lr.257 Dependencies: Magritte-Model-lr.303, Magritte-Tests-lr.131, Magritte- Seaside-lr.253, Magritte-Morph-tg.39 - remove #attributes send in MACssRenderer Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080110/1421643d/smime.bin From merlyn at stonehenge.com Thu Jan 10 23:12:50 2008 From: merlyn at stonehenge.com (Randal L. Schwartz) Date: Thu, 10 Jan 2008 14:12:50 -0800 Subject: DNU in my code from latest Magritte In-Reply-To: (Lukas Renggli's message of "Thu, 10 Jan 2008 23:04:23 +0100") References: <86abnd8i90.fsf@blue.stonehenge.com> <8663y18hzc.fsf@blue.stonehenge.com> Message-ID: <86wsqh72jh.fsf@blue.stonehenge.com> >>>>> "Lukas" == Lukas Renggli writes: Lukas> I know, I am responsible for all bugs you might encounter. Lukas> Name: Magritte-All-lr.258 Lukas> Author: lr Lukas> Time: 10 January 2008, 11:03:26 pm Lukas> UUID: a708253b-3616-4fb4-9107-d17b2d6b37fd Lukas> Ancestors: Magritte-All-lr.257 Lukas> Dependencies: Magritte-Model-lr.303, Magritte-Tests-lr.131, Magritte- Lukas> Seaside-lr.253, Magritte-Morph-tg.39 Lukas> - remove #attributes send in MACssRenderer So, I see Magritte-Model updated I presume a few hours ago, which I loaded, and got my DNU. What is Magritte-All? Does it also contain code? Should I be loading it instead? -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training! From girba at iam.unibe.ch Sat Jan 12 23:02:24 2008 From: girba at iam.unibe.ch (Tudor Girba) Date: Sat, 12 Jan 2008 23:02:24 +0100 Subject: css problem with firefox Message-ID: <9FE64DA5-3380-486E-9CD7-C1BB00C77145@iam.unibe.ch> Using the latest version of Pier I did the followings: - added a file that I named "css" without any content - make the stylesheet of the root page to point to the css file - edit design - added "ul{font-size: 20pt}" to the stylesheet - on Safari and Opera it works Ok - on Firefox there is no effect I believe there should be a problem in how the css is linked with the page, but I have no idea what it actually is. Any hints? Cheers, Doru -- www.tudorgirba.com www.tudorgirba.com/blog "In a world where everything is moving ever faster, one might have better chances to win by moving slower." From ch.l.ngre at online.de Sun Jan 13 23:04:14 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Sun, 13 Jan 2008 23:04:14 +0100 Subject: Magritte Component in SULightbox Message-ID: <478A8ADE.6040401@online.de> Hi, I am trying to show a Magritte Component in a SULightbox. I replaced #call: with #lightbox: and I'm getting the form in a Lightbox, but it looks like some sort of container is missing... I see many small lightboxes - one for each text-entry / button etc. Labels are rendered on the darkened background. Scriptaculous-lr.238 Magritte-Seaside-lr.251 Thank you, Christoph From renggli at iam.unibe.ch Mon Jan 14 16:39:17 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 14 Jan 2008 16:39:17 +0100 Subject: css problem with firefox In-Reply-To: <9FE64DA5-3380-486E-9CD7-C1BB00C77145@iam.unibe.ch> References: <9FE64DA5-3380-486E-9CD7-C1BB00C77145@iam.unibe.ch> Message-ID: <4FF0410C-57AE-4625-8DB7-4D0A58631F6A@iam.unibe.ch> Hi Doru, thanks for reporting. This is fixed in Pier-Seaside-lr.248. Cu, Lukas On Jan 12, 2008, at 23:02 , Tudor Girba wrote: > Using the latest version of Pier I did the followings: > - added a file that I named "css" without any content > - make the stylesheet of the root page to point to the css file > - edit design > - added "ul{font-size: 20pt}" to the stylesheet > - on Safari and Opera it works Ok > - on Firefox there is no effect > > I believe there should be a problem in how the css is linked with the > page, but I have no idea what it actually is. Any hints? > > Cheers, > Doru > > -- > www.tudorgirba.com > www.tudorgirba.com/blog > > "In a world where everything is moving ever faster, > one might have better chances to win by moving slower." > > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/3f626356/smime.bin From girba at iam.unibe.ch Mon Jan 14 16:53:15 2008 From: girba at iam.unibe.ch (Tudor Girba) Date: Mon, 14 Jan 2008 16:53:15 +0100 Subject: css problem with firefox In-Reply-To: <4FF0410C-57AE-4625-8DB7-4D0A58631F6A@iam.unibe.ch> References: <9FE64DA5-3380-486E-9CD7-C1BB00C77145@iam.unibe.ch> <4FF0410C-57AE-4625-8DB7-4D0A58631F6A@iam.unibe.ch> Message-ID: <04A30054-5CC4-4A4A-B874-E9C428BD2213@iam.unibe.ch> Thanks. CU :), Doru On Jan 14, 2008, at 4:39 PM, Lukas Renggli wrote: > Hi Doru, > > thanks for reporting. This is fixed in Pier-Seaside-lr.248. > > Cu, > Lukas > > On Jan 12, 2008, at 23:02 , Tudor Girba wrote: > >> Using the latest version of Pier I did the followings: >> - added a file that I named "css" without any content >> - make the stylesheet of the root page to point to the css file >> - edit design >> - added "ul{font-size: 20pt}" to the stylesheet >> - on Safari and Opera it works Ok >> - on Firefox there is no effect >> >> I believe there should be a problem in how the css is linked with the >> page, but I have no idea what it actually is. Any hints? >> >> Cheers, >> Doru >> >> -- >> www.tudorgirba.com >> www.tudorgirba.com/blog >> >> "In a world where everything is moving ever faster, >> one might have better chances to win by moving slower." >> >> >> >> _______________________________________________ >> SmallWiki, Magritte, Pier and Related Tools ... >> https://www.iam.unibe.ch/mailman/listinfo/smallwiki > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- www.tudorgirba.com www.tudorgirba.com/blog "No matter how many recipes we'll know, we'll still value a chef." From michal.perutka at gmail.com Mon Jan 14 16:58:19 2008 From: michal.perutka at gmail.com (Michal Perutka) Date: Mon, 14 Jan 2008 16:58:19 +0100 Subject: MADescription>>cssClass: In-Reply-To: <4B55BFD6-48FA-413F-A89E-E07FD28EF5D7@iam.unibe.ch> References: <4B55BFD6-48FA-413F-A89E-E07FD28EF5D7@iam.unibe.ch> Message-ID: > > Ok, that was a dark area in Magritte. This is fixed in: > > Name: Magritte-Seaside-lr.252 > Author: lr > Time: 10 January 2008, 4:26:27 pm > UUID: 844438d1-df1f-491d-95bc-ddbee1ae6da1 > Ancestors: Magritte-Seaside-lr.251 > > - remove #attributes from description (too fuzzy and too generic), but > added #cssClasses that will be used by the updated renderer > Oh, quite a radical fix ;-) I had to redesign my app as I was using attributes frequently. But, no matter, it's already done. Cheers, Michal -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/03bab2f0/attachment.html From renggli at iam.unibe.ch Mon Jan 14 17:11:13 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 14 Jan 2008 17:11:13 +0100 Subject: MADescription>>cssClass: In-Reply-To: References: <4B55BFD6-48FA-413F-A89E-E07FD28EF5D7@iam.unibe.ch> Message-ID: > Oh, quite a radical fix ;-) I had to redesign my app as I was using > attributes frequently. But, no matter, it's already done. Sorry for the trouble. I hope you can still archive the same effects with #cssClass: ? I think this is much cleaner from a mode-view separation perspective. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/5efacaeb/smime.bin From renggli at iam.unibe.ch Mon Jan 14 17:19:09 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 14 Jan 2008 17:19:09 +0100 Subject: Magritte Component in SULightbox In-Reply-To: <478A8ADE.6040401@online.de> References: <478A8ADE.6040401@online.de> Message-ID: <385E5900-FD1C-4EC1-B22B-8F37D416AC7E@iam.unibe.ch> You have to provide some stylesheet for your lightbox. The included style only provides a minimum functionality. Try to add something like: #lightbox { padding: 2em; background-color: #eee; } Cheers, Lukas On Jan 13, 2008, at 23:04 , Ch Lamprecht wrote: > Hi, > > I am trying to show a Magritte Component in a SULightbox. > > I replaced #call: with #lightbox: and I'm getting the form in a > Lightbox, but it looks like some sort of container is missing... > I see many small lightboxes - one for each text-entry / button etc. > Labels are rendered on the darkened background. > > Scriptaculous-lr.238 > Magritte-Seaside-lr.251 > > Thank you, Christoph > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/90bd713e/smime-0001.bin From renggli at iam.unibe.ch Mon Jan 14 17:21:57 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 14 Jan 2008 17:21:57 +0100 Subject: Pier Sites Message-ID: Hi, I am looking for Pier instances, so that I can add them to the list of online sites on the official Pier web-site. Currently this is http://lukas-renggli.ch/smalltalk/pier , but will maybe move to its own domain later on. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/a611f6ef/smime.bin From ch.l.ngre at online.de Mon Jan 14 17:43:59 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Mon, 14 Jan 2008 17:43:59 +0100 Subject: Magritte Component in SULightbox In-Reply-To: <385E5900-FD1C-4EC1-B22B-8F37D416AC7E@iam.unibe.ch> References: <478A8ADE.6040401@online.de> <385E5900-FD1C-4EC1-B22B-8F37D416AC7E@iam.unibe.ch> Message-ID: <478B914F.5040108@online.de> Lukas Renggli wrote: > You have to provide some stylesheet for your lightbox. The included > style only provides a minimum functionality. Try to add something like: > > #lightbox { > padding: 2em; > background-color: #eee; > } Thank you, Christoph From merlyn at stonehenge.com Mon Jan 14 18:20:41 2008 From: merlyn at stonehenge.com (Randal L. Schwartz) Date: Mon, 14 Jan 2008 09:20:41 -0800 Subject: MADescription>>cssClass: In-Reply-To: (Lukas Renggli's message of "Mon, 14 Jan 2008 17:11:13 +0100") References: <4B55BFD6-48FA-413F-A89E-E07FD28EF5D7@iam.unibe.ch> Message-ID: <86bq7omihi.fsf@blue.stonehenge.com> >>>>> "Lukas" == Lukas Renggli writes: Lukas> I hope you can still archive the same effects with #cssClass: ? I think Lukas> this is much cleaner from a mode-view separation perspective. Yeah, it did seem a bit odd that I was asking a description to know what its CSS class was. :) -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training! From ducasse at iam.unibe.ch Mon Jan 14 19:51:35 2008 From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=) Date: Mon, 14 Jan 2008 19:51:35 +0100 Subject: Pier Sites In-Reply-To: References: Message-ID: <178A45CA-B405-4918-9879-4AF4E568DF44@iam.unibe.ch> thinlizzy ? the one where avi helped alan ? Stef On Jan 14, 2008, at 5:21 PM, Lukas Renggli wrote: > Hi, > > I am looking for Pier instances, so that I can add them to the list > of online sites on the official Pier web-site. Currently this is http://lukas-renggli.ch/smalltalk/pier > , but will maybe move to its own domain later on. > > Cheers, > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From renggli at iam.unibe.ch Mon Jan 14 20:21:23 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Mon, 14 Jan 2008 20:21:23 +0100 Subject: Pier Sites In-Reply-To: <178A45CA-B405-4918-9879-4AF4E568DF44@iam.unibe.ch> References: <178A45CA-B405-4918-9879-4AF4E568DF44@iam.unibe.ch> Message-ID: <95AE154F-A5E0-48C4-8450-82E97F8676C1@iam.unibe.ch> > thinlizzy ? Where is that? > the one where avi helped alan ? The LogoWiki is not online amore, I am afraid. It is a free Pier extension though, so anybody can turn his Pier instance into a LogoWiki by loading the package from http://source.lukas-renggli.ch/pieraddons . Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080114/c93c8d83/smime.bin From ducasse at iam.unibe.ch Tue Jan 15 20:30:24 2008 From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=) Date: Tue, 15 Jan 2008 20:30:24 +0100 Subject: Pier Sites In-Reply-To: <95AE154F-A5E0-48C4-8450-82E97F8676C1@iam.unibe.ch> References: <178A45CA-B405-4918-9879-4AF4E568DF44@iam.unibe.ch> <95AE154F-A5E0-48C4-8450-82E97F8676C1@iam.unibe.ch> Message-ID: <84DCC5F8-4A9A-42DF-840F-51533BB2E735@iam.unibe.ch> may be ask kim rose or scott because I cannot find it. stef On Jan 14, 2008, at 8:21 PM, Lukas Renggli wrote: >> thinlizzy ? > > Where is that? > >> the one where avi helped alan ? > > The LogoWiki is not online amore, I am afraid. It is a free Pier > extension though, so anybody can turn his Pier instance into a > LogoWiki by loading the package from http://source.lukas-renggli.ch/pieraddons > . > > Cheers, > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From ducasse at iam.unibe.ch Tue Jan 15 20:34:35 2008 From: ducasse at iam.unibe.ch (=?ISO-8859-1?Q?st=E9phane_ducasse?=) Date: Tue, 15 Jan 2008 20:34:35 +0100 Subject: Pier Sites In-Reply-To: <84DCC5F8-4A9A-42DF-840F-51533BB2E735@iam.unibe.ch> References: <178A45CA-B405-4918-9879-4AF4E568DF44@iam.unibe.ch> <95AE154F-A5E0-48C4-8450-82E97F8676C1@iam.unibe.ch> <84DCC5F8-4A9A-42DF-840F-51533BB2E735@iam.unibe.ch> Message-ID: may be this what they called LogoWiki I was never clear on that On Jan 15, 2008, at 8:30 PM, st?phane ducasse wrote: > may be ask kim rose or scott because I cannot find it. > > stef > On Jan 14, 2008, at 8:21 PM, Lukas Renggli wrote: > >>> thinlizzy ? >> >> Where is that? >> >>> the one where avi helped alan ? >> >> The LogoWiki is not online amore, I am afraid. It is a free Pier >> extension though, so anybody can turn his Pier instance into a >> LogoWiki by loading the package from http://source.lukas-renggli.ch/pieraddons >> . >> >> Cheers, >> Lukas >> >> -- >> Lukas Renggli >> http://www.lukas-renggli.ch >> >> >> _______________________________________________ >> SmallWiki, Magritte, Pier and Related Tools ... >> https://www.iam.unibe.ch/mailman/listinfo/smallwiki > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki From kbrown at mac.com Tue Jan 15 21:50:56 2008 From: kbrown at mac.com (Ken G. Brown) Date: Tue, 15 Jan 2008 13:50:56 -0700 Subject: Pier Sites In-Reply-To: References: Message-ID: Maybe these links will help find what you want regarding tinlizzie: Ken G. Brown At 8:34 PM +0100 1/15/08, smallwiki-request at iam.unibe.ch apparently wrote: >Date: Tue, 15 Jan 2008 20:34:35 +0100 >From: st?phane ducasse >Subject: Re: Pier Sites >To: "Magritte, Pier and Related Tools ..." >Message-ID: >Content-Type: text/plain; charset=ISO-8859-1; format=flowed > >may be this what they called LogoWiki > >I was never clear on that > >On Jan 15, 2008, at 8:30 PM, st?phane ducasse wrote: > >> may be ask kim rose or scott because I cannot find it. >> > > stef > > On Jan 14, 2008, at 8:21 PM, Lukas Renggli wrote: > > > >>> thinlizzy ? > >> > >> Where is that? From ch.l.ngre at online.de Thu Jan 17 08:47:10 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Thu, 17 Jan 2008 08:47:10 +0100 Subject: MAMultilpleOption vs. MAToManyRelation Message-ID: <478F07FE.5050508@online.de> Hi, if an Object holds a collection of other Objects (like a group of Persons), and these (Persons) should not be created/edited but a list to choose from should be presented - would a MAMultipleOptionDescription be appropriate? If so, how can the options presented be made dynamic. - I found, that they are getting cached in the Default DescriptionBuilder. Or are there options to pass to a MAToManyRelationDescription, so that it does not show editors, but choices instead? Thanks, Christoph From renggli at iam.unibe.ch Thu Jan 17 09:55:17 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 17 Jan 2008 09:55:17 +0100 Subject: MAMultilpleOption vs. MAToManyRelation In-Reply-To: <478F07FE.5050508@online.de> References: <478F07FE.5050508@online.de> Message-ID: <948DE63B-743C-4845-BE9A-0396CB14673C@iam.unibe.ch> > if an Object holds a collection of other Objects (like a group of > Persons), and > these (Persons) should not be created/edited but a list to choose > from should be > presented - would a MAMultipleOptionDescription be appropriate? If > so, how can > the options presented be made dynamic. There are several ways to make properties of descriptions dynamic. 1. Override #description on the instance-side and build your descriptions programmatically. Don't forget to copy the descriptions, in case you call super to get modify the default behavior. 2. Pass the property a block that you send #asDynamicObject. See the senders of this message for examples. This approach is simple and works in most cases, there are some possible drawbacks however: (1) the implementation uses a proxy object that might not get properly resolved in all cases, (2) the proxy object references a block context that might not be serializeable with your preferred persistency mechanism, (3) proxies are hard to debug, and (4) from within the block you don't have access to the described object. PS: I started a Magritte FAQ at http://www.lukas-renggli.ch/smalltalk/magritte/faq . Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 3774 bytes Desc: not available Url : http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080117/62ffea90/smime.bin From cdrick65 at gmail.com Thu Jan 17 15:04:11 2008 From: cdrick65 at gmail.com (cdrick) Date: Thu, 17 Jan 2008 15:04:11 +0100 Subject: how to access/view hidden structure Message-ID: Hi Is there a way to access hidden structure (if checked in the settings). Until now, I'm doing internal links, but if I'm sure I have some unused hidden structure... Also, what's your way to make ling to internal struct ? I only use the wiki syntax. Would you find useful to have another category (structure) when we make a link that display the full tree of struc (the one you're authorized to view and the hidden ones) ? Maybe such a thing already exists... Thanks C?drick -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080117/ba5d81c9/attachment.html From cdrick65 at gmail.com Thu Jan 17 23:20:17 2008 From: cdrick65 at gmail.com (cdrick) Date: Thu, 17 Jan 2008 23:20:17 +0100 Subject: how to access/view hidden structure In-Reply-To: References: Message-ID: no idea ? I rephrase. How can I find pages...structures that are hidden from the menu in pier ? Thanks -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080117/b6f11d37/attachment.html From ch.l.ngre at online.de Sat Jan 19 22:47:13 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Sat, 19 Jan 2008 22:47:13 +0100 Subject: MAMultilpleOption vs. MAToManyRelation In-Reply-To: <948DE63B-743C-4845-BE9A-0396CB14673C@iam.unibe.ch> References: <478F07FE.5050508@online.de> <948DE63B-743C-4845-BE9A-0396CB14673C@iam.unibe.ch> Message-ID: <47926FE1.6020104@online.de> Lukas Renggli schrieb: >> if an Object holds a collection of other Objects (like a group of >> Persons), and >> these (Persons) should not be created/edited but a list to choose >> from should be >> presented - would a MAMultipleOptionDescription be appropriate? If >> so, how can >> the options presented be made dynamic. > > > There are several ways to make properties of descriptions dynamic. > > 1. Override #description on the instance-side and build your > descriptions programmatically. Don't forget to copy the descriptions, > in case you call super to get modify the default behavior. > > 2. Pass the property a block that you send #asDynamicObject. See the > senders of this message for examples. This approach is simple and works > in most cases, there are some possible drawbacks however: (1) the > implementation uses a proxy object that might not get properly resolved > in all cases, (2) the proxy object references a block context that > might not be serializeable with your preferred persistency mechanism, > (3) proxies are hard to debug, and (4) from within the block you don't > have access to the described object. Hi, thanks a lot for the suggestions! I tried some things and would like to share my experiences here: [1.] was obvious (after having worked through your exercises at least) but not very elegant. I started experimenting with asDynamicObject but in case of MASingleOptionDescription >> optionsAndLabels: it was not of any use, as the Collection passed in gets copied, and so the proxy object is lost. Now simply trying to use MASingleOptionDescription >> options: (providing the label through model >> printString) raised other questions: The defaultReference used by the MAOptionDescriptions is a MAStringDescription which tells a visitor to use the visitStringDescription: method. In case of the MAStringWriter this is implemented as: MaStringWriter >> visitStringDescription: aDescription self stream nextPutAll: self object This obviously fails in cases where object is no String. visitElementDescription adds asString which works : MaStringWriter >> visitElementDescription: aDescription self stream nextPutAll: self object asString MAElementDescription however isAbstract and therefore can't be used for MASingleOptionDescription >> reference: I could work around by creating a new description class and corresponding MAStringWriter subclass then setting the reference attribute to it: MASingleOptionDescription new label: 'Label' ; reference: MAMyStringDescription new; selectorAccessor: #myAccessor; options: [WAPersonManager persons]asDynamicObject; Then MAPersonModel >> printString returns lastName or something useful. That way everything works fine and the choices are dynamic (hitting reload in the browser updates the list) To make this work out of the box, maybe 'asString' should be added to MaStringWriter >> visitStringDescription: - Or a better defaultReference could be choosen for MAOptionDescription subclasses. I hope this made sense... Cheers, Christoph From ch.l.ngre at online.de Mon Jan 21 12:29:04 2008 From: ch.l.ngre at online.de (Ch Lamprecht) Date: Mon, 21 Jan 2008 12:29:04 +0100 Subject: MAMultilpleOption vs. MAToManyRelation In-Reply-To: <47926FE1.6020104@online.de> References: <478F07FE.5050508@online.de> <948DE63B-743C-4845-BE9A-0396CB14673C@iam.unibe.ch> <47926FE1.6020104@online.de> Message-ID: <47948200.4090104@online.de> Ch Lamprecht wrote: > Lukas Renggli schrieb: > >>>if an Object holds a collection of other Objects (like a group of >>>Persons), and >>>these (Persons) should not be created/edited but a list to choose >>>from should be >>>presented - would a MAMultipleOptionDescription be appropriate? If >>>so, how can >>>the options presented be made dynamic. >> >> >>There are several ways to make properties of descriptions dynamic. >> >>1. Override #description on the instance-side and build your >>descriptions programmatically. Don't forget to copy the descriptions, >>in case you call super to get modify the default behavior. >> >>2. Pass the property a block that you send #asDynamicObject. See the >>senders of this message for examples. This approach is simple and works >>in most cases, there are some possible drawbacks however: (1) the >>implementation uses a proxy object that might not get properly resolved >>in all cases, (2) the proxy object references a block context that >>might not be serializeable with your preferred persistency mechanism, >>(3) proxies are hard to debug, and (4) from within the block you don't >>have access to the described object. > > > > I could work around by creating a new description class and corresponding > MAStringWriter subclass then setting the reference attribute to it: > > MASingleOptionDescription new > label: 'Label' ; > reference: MAMyStringDescription new; > selectorAccessor: #myAccessor; > options: [WAPersonManager persons]asDynamicObject; > > Then MAPersonModel >> printString returns lastName or something useful. > > That way everything works fine and the choices are dynamic (hitting reload in > the browser updates the list) > > I hope this made sense... Cheers, Christoph Hmm, maybe it did not... Is #reference: meant for 'public' use with MA*OptionDescription ? Thank you, Christoph From jason.johnson.081 at gmail.com Wed Jan 23 08:19:17 2008 From: jason.johnson.081 at gmail.com (Jason Johnson) Date: Wed, 23 Jan 2008 08:19:17 +0100 Subject: Strange behavior with MAToManyRelationDescription In-Reply-To: <3FC3F4A7-B28E-4D7D-8E74-7E2252928729@iam.unibe.ch> References: <3FC3F4A7-B28E-4D7D-8E74-7E2252928729@iam.unibe.ch> Message-ID: Thanks for the response. Sorry I was so long in getting back to this. December is a pretty insane month for me at the moment. :) (comments below) On Nov 29, 2007 8:00 AM, Lukas Renggli wrote: > > > This is where I need the custom memento (and tell me if I'm doing this > > wrong): If you click save that winds up calling "commit" on the > > memento and actually writing the data to the object, but this new > > object wont be registered in the class side collection (and therefor > > will simply cease to exist when this session expires). So my custom > > memento simply lets everything work as normal and then calls "Person > > register: self model" to get the new object registered. > > Sure that's a possibility, but why not use the return value if you are > using #call: or set an #onAnswer: handler if you are using composition? Well, the thing is, I'm using a shortcut here. :) The manager just basically provides a web page that allows one to use the descriptions of other two. It doesn't use call because it doesn't want anything back. Maybe an example will make it clearer. Manager>>renderContentOn: html html render: self component >>component ^ component ifNil: [ component := ((self housesComponent, self peopleComponent) asComponentOn: self) addForm: #(); "I don't want any buttons on this form since submitting it would make no sense, I do nothing with the values here" yourself ] "since the House and Person classes manage their own instances" >>housesComponent "gives a nice view for managing houses" ^ MAToManyRelationDescription new classes: (Array with: House); label: 'Houses'; accessor: (MASelectorAccessor new readSelector: #houses; writeSelector: nil; yourself); default: Array new; priority: 20; yourself >>peopleComponent "ditto for people" ^ MAToManyRelationDescription new classes: (Array with: Person); label: 'People'; accessor: (MASelectorAccessor new readSelector: #people; writeSelector: nil; yourself); default: Array new; priority: 20; yourself >>houses ^ House instances >>people ^ Person instances > So in your manager you have something along: > > house := self call: House new asComponent. > house ifNotNil: [ House register: house ]. > > or (in case of composition): > > editor := House new asComponent onAnswer: [ :house | > house ifNotNil: [ House register: house ]. > editor := nil ] This I tried. It doesn't work when I put it on #houseComponent or #storeComponent since I'm using #, and when I put it in #component it seems to be ignored. > > The problem happens when "save" returns on Person and you go back to > > the manager page. You see your nice new Person in the report table as > > expected, but right below is the same Person again! And if you edit > > either one the changes reflect in both. And the really hairy thing > > is, even though House is almost exactly the same in every way (I can > > find no relevant difference) it doesn't have this behavior. > > Looks like your memento method gets called twice? Why not check if the > object is already present? Why not use a Set then? Ah but I do. Since I'm in the momento I have to check if there already exists an object that has the same value as the ones cached in the momento. I had a #break call in the momento and I'm sure it's not getting called twice, which wouldn't hurt anything if it did. It seems like the MA machinery is doing some caching of some kind. The part that baffles me is that it doesn't happen for Houses, even though they are fundamentally exactly the same. I suppose this would be so much easier if I just let Magritte manage everything completely, but this code gets managed in other places then on these web pages. > > > Cheers, > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki > From renggli at iam.unibe.ch Thu Jan 24 12:59:18 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 24 Jan 2008 12:59:18 +0100 Subject: MAMultilpleOption vs. MAToManyRelation In-Reply-To: <47926FE1.6020104@online.de> References: <478F07FE.5050508@online.de> <948DE63B-743C-4845-BE9A-0396CB14673C@iam.unibe.ch> <47926FE1.6020104@online.de> Message-ID: > I could work around by creating a new description class and > corresponding > MAStringWriter subclass then setting the reference attribute to it: > > MASingleOptionDescription new > label: 'Label' ; > reference: MAMyStringDescription new; > selectorAccessor: #myAccessor; > options: [WAPersonManager persons]asDynamicObject; What I would use here is the description of persons, given that the options are instances of Person: ... reference: Person description; ... > To make this work out of the box, maybe 'asString' should be added to > MaStringWriter >> visitStringDescription: I agree, the latest commit includes your suggestion. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Thu Jan 24 12:59:24 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Thu, 24 Jan 2008 12:59:24 +0100 Subject: how to access/view hidden structure In-Reply-To: References: Message-ID: <4DC71956-CF42-4136-9C28-9C7F70DEB1BE@iam.unibe.ch> > I rephrase. How can I find pages...structures that are hidden from > the menu in pier ? This is one of the rare cases where REST can be useful, but only if you know the URL by heart. Another possibility is to open the 'Browse' view on the parent structure. This view is part of the Unix Security plugin, it should probably move into the core package. Yet another possibility is to use OmniBrowser for Pier. Or as a last resort the inspector. That's not really a solution, though. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From tjohnson at iwu.edu Fri Jan 25 02:11:17 2008 From: tjohnson at iwu.edu (Mr. Timothy C. Johnson) Date: Thu, 24 Jan 2008 19:11:17 -0600 (CST) Subject: PRPierLibrary gone? Message-ID: <901634.1201223477842.JavaMail.tjohnson@iwu.edu> Hello, I installed all the latest packages from Monticello into a Seaside-mini 2.8 image. Everything was working well, except I had no image files. Trying to load Pier-Styles-Kph.2.mcz results in the following dependency missing: PRPierLibrary. I thought that would be available from Pier-Seaside but it is not, as of Pier-Seaside-lr.251.mcz. What can I do? Something seems to already work because when I go to 'http://localhost:8080/seaside/files/PRBlueprintLibrary/pier.png' it loads and displays just fine. Thanks, Tim From renggli at iam.unibe.ch Fri Jan 25 08:25:52 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Fri, 25 Jan 2008 08:25:52 +0100 Subject: PRPierLibrary gone? In-Reply-To: <901634.1201223477842.JavaMail.tjohnson@iwu.edu> References: <901634.1201223477842.JavaMail.tjohnson@iwu.edu> Message-ID: <38ECCB90-1479-47EF-951F-50787BB6BD11@iam.unibe.ch> Please you the latest version from Package Universe or Squeak Map. As it says in the Monticello commit comments, the very latest versions on my Squeak Source are in the middle of a transition and break certain things. Thanks, Lukas On Jan 25, 2008, at 02:11 , Mr. Timothy C. Johnson wrote: > Hello, > > I installed all the latest packages from Monticello into a Seaside- > mini > 2.8 image. Everything was working well, except I had no image files. > Trying to load Pier-Styles-Kph.2.mcz results in the following > dependency missing: PRPierLibrary. I thought that would be available > from Pier-Seaside but it is not, as of Pier-Seaside-lr.251.mcz. > > What can I do? > > Something seems to already work because when I go to > 'http://localhost:8080/seaside/files/PRBlueprintLibrary/pier.png' it > loads and displays just fine. > > Thanks, > Tim > > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch From tjohnson at iwu.edu Fri Jan 25 16:25:39 2008 From: tjohnson at iwu.edu (Tim Johnson) Date: Fri, 25 Jan 2008 09:25:39 -0600 Subject: PRPierLibrary gone? In-Reply-To: <38ECCB90-1479-47EF-951F-50787BB6BD11@iam.unibe.ch> Message-ID: Hi Lukas, Thank you. I suspected that. Unfortunately there is no SqueakMap or Package Universe in the images supplied for download on Seaside's web site (seaside-mini). My efforts with 3.8-6665 the night before resulted in my not being able to connect to localhost. I suspect I needed to do a WARegistry>>clearAllHandlers or something, but I could not find this anywhere in the docs for Pier or through vague Google searches so I quit out of frustration and threw away my work. Will try again with 3.8. Of course 3.7 won't work because SqueakMap can't update from 2.0 to 2.2. And 3.9... I just don't like. Maybe I *have* to use it? I know I had Pier working a few months ago but I can't remember which image I was using. Oh, and somewhere in your docs or on the list someone states that Damien Cassou's squeak-dev and squeak-web images are not adequate for Pier/Magritte because they are missing things or aren't stable yet. Is that not true anymore? Somewhere else it says people *can* use these. I am very confused. I just want to run Squeak & Komanche & Seaside & Pier in an attempt to set up a wiki for the computer labs I run... I had it working in October but ran out of time and now I'm picking it up again and can't get it to work. Again, thanks! - Tim On 1/25/08 1:25 AM, "Lukas Renggli" wrote: > Please you the latest version from Package Universe or Squeak Map. > > As it says in the Monticello commit comments, the very latest versions > on my Squeak Source are in the middle of a transition and break > certain things. > > Thanks, > Lukas > > On Jan 25, 2008, at 02:11 , Mr. Timothy C. Johnson wrote: > >> Hello, >> >> I installed all the latest packages from Monticello into a Seaside- >> mini >> 2.8 image. Everything was working well, except I had no image files. >> Trying to load Pier-Styles-Kph.2.mcz results in the following >> dependency missing: PRPierLibrary. I thought that would be available >> from Pier-Seaside but it is not, as of Pier-Seaside-lr.251.mcz. >> >> What can I do? >> >> Something seems to already work because when I go to >> 'http://localhost:8080/seaside/files/PRBlueprintLibrary/pier.png' it >> loads and displays just fine. >> >> Thanks, >> Tim >> >> >> >> _______________________________________________ >> SmallWiki, Magritte, Pier and Related Tools ... >> https://www.iam.unibe.ch/mailman/listinfo/smallwiki From firerox at centrum.cz Fri Jan 25 18:25:03 2008 From: firerox at centrum.cz (firerox@centrum.cz) Date: Fri, 25 Jan 2008 18:25:03 +0100 Subject: pier Message-ID: <200801251825.20680@centrum.cz> Hi all, I want to write to my textArea for example ""hi all"". And then I would like to see hi all in bold exactly like in WikiSystem. I used PRDocumentParser parse: '""hi all""', but this method returns PRDocument. But I want reneder this string in renederContentOn: method Please help me! Best regards Marek Fiala From renggli at iam.unibe.ch Sat Jan 26 08:06:56 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sat, 26 Jan 2008 08:06:56 +0100 Subject: PRPierLibrary gone? In-Reply-To: References: Message-ID: <0AB30310-964B-4A3F-A2F7-8DB46752C1DA@iam.unibe.ch> > Thank you. I suspected that. Unfortunately there is no SqueakMap or > Package Universe in the images supplied for download on Seaside's > web site (seaside-mini). Ok, I understand. Please try a slightly older version then. For example Pier-All-lr.271 and its dependencies. > My efforts with 3.8-6665 the night before resulted in my not being > able to > connect to localhost. I suspect I needed to do a > WARegistry>>clearAllHandlers or something, but I could not find this > anywhere in the docs for Pier or through vague Google searches so I > quit out > of frustration and threw away my work. That cleans the session cache. What you probably want is WADispatcher resetAll that re-initializes the applications. > ry again with 3.8. Of course 3.7 won't work because SqueakMap can't > update from 2.0 to 2.2. And 3.9... I just don't like. Maybe I > *have* to > use it? I know I had Pier working a few months ago but I can't > remember > which image I was using. I am using 3.9 for all my development and deployed work. > Oh, and somewhere in your docs or on the list someone states that > Damien > Cassou's squeak-dev and squeak-web images are not adequate for Pier/ > Magritte > because they are missing things or aren't stable yet. Is that not > true > anymore? Somewhere else it says people *can* use these. This has been fixed a long time ago. The images of Damien use the latest version from Package Universe. > I am very confused. I just want to run Squeak & Komanche & Seaside > & Pier > in an attempt to set up a wiki for the computer labs I run... I had > it > working in October but ran out of time and now I'm picking it up > again and > can't get it to work. Take the Web image of Damien. It has everything loaded. Lukas -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Sat Jan 26 08:15:05 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sat, 26 Jan 2008 08:15:05 +0100 Subject: pier In-Reply-To: <200801251825.20680@centrum.cz> References: <200801251825.20680@centrum.cz> Message-ID: <8DC36D36-063A-4B98-912B-8EF4FD3F5FC9@iam.unibe.ch> > I used PRDocumentParser parse: '""hi all""', > but this method returns PRDocument. > > But I want reneder this string in renederContentOn: method document := PRDocumentParser parse: '""hi all""'. The parser returns a document, this is the root of an AST representing your input. It is independent from the actual output, but it can be easily transformed into about any other format. For example plain text: PRTextWriter write: document --> 'hi all' Back to wiki syntax: PRWikiWriter write: document --> '""hi all""' Latex (you need a special plugin here): PRLatexWriter write: document --> '\textbf{hi all}' And, of course, also to HTML (you need Seaside and Pier-Seaside here): PRViewRenderer new start: document in: aComponent on: html --> 'hi all' Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From firerox at centrum.cz Sat Jan 26 11:39:42 2008 From: firerox at centrum.cz (firerox@centrum.cz) Date: Sat, 26 Jan 2008 11:39:42 +0100 Subject: pier In-Reply-To: <8DC36D36-063A-4B98-912B-8EF4FD3F5FC9@iam.unibe.ch> References: 200801251825.20680@centrum.cz> <200801251825.20680@centrum.cz> <8DC36D36-063A-4B98-912B-8EF4FD3F5FC9@iam.unibe.ch> Message-ID: <200801261139.23887@centrum.cz> Hi, thank you for your answer. It works! But I have still one problem. When I type new line in text area, PRViewRenderer show me whole text in one line. But when i try edit again, in text area i see multiple line text. Where is a problem? Thank you for any help best regards Marek Fiala > >> I used PRDocumentParser parse: '""hi all""', >> but this method returns PRDocument. >> >> But I want reneder this string in renederContentOn: method > > document := PRDocumentParser parse: '""hi all""'. > >The parser returns a document, this is the root of an AST representing >your input. It is independent from the actual output, but it can be >easily transformed into about any other format. > >For example plain text: > > PRTextWriter write: document --> 'hi all' > >Back to wiki syntax: > > PRWikiWriter write: document --> '""hi all""' > >Latex (you need a special plugin here): > > PRLatexWriter write: document --> '\textbf{hi all}' > >And, of course, also to HTML (you need Seaside and Pier-Seaside here): > > PRViewRenderer new > start: document > in: aComponent > on: html --> 'hi all' > >Cheers, >Lukas > >-- >Lukas Renggli >http://www.lukas-renggli.ch > > >_______________________________________________ >SmallWiki, Magritte, Pier and Related Tools ... >https://www.iam.unibe.ch/mailman/listinfo/smallwiki > From renggli at iam.unibe.ch Sat Jan 26 12:15:01 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sat, 26 Jan 2008 12:15:01 +0100 Subject: pier In-Reply-To: <200801261139.23887@centrum.cz> References: 200801251825.20680@centrum.cz> <200801251825.20680@centrum.cz> <8DC36D36-063A-4B98-912B-8EF4FD3F5FC9@iam.unibe.ch> <200801261139.23887@centrum.cz> Message-ID: <713F1F10-CF1E-499E-A7A3-D661A5C906F8@iam.unibe.ch> > It works! > But I have still one problem. > > When I type new line in text area, > PRViewRenderer show me whole text in one line. That's a new feature. You need to add an empty line in-between to get a new paragraph started. This behavior is similar to the one you have in the LaTeX markup language for paragraphs. For example in my image the following test passes: string := 'foo' , String cr , 'bar'. document := PRDocumentParser parse: string. self assert: (PRWikiWriter write: document) = string > But when i try edit again, in text area i see multiple line text. There is something wrong. Do the Pier tests pass? If not try to evaluate: PRDocumentParser classPool removeKey: #LineMatcher; removeKey: #LineMatcherDefault; removeKey: #TextMatcher If that doesn't help, could you show some code so that we can try to reproduce the problem? Lukas -- Lukas Renggli http://www.lukas-renggli.ch From firerox at centrum.cz Sat Jan 26 13:37:08 2008 From: firerox at centrum.cz (firerox@centrum.cz) Date: Sat, 26 Jan 2008 13:37:08 +0100 Subject: pier In-Reply-To: <713F1F10-CF1E-499E-A7A3-D661A5C906F8@iam.unibe.ch> References: 00801251825.20680@centrum.cz> <200801251825.20680@centrum.cz> <200801261139.23887@centrum.cz> <713F1F10-CF1E-499E-A7A3-D661A5C906F8@iam.unibe.ch> Message-ID: <200801261337.28615@centrum.cz> Hi Lukas, >For example in my image the following test passes: > > string := 'foo' , String cr , 'bar'. > document := PRDocumentParser parse: string. > self assert: (PRWikiWriter write: document) = string Yes, it passes me too. But maybe I think somthing other> PRViewRenderer new start: (PRDocumentParser parse: 'foo' , String cr , 'bar') in: self on: html. this code show me: foobar But I would like to see: foo bar Thank you. Marek >> It works! >> But I have still one problem. >> >> When I type new line in text area, >> PRViewRenderer show me whole text in one line. > >That's a new feature. You need to add an empty line in-between to get >a new paragraph started. This behavior is similar to the one you have >in the LaTeX markup language for paragraphs. > >For example in my image the following test passes: > > string := 'foo' , String cr , 'bar'. > document := PRDocumentParser parse: string. > self assert: (PRWikiWriter write: document) = string > >> But when i try edit again, in text area i see multiple line text. > >There is something wrong. Do the Pier tests pass? If not try to >evaluate: > > PRDocumentParser classPool > removeKey: #LineMatcher; > removeKey: #LineMatcherDefault; > removeKey: #TextMatcher > >If that doesn't help, could you show some code so that we can try to >reproduce the problem? > >Lukas > >-- >Lukas Renggli >http://www.lukas-renggli.ch > > >_______________________________________________ >SmallWiki, Magritte, Pier and Related Tools ... >https://www.iam.unibe.ch/mailman/listinfo/smallwiki > From renggli at iam.unibe.ch Sat Jan 26 14:59:49 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sat, 26 Jan 2008 14:59:49 +0100 Subject: pier In-Reply-To: <200801261337.28615@centrum.cz> References: 00801251825.20680@centrum.cz> <200801251825.20680@centrum.cz> <200801261139.23887@centrum.cz> <713F1F10-CF1E-499E-A7A3-D661A5C906F8@iam.unibe.ch> <200801261337.28615@centrum.cz> Message-ID: <21CEC9FC-16B4-414C-84E8-FDAC141ED4EC@iam.unibe.ch> >> But I would like to see: > foo > bar Then you have to press enter twice: foo bar Also see the documentation of the syntax in Pier: "Paragraphs: A single newline has no effect on the layout. An empty line starts a new paragraph." Lukas > > > Thank you. > > Marek > >>> It works! >>> But I have still one problem. >>> >>> When I type new line in text area, >>> PRViewRenderer show me whole text in one line. >> >> That's a new feature. You need to add an empty line in-between to get >> a new paragraph started. This behavior is similar to the one you have >> in the LaTeX markup language for paragraphs. >> >> For example in my image the following test passes: >> >> string := 'foo' , String cr , 'bar'. >> document := PRDocumentParser parse: string. >> self assert: (PRWikiWriter write: document) = string >> >>> But when i try edit again, in text area i see multiple line text. >> >> There is something wrong. Do the Pier tests pass? If not try to >> evaluate: >> >> PRDocumentParser classPool >> removeKey: #LineMatcher; >> removeKey: #LineMatcherDefault; >> removeKey: #TextMatcher >> >> If that doesn't help, could you show some code so that we can try to >> reproduce the problem? >> >> Lukas >> >> -- >> Lukas Renggli >> http://www.lukas-renggli.ch >> >> >> _______________________________________________ >> SmallWiki, Magritte, Pier and Related Tools ... >> https://www.iam.unibe.ch/mailman/listinfo/smallwiki >> > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch From cdrick65 at gmail.com Sun Jan 27 10:42:18 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 10:42:18 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) Message-ID: Hi, My import through SIXX is OK. I can load it but when I try to login, the squeak image crash. Does someone have experienced the same problem or have an idea of the problem origin ? only updating the image causes trouble too. Thanks C?drick -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/fa84b02d/attachment.html From renggli at iam.unibe.ch Sun Jan 27 10:49:27 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 10:49:27 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: Message-ID: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> > My import through SIXX is OK. > I can load it but when I try to login, the squeak image crash. Do you have a crash-dump? Do you have block-contexts or MADynamicObject instances in your serialized mode? If you do, you have to get rid of them beforehand: PRComponent allInstancesDo: [ :each | each reset ] Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From cdrick65 at gmail.com Sun Jan 27 11:07:11 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 11:07:11 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> Message-ID: > > > Do you have a crash-dump? no... it's like a buffer overflow... showing only in the bottom part of the image screen... I've seen a line about BlockContext... Do you have block-contexts or MADynamicObject instances in your > serialized mode? If you do, you have to get rid of them beforehand: > > PRComponent allInstancesDo: [ :each | each reset ] I'll try to make a kernel backup after making that... Is it: PRComponent allInstancesDo: [ :each | each reset ] or PRComponent allSubInstancesDo: [ :each | each reset ] Thanks Lukas -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/4f234655/attachment-0001.html From renggli at iam.unibe.ch Sun Jan 27 11:12:48 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 11:12:48 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> Message-ID: > Is it: > PRComponent allInstancesDo: [ :each | each reset ] > or > PRComponent allSubInstancesDo: [ :each | each reset ] Either one should be fine. I guess you don't have any subinstances, because PRComponent has no subclasses (unless you've created one) ;-) > > > Thanks Lukas > > > _______________________________________________ > SmallWiki, Magritte, Pier and Related Tools ... > https://www.iam.unibe.ch/mailman/listinfo/smallwiki -- Lukas Renggli http://www.lukas-renggli.ch From cdrick65 at gmail.com Sun Jan 27 13:18:38 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 13:18:38 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> Message-ID: > > Either one should be fine. I guess you don't have any subinstances, > because PRComponent has no subclasses (unless you've created one) ;-) oups, right... :) so I tried and still the same... It's not specific to the login... also validating an edit makes the image crash.. maybe I did something bad... to serialize: |xmlString | PRComponent allInstancesDo: [:ea | ea reset]. xmlString := (PRKernel instanceNamed: 'Wiki Viewty') sixxOn: (FileStream fileNamed: 'saveKernelViewty.xml') to load: PRKernel instances add: (Object readSixxFrom: (FileDirectory default readOnlyFileNamed: 'saveKernelViewty.xml')) I put the image here if you have a moment ... ( http://cdrick.free.fr/sq3.10-7159web08.01.1.zip) C?drick -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/3b9f83da/attachment.html From renggli at iam.unibe.ch Sun Jan 27 13:49:58 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 13:49:58 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> Message-ID: <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> > I put the image here if you have a moment ... (http://cdrick.free.fr/sq3.10-7159web08.01.1.zip > ) Aha, I know that problem. You have to reset the persistency using something like this (before serialization and after loading): kernel persistency: PRNullPersistency new The persistency references a semaphore and semaphores obviously don't like to be serialized. Btw, I saw that you have PRImagePersistency set in your image. This will sooner or later result in corrupted sources and changes, if you manually change code or load packages in the image. You have been warned! Lukas -- Lukas Renggli http://www.lukas-renggli.ch From cdrick65 at gmail.com Sun Jan 27 14:03:00 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 14:03:00 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> Message-ID: > > > Aha, I know that problem. You have to reset the persistency using > something like this (before serialization and after loading): > > kernel persistency: PRNullPersistency new Arf, at one point I wanted to try that ! but I forgot :) Thanks Lukas... The persistency references a semaphore and semaphores obviously don't > like to be serialized. > > Btw, I saw that you have PRImagePersistency set in your image. This > will sooner or later result in corrupted sources and changes, if you > manually change code or load packages in the image. You have been > warned! uhm ! What the alternative then... change the persistency to PRNullPersistency, makes changes and updates and then redo image persistency. I've seen there are evolutions to ImagePersistency in the last commits... C?drick ps: btw, I found pier really great as a user :) I serve it at home ans this is working nicely. I have suggestions as an admin page per default but I'llcome back later for that ... :) -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/7ad13a18/attachment.html From cdrick65 at gmail.com Sun Jan 27 14:03:57 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 14:03:57 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> Message-ID: are the last commits stable ? I saw some should not be loaded. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/7ce6e364/attachment.html From renggli at iam.unibe.ch Sun Jan 27 14:08:01 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 14:08:01 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> Message-ID: <69276EC3-766B-49F2-B016-7A2E3BEB67DD@iam.unibe.ch> > What the alternative then... change the persistency to > PRNullPersistency, makes changes and updates and then redo image > persistency. Yes, it is dangerous that this suddenly triggers a snapshot when you are in the middle of doing something. Depending on when this happens (what is quite likely after all) your changes are screwed ... > I've seen there are evolutions to ImagePersistency in the last > commits... As far as I know these changes are made by Doru. He made it easier to make Pier be deployed on seasidehosting.st where there are certain disc limits. I didn't had a look at his changes though. > ps: btw, I found pier really great as a user :) I serve it at home > ans this is working nicely. I have suggestions as an admin page per > default but I'llcome back later for that ... :) Yes, we are working on that. Also on a decent default layout. Etc. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From renggli at iam.unibe.ch Sun Jan 27 14:08:44 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 14:08:44 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> Message-ID: > are the last commits stable ? It is just that the stylesheet is in the middle of a transition. Shouldn't be dangerous if you don't depend on it. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch From cdrick65 at gmail.com Sun Jan 27 14:14:10 2008 From: cdrick65 at gmail.com (cdrick) Date: Sun, 27 Jan 2008 14:14:10 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: <69276EC3-766B-49F2-B016-7A2E3BEB67DD@iam.unibe.ch> References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> <69276EC3-766B-49F2-B016-7A2E3BEB67DD@iam.unibe.ch> Message-ID: 2008/1/27, Lukas Renggli : > > > What the alternative then... change the persistency to > > PRNullPersistency, makes changes and updates and then redo image > > persistency. > > Yes, it is dangerous that this suddenly triggers a snapshot when you > are in the middle of doing something. Depending on when this happens > (what is quite likely after all) your changes are screwed ... is it also the case if one page is edited ? -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.iam.unibe.ch/pipermail/smallwiki/attachments/20080127/c4e4411f/attachment.html From renggli at iam.unibe.ch Sun Jan 27 14:28:25 2008 From: renggli at iam.unibe.ch (Lukas Renggli) Date: Sun, 27 Jan 2008 14:28:25 +0100 Subject: [Pier] Image crash after a kernel import (through SIXX) In-Reply-To: References: <27D8FF37-809D-4808-9192-859A4DFAC33A@iam.unibe.ch> <5A2F8995-557C-4B30-BBE8-3E4350104048@iam.unibe.ch> <69276EC3-766B-49F2-B016-7A2E3BEB67DD@iam.unibe.ch> Message-ID: > 2008/1/27, Lukas Renggli : > What the > alternative then... change the persistency to > > PRNullPersistency, makes changes and updates and then redo image > > persistency. > > Yes, it is dangerous that this suddenly triggers a snapshot when you > are in the middle of doing something. Depending on when this happens > (what is quite likely after all) your changes are screwed ... > > is it also the case if one page is edited ? No, that goes trough the persistency mechanism of Pier, that's not an issue. It is just editing source code. Cheers, Lukas -- Lukas Renggli http://www.lukas-renggli.ch