Hi,
When loading the latest and greatest Sport 2.031; Swazoo 2.3beta2.6; Aida6.4-janko.5; scribo1.0-beta-janko.9 I get the DNU from my previous post which is caused by Aida6.4-janko5 having changed Object>>firstRequestFromStack .............................. isBlock not and: [(stContext receiver isKindOf: AIDASite) and: [stContext arguments size > 1 "------------- this condition was added" and: [(stContext at: 2) isKindOf: HTTPRequest]] ] ................. where stContext is a MethodContext which does not understand #arguments.
This also happens when scribo is not loaded. So Aida6.4-final.1.mcz seems the right version of Aida to use. In this configuration i can log into the Aidademo site on port 8888.
If I load either the beta version or the previous version of scribo (jm 192) on top of this I get the attached debugger. Climbing up the stack ends up trying to send #isFromMSIE to an instance of Writestream which DNU's.
BTW in Swazoo 2.3beta2.6 a lot of the swazoo tests fail. I use an older version in trunk where much less tests fail, but that can be left for later.
I understand too little of this climbing up the stack to feel confident debugging that. Janko?
Cheers
Herbert
On 21. 06. 2012 21:55, Herbert König wrote:
Hi,
When loading the latest and greatest Sport 2.031; Swazoo 2.3beta2.6; Aida6.4-janko.5; scribo1.0-beta-janko.9 I get the DNU from my previous post which is caused by Aida6.4-janko5 having changed Object>>firstRequestFromStack .............................. isBlock not and: [(stContext receiver isKindOf: AIDASite) and: [stContext arguments size > 1 "------------- this condition was added"
And it seems Pharo understands that while Squeak not?
and: [(stContext at: 2) isKindOf: HTTPRequest]] ]
................. where stContext is a MethodContext which does not understand #arguments.
This also happens when scribo is not loaded. So Aida6.4-final.1.mcz seems the right version of Aida to use. In this configuration i can log into the Aidademo site on port 8888.
If I load either the beta version or the previous version of scribo (jm 192) on top of this I get the attached debugger. Climbing up the stack ends up trying to send #isFromMSIE to an instance of Writestream which DNU's.
Can you just skip usage of tham method by temporary commenting it out?
BTW in Swazoo 2.3beta2.6 a lot of the swazoo tests fail. I use an older version in trunk where much less tests fail, but that can be left for later.
I understand too little of this climbing up the stack to feel confident debugging that. Janko?
Normally you should not cope with that, this is really internal stuff. But when some method is directly or indireclty using #first*FromStack methods, you should know the basic idea and consequences, as said specially that you cannot debug such method call because those #first*FromStack methods will fail finding desired reference from the calling stack.
Best regards Janko
Cheers
Herbert
Webteam mailing list Webteam@lists.squeakfoundation.org http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/webteam
Am 21.06.2012 22:04, schrieb Janko Mivšek:
On 21. 06. 2012 21:55, Herbert König wrote:
Hi,
When loading the latest and greatest Sport 2.031; Swazoo 2.3beta2.6; Aida6.4-janko.5; scribo1.0-beta-janko.9 I get the DNU from my previous post which is caused by Aida6.4-janko5 having changed Object>>firstRequestFromStack .............................. isBlock not and: [(stContext receiver isKindOf: AIDASite) and: [stContext arguments size > 1 "------------- this condition was added"
And it seems Pharo understands that while Squeak not?
I have no knowledge of Pharo, too little time to follow another Smalltalk.
Can you just skip usage of tham method by temporary commenting it out?
I made it return false in case it's used in other places and it works now. I get three additional links on the bottom of the welcome page. They all seem to work but after I looked back at the Squeak image I got several more debuggers.
They all seem to originate from WikiPage>>preferedUrl in ifTrue: [^'/wiki/', self title replaceSpecialCharacters asLowercase, '.html'] where title is a ByteString which DNU #replaceSpecialCharacters
I continue commenting out and report where that gets me.
Normally you should not cope with that, this is really internal stuff. But when some method is directly or indireclty using #first*FromStack methods, you should know the basic idea and consequences, as said specially that you cannot debug such method call because those #first*FromStack methods will fail finding desired reference from the calling stack.
I understood the last point.
Thanks,
Herbert
webteam@lists.squeakfoundation.org