seaside August 2005

seaside@lists.squeakfoundation.org

54 participants
65 discussions

Re: [Seaside] Template support in Seaside 2.5
by Adnan Zaman 07 Feb '06

07 Feb '06

Hi Alain, Your fix works so thanks for that. However, I see that my html text was not "rendered" as you would expect to see in a browser. All I saw was the same html text that I wanted to be rendered. Not sure if this has anything to do with your another email in which you asked about rendering problem in 2.5. - Adnan >>Date: Tue, 21 Jun 2005 22:42:30 +0200 >>From: Alain Fischer <alain.fischer(a)bluewin.ch> >>Subject: Re: [Seaside] Template support in Seaside 2.5 >>To: "The Squeak Enterprise Aubergines Server - general discussion." >> <seaside(a)lists.squeakfoundation.org> >>Message-ID: <daafe26dd5ae1e8b1e8f16ded82f3fdf(a)bluewin.ch> >>Content-Type: text/plain; charset=ISO-8859-1; format=flowed >> >>Hi Adnan, >> >>For the first issue, I doesn't got a dependency warning. >>Here is how I setup my image: >>- downloaded prebuild image from Avi >>(http://beta4.com/seaside2/Seaside-2.5.zip) >>- used SqueakMap Package Loader to install SmaCC Smalltalk >>Compiler-Compiler-Runtime version: 4 >>- opened a repository on the disk space containing >>SeasideTemplate-AFi.1.mcz >>- loaded this package in the image. >> >>For the second issue here is a fix: >> >>STTTemplate>>compileIn: aClass >> >> | methodNode | >> methodNode _ Compiler new >> compile: self asSmalltalkMethod >> in: aClass >> notifying: nil >> ifFail: [^ nil]. >> aClass >> addSelector: methodNode selector >> withMethod: (methodNode generate: #(0 0 0 0)). >> aClass organization >> classify: methodNode selector >> under: 'generated'. >> "aClass compileProgrammatically: self asSmalltalkMethod classified: >>'generated'" >> "| node method | >> node := Compiler new >> compile: self asSmalltalkMethod >> in: aClass >> notifying: nil >> ifFail: [self error: 'Syntax Error']. >> method := node generate: #(0 0 0 0). >> aClass addSelector: node selector withMethod: method." >> >> >>and it should work. >> >>If you try to run STTTest in Test Runner, you will notice: >>'13 run, 12 passes, 0 expected failures, 0 failures, 1 errors, 0 >>unexpected passes' >> >> >>Alain __________________________________ Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone. http://mobile.yahoo.com/learn/mail

4 4

Version 2.5b6 for VW is ready
by Bany, Michel 06 Nov '05

06 Nov '05

I just published Seaside 2.5b6 for VW on the public repository. It is based upon Seaside2.5b6-avi.8.mcz 15-Jan-2005 16:15 i.e. the very latest version of 2.5b6. This port includes all the fixes that were discussed here: - file upload - lost halo icons and missing styles Note: I also fixed Seaside 2.5b5 on the public repository. For the fun, I added the Seaside logo on some of the pages. Bundle to load : either SeasideForWebToolkit or SeasideForSwazoo. Enjoy, Michel.

3 4

Nested form tags
by Bany, Michel 04 Nov '05

04 Nov '05

I came accross some weird situation when using a component that is a composition of other components. A subcomponent may in turn be a composition of some more components. To make things even more complex, some deeply nested subcomponents may call some other component. Some of the subcomponents may contain some input html tag and as such will require to be contained in a form html tag. In such a complex construct it may be difficult to decide exactly what component should have the responsibility to generate the form tag. Also, it happened frequently to me that form tags become nested especially when a subcomponent calls another component that happens to be a dialog. Therefore I am making the following suggestion : Seaside should prevent form tags from being nested. It should discard any form tag that occurs within a previously opened form tag. With this in place, I can design a component that needs a form with a form construct knowing that if my component happens to be used inside a form, then my form construct will be discarded. Otherwise my form construct will be preserved. If my component calls a dialog component, the form generated by the dialog component will be discarded if my component happens to be used from inside a form. Michel.

6 6

[ANN] the Bicephale project
by Michal 12 Oct '05

12 Oct '05

Hi all - I would like to announce the first version of the Bicephale project: <http://software.auf.net/bicephale/Bicephale3-mist.1.mcz> The goal of Bicephale is to change Seaside to make it handle elegantly the display of static, long-lived objects. As it stands, Seaside is very powerful at handling sessions - sequences of pages that represent a dynamic operation and for whom bookmarkability and long-term persistence is not an issue. Seaside is however somewhat clunky at displaying simple, unchanging pages with stable and bookmarkable urls. Since most applications with sessions also need such stable pages to display the objects they are editing, Seaside suffers from this even on its home turf. To take an example, if you view a blog entry with Smallblog - the Seaside blog application - you get a different url every time you view the entry. This creates several problems: (i) such pages cannot be easily bookmarked although they should be (Smallblog resorts to a workaround for this); (ii) they cannot be indexed although one would want them to; (iii) such pages 'expire' after a while, and clicking on a link after expiration has unexpected effects, confusing the user. Almost every Seaside application suffers from this, whether it is a store, a ticketing application or an academic article repository. There are currently workarounds for these issues, but Bicephale aims at solving the problem at its root, in an elegant and principled way. It does that by introducing static, long-lived callbacks (following up on these two February 2005 posts on the Seaside list by Avi and myself: <http://lists.squeakfoundation.org/pipermail/seaside/2005-February/004400.ht…> <http://lists.squeakfoundation.org/pipermail/seaside/2005-February/004436.ht…> ) These lasting callbacks are rendered by a new type of components (WAStaticComponent) that live outside of sessions but within the application. Apart from having stable urls, these components behave just like session components: i. they use standard Seaside rendering methods - e.g. they render with the usual #renderContentOn: and link to another object with 'html anchor: anObject text: aString'. ii. they can #call: traditional dynamic components (WAComponent) and receive an #answer: from them. A static component can for instance include a (stable) link to edit the object currently displayed. That link will trigger a standard seaside login dialogue and based on the answer, it initiates a traditional Seaside editing session. iii. the handling of urls is largely transparent to the developper. Much of Seaside's appeal is the ability to concentrate on the logic of the application and have the framework automatically handle the details of the underlying http transaction, including the urls used. This is the same with static components. A Bicephale application (WABicephale) has therefore two root components: the usual session root component and the new root static component (hence the Bicephale name). By default the application renders the static component, and sessions are used only when a #call: is issued. This first version implements all of the above, but in a preliminary and experimental way. This means that it is now almost trivial to display static object with automatically generated fixed urls and have them call/answer dynamic components. At the same time, I am still experimenting with various implementations, and looking for the right abstractions -- particularly in the handling of urls/component composition. Since almost every design decision is still open to drastic revision, I would appreciate feedback! Finally, there is currently no UI or documentation. I will however post a sample application, and if there is interest, I will provide some basic documentation. Michal ps. there are no modification to base Seaside classes in this package - it only adds new classes. It should hence be harmless to install it into your Seaside image, but this being an early prototype, caveat emptor.

5 7

Re: [ShoreComponents] WebTree, SVG demo and dynamic files
by Jan B. Krejčí 26 Sep '05

26 Sep '05

> From: radoslav hodnicak <rh(a)4096.sk> > > 3) a decorator allowing to preserve page scrolling coordinates between > > submits, so if you have scrolled 5 feet down to find the button you > > want to click, you will see the same page location after the response > > is loaded from server. I saw this in .NET and I like it very much. > > Do you know how is it implemented in .not? Page anchors won't scroll > vertically (at least not in IE). I'd like to have this one too so I'm > willing to implement it. sure. please have a look into enclosed HTML file. It's done this way: 1) reister onScroll handler to <body> 2) in the handler store the coordinates into hidden form fields 3) make sure they are submitted with each clickable component you need 4) put a script at the end of the page causing the scroll position is restored after the page is loaded (or use <body onLoad="..."> that's all > I will try to pick some interesting pieces from my app and add it to the > collection but really, can we pick a better name than ShoreComponents? It > hurts my fingers. Something shorter, like Seadgets or whatever... uh, well, I do not really insist on "ShoreComponents" name, but please consider what impact would that have. How about creating some voting component, expose it in the demo and count other's opinions? :) -- .:jbk:. Jan B. Krejci

11 12

RE: [Seaside] BUG: Basic authentication
by Mart-Mari Breedt 07 Sep '05

07 Sep '05

Thank you for your reply. We are using Seaside 2.6a2, but the problem also occurs in Seaside 2.6a1. I added WAAuthConfiguration to the basic counter application and created a second entry point to the counter application (counter2) and added WAAuthConfiguartion to it as well. I specified the username and password as being admin and seaside respectively on both. I tried accessing the application from different browsers, but I get the same behavior in Firefox, Opera and IE. The user is prompted for a password when the first entry point is accessed, but the user is definitely not prompted when trying to access the second entry point. I can see the realms are indeed different for each entry point since the realm is displayed when prompted for a password. I.e. when I access the entry points in a reversed order from a new browser session I can see the counter2 realm is different to the counter realm. Thank you, Mart-Mari -----Original Message----- From: Julian Fitzell [mailto:julian@beta4.com] Sent: 31 Augustus 2005 04:40 To: The Squeak Enterprise Aubergines Server - general discussion. Subject: Re: [Seaside] BUG: Basic authentication Hmm... I can't reproduce this obviously with Seaside 2.5. The two applications obviously have different basePath's and it is the basePath that Seaside uses by default as the realm for HTTP Basic Authentication. As long as the applications are providing different authentication realms, the user should be reprompted for a password. If this is not the case, this is obviously a browser bug. Can you confirm which version of seaside you're using and perhaps how you are setting up the authenticated applications... is there any reason they would actually be using the same realm? Julian Quoting Mart-Mari Breedt <breedt_m(a)aircom.co.za>: > Hallo all, > > > > We have the following problem. Consider you have two instances of the > same seaside application, each set up with basic authentication. > Example, you have two instances of the counter application (named > counter1 and counter2) with username=admin and password=seaside on both. > > > > > When you log on to seaside/counter1, you would be prompted for a > username and password. (Which is correct...) When you now log on to > seaside/counter2 (from the same browser session) you would NOT be > prompted for a username and password. The WARequest object contains the > previous username and password and since it is the same as the username > and password for this application, you are automatically validated. This > (We believe) is a bug, since you should have been prompted for a new > username and password for two reasons. One: Because you are starting a > new session and Two: Because you are accessing a totally different > application instance. > > > > Does anyone have any ideas or previous experiences on fixing this? > > > > Thank you, > > > > Mart-Mari > > > > > > -- julian(a)beta4.com Beta4 Productions (http://www.beta4.com) _______________________________________________ Seaside mailing list Seaside(a)lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/seaside

6 7

WADateSelector FIX
by Dmitry Dorofeev 31 Aug '05

31 Aug '05

Hi all, Hopefully i found a bug and fixed it. It was impossible to choose some dates (like 31 December and many others) becouse of incorrect month validation. Menu for months is generated as 1 to: 12 but month check was >=0 and < 12. fix is attached. -Dmitry. 'From Squeak3.8 of ''5 May 2005'' [latest update: #6665] on 31 August 2005 at 1:20:54 pm'! !WADateSelector methodsFor: 'as yet unclassified' stamp: 'DVD 8/31/2005 13:18'! privateIsValidDay: theDay monthNumber: theMonth year: theYear | maxDay | theMonth > 0 & (theMonth <= 12) ifFalse: [^false]. maxDay := Month daysInMonth: theMonth forYear: theYear. ^theDay > 0 & (theDay <= maxDay)! !

3 3

Help! ByteString is out of control
by Jason Rogers 31 Aug '05

31 Aug '05

I posted this on the Squeak dev list, but it dawned on me that I should cross post here as well in case it was a Seaside isse. I have an image that is serving up logFiles using Seaside and Comanche. It has been steadily growing in size. It's now up to 49+MB. Recently it started hanging. I have been able to run a SpaceTally on it and found that 33.1+MB are used for instances of ByteString. I haven't been able to track down all 99000+ instances, but I am almost sure that some of them are supposed to be temporary and are not getting garbageCollected. Has anyone else seen something similar to this? (Radoslav says he is seeing the same behavior) -- Jason Rogers "I am crucified with Christ: nevertheless I live; yet not I, but Christ liveth in me: and the life which I now live in the flesh I live by the faith of the Son of God, who loved me, and gave himself for me." Galatians 2:20

3 4

OmniBase-mjr.4
by Michael Roberts 31 Aug '05

31 Aug '05

I've attached OmniBaseTests based on OmniBaseTester. I knocked this up quickly and hope it might be useful in chasing down some corruption issues. Feel free to fix/improve etc. Cheers, Mike From the comment: "Provides a version of OmniBaseTester written as a TestCase that is hopefully semantically equivalent. Adds a simple contention test in #test16 that does the image spawning for you using OSProcess. After merging, set #tmpDir: to something sensible for your platform."

2 1

Issue with #checkForCookies
by German Viscuso 31 Aug '05

31 Aug '05

Hello list. When I try to use #checkForCookies in a seaside component I obtain the following message in the Transcript: "Warning: You have tried to set a cookie and a redirect in the same page, responding to the request /seaside/go/myapp. Some browsers will not handle this correctly." Has this issue been already reported or solved? (I'm using VW 7.3 nc and the web toolkit) My test code is: --------------------------------------------------------------------------- renderContentOn: html self session checkForCookies ifTrue: [ self inform: 'Cookies enabled!'. self answer: 'ok'] ifFalse: [ self inform: 'You must have cookies enabled to proceed.'. self answer: nil ] --------------------------------------------------------------------------- Regards and thanks in advance, GV

3 4

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

seaside August 2005