seaside-dev November 2007

seaside-dev@lists.squeakfoundation.org

11 participants
19 discussions

submitFormNamed: weirdness
by Michael Lucas-Smith 28 Nov '07

28 Nov '07

Hi All, Just recently I hit a weird bug with anchor's submitting forms.. specifically using #submitFormNamed: If you do something like: html anchor callback: [self save]; submitFormNamed: 'form'; with: 'Save'. You'll have the web browser fire off a POST followed immediately by a GET on two separate connections making it impossible for Seaside to know which should be processed first. Often the GET will finish first, causing the browser to redirect to a different page and cancel the POST command - which may or may not have run yet. I saw this appear in the form of a bad demo at OCSTUG where the POST submits the form and the GET writes to the database. However, depending on timing, the write to the database would have empty values. If you use #ignoreURL to put a 'return false' on the end of the onclick you quickly discover that the anchor's callback is never run.. which is puzzling when I found this thread: http://www.nabble.com/Anchor-with-submit-and-callback-t2648056.html from 2006 describing how submitFormNamed: should preserve the action id of the previously setup callback. This code doesn't exist in Seaside 2.8 final. Right now a workable hack is: html anchor callback: [self save]; submitFormNamed: 'form'; onClick: 'window.location = this.href'; ignoreURL; with: 'Save' Which is just ugly as :) and it does mean the browser is still very likely to terminate the second POST connection on the server side which is likely to raise a "Client disconnect" sort of exception (which is probably caught by the server - however, that's sort of besides the point). Cheers, Michael

4 9

[2.9] Seaside-Core-pmm.9 and friends
by Philippe Marschall 24 Nov '07

24 Nov '07

Hi This includes a rather big list of changes. I'm sorry I didn't cut it up into smaller patches. The fist thing it includes is a new string conversion method #toString which replaces #displayString and #asString. Semantics are similar to #asString in Squeak. WAPlatfromTest >> #testToString tests them. The second thing that it introduces are classes for mime types and locales. The reasoning behind this is the same as having a first class URL objects instead of strings. There is also experimental code in there for content and language negotiation. Most of the existing code should be unaffected by this and can still use strings if he wants to just like with URLs. The two exceptions that come into my mind are server adaptors and senders of WAFile >> #contentType. They must be aware that WAMimeType objects are returned. While we are breaking the existing server adapter interface it might be a good time to include #remoteAddress to WARequest. So WASessionProtector would not have to delegate to #nativeRequest which is against the spirit of WAReqest. Cheers Philippe

1 0

Seaside One-Click Experience VM Suggestion
by Ron Teitelbaum 17 Nov '07

17 Nov '07

Hi all, I'm not sure if you saw John's comment on weekly squeak: http://news.squeak.org/2007/11/10/seaside-one-click-experience/#comment-2439 5 Ron Teitelbaum

2 1

spam
by Avi Bryant 13 Nov '07

13 Nov '07

Is anyone else seeing an unbelievable amount of spam on the seaside and seaside-dev lists? It may just be me, but very little of it gets filtered out (compared to my inbox where it nearly all does). I can't figure out what settings I need to change to solve this - afaict, we're already only accepting messages from registered members, which the spammers obviously aren't... Avi

1 0

implementation classes and initialization
by Philippe Marschall 11 Nov '07

11 Nov '07

Hi One of the main consequences of breaking Seaside into modules is that SeasidePlatformSupport does no longer include the implementation. Instead it delegates to an implementation that will be provided later by an other package. An important consequence of this is that during the initialization of the Seaside-Core package the implementation is not yet present. This means you can't automatically register components in the Seaside-Core package in the #initialize method on the class side. I moved this code to #initializeClass methods which I send during the initialization of the Seaside-Squeak-Core package. Some clarification about the Seaside-Sport-Core package: The idea is not that we move Seaside to SPort and for example replace the chronology classes with their SPort equivalents. Instead it is only a partial implementation of SeasidePlatformSupport using SPort. It is not a full implementation because SPort does not provide all the functionality we need. The target audience for this is not existing ports but new ports that already have SPort. This reduces the amount of implementation code they have to write. Existing ports are of course free to move to SPort if they want to. Cheers Philippe

1 0

Server Adapters in 2.9
by Philippe Marschall 11 Nov '07

11 Nov '07

Hi When working on the Swazoo 2 server adapter I realized a lot of stuff that is wrong with WAKom. I ended up duplicating stuff from WAKom or writing stuff that would make sense for WAKom as well. Now that we have modules in Seaside and WAKom is in its own module I had time to refactor this stuff. The common port of this is in Seaside-Adapters-Core. Since we need to do platform specific stuff (encoding) there has to be a Squeak implementation which is in Seaside-Squeak-Adapters. I changed WAKom (Seaside-Squeak-Kom) and Seaside for Swazoo (Seaside-Adapters-Swazoo) to use this common code. A lot of the ugliness of the Kom hierarchy is now gone and in theory we could even get rid of WAKomEncoded* entirely and toggle encoding in WAKom on demand. The idea is not that everybody switches to use this code but if you're doing a new port or server adapter you might want to have a look at it. There has been some interest in a Swazoo 2 adapter for other dialects than Squeak and this should make it easier. Cheers Philippe

1 0

Seaside2.8a1-rjl.495
by Philippe Marschall 10 Nov '07

10 Nov '07

This looks good to me there are however two buts. The first is WADocumentHandler >> #response and WAResponse class >> #document:mimeType: should IMHO be changed to use #cacheForever. The second is whether this is not too late for 2.8 now that we are under code freed and should wait for 2.9. Any opinions? Cheers Philippe

2 4

Seaside 2.9 development started
by Philippe Marschall 10 Nov '07

10 Nov '07

Hi The main goal of this years Opertation STABILO [1] is to break Seaside into modules. The main idea is that porters no longer need to rip all the Squeak stuff out. Instead they can load the core Seaside which is supposed to be the same on all dialects and add thier own implementatin classes. This will hurt for a moment until all your tools are updated but should be much better for the future especially with all the ports to come (VAST, GST, Ambrai, you name it). SeasidePlatformSupport got more ugly but smaller since some methods were only used for the Squeak implementation. There is also an experimental implementation of SeasidePlatform support based on SPort. All this work is based on lr.522 so the very latest is not yet in. Thanks to the Swiss Army for giving their soldiers an opportunity to provide a service to the community: hacking Seaside. More will follow, there are two weeks to go and not much else to do. [1] http://www.news.admin.ch/dienstleistungen/00009/index.html?lang=de&event-id… Cheers Philippe

4 3

Seaside 2.8a1-eam.528 for Dolphin Smalltalk
by Esteban A. Maringolo 09 Nov '07

09 Nov '07

Hi all, I've just released a preview version of the ongoing port of Seaside for Dolphin I've been working on. You can find the instructions, notes, and download link from: <http://dolphinseaside.blogspot.com/2007/11/seaside-28a1-eam528-preview-vers…> Best regards, -- Esteban A. Maringolo eMaringolo(a)gmail.com

1 0

Handling of session expiration
by Esteban A. Maringolo 08 Nov '07

08 Nov '07

Hi all, I've been stressing Seaside on Dolphin, with very good results, but as the test goes by the amount of garbage grows, and is never claimed back, leaving some continuations referenced by action handlers which will never be activated. If I place a background process cleaning all the handlers every 30 seconds, the image runs very well, with no unnecesary continuations, nor nothing undesired. [[Processor sleep: 30000. WARegistry clearAllHandlers. true ] repeat] forkAt: Processor userBackgroundPriority . How do you handle the "cleaning" of undesired/expired WARequestHandler's? (and all of its subclasses). I've run the same test against Seaside on Squeak, and evaluating "Continuation allSubinstances" answer a very large list of instances too, so I believe this issue is not Dolphin-only. Best regards, -- Esteban A. Maringolo eMaringolo(a)gmail.com

3 5

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

seaside-dev November 2007