printOn: a WAUrlEncoder can get interesting when no WACurrentContext on stack

[GoogleCodeExporter]

1. put a halt in WAUrlTest>>testParseAmbiguous as follows:
testParseAmbigous
    self ambigousExamples do: [ :each |
        self 
            assert: (WAUrl halt absolute: each first) toString = each last
            description: each last resumable: true ]

2. run the test and step into WAUrl>>initializeFromString:
3. After stepping over 'self initialize', select self and you'll get an
error, because there is no WACurrentRequest on the stack .... '<error in
printString: evaluate "self printString" to debug>'...

Presumably the folks that use WACurrentRequest to get values should do nil
checks to avoid getting such errors ... 

Original issue reported on code.google.com by [email protected] on 19 May 2008 at 11:08

[GoogleCodeExporter]

The issue is probably in WAUrlEncoder >> #initializeOn:table: . The question is 
what
do we do if we know that there is no request context. Either raise an error or 
use a
NullCodec. I'd rather raise an error.

Original comment by [email protected] on 31 May 2008 at 12:26

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Ah, just stumbled across this bug. I'd rather use a NullCodec otherwise doing
development just seems painful. In fact, I already committed this change in:

---------------
Name: Seaside-Core-jf.183
Author: jf
Time: 29 July 2008, 12:25:59 pm
UUID: 7dc4db18-c648-cd4d-81c8-5c485b9a3d0f
Ancestors: Seaside-Core-jf.182

Have WAUrlEncoder use a null codec if it can't find any other in the request 
context.
------------------

As per our discussion on the list, though, this introduces a dependency the 
wrong way
with Seaside-Adapters-Core. So either we pull WANullCodec and WACodec into core 
or we
implement WAUrl>>printOn: so it doesn't need a codec and does an approximation 
the
dev tools can use and WAUrl>>toString so it uses the original codec 
implementation.

Or we decide it doesn't matter that you can't easily work with WAUrl objects in 
the
debugger/inspector... seems a shame though.

Original comment by [email protected] on 5 Aug 2008 at 2:04

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Preventing an error during printOn: would be the first order of business, being 
able
to get a meaningful printString during debugging/inspecting is secondary, but 
still
pretty useful ... as it is you have to inspect the WAUrl and reconstruct the url
string from it's pieces, so a null codec would be more useful than not ... 
another
argument for including a codec in the core

Original comment by [email protected] on 5 Aug 2008 at 4:07

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Ok, you guys convinced me.

Original comment by [email protected] on 5 Aug 2008 at 5:10

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Ok, WACodec is in core now. Seaside-Adapters-Core, Seaside-Adapters-Squeak and
Seaside-Adapters-Test are all no more. Same goes for WAGenericCodec.
SeasidePlatformSupport >> #utf8Codec has been replaced with #genericCodedClass.

Name: Seaside-Core-pmm.184
Author: pmm
Time: 5 August 2008, 11:04:09 pm
UUID: 725e41b4-a0ad-4dbd-b45a-0a46cc4f6ca2
Ancestors: Seaside-Core-jf.183

- moved WACoded to Seaside-Core

Name: Seaside-Squeak-Core-pmm.52
Author: pmm
Time: 5 August 2008, 11:04:41 pm
UUID: bfddc629-387d-4b23-997e-ce8ad626c871
Ancestors: Seaside-Squeak-Core-jf.51

- moved WACoded to Seaside-Core

Name: Seaside-Squeak-Tests-pmm.7
Author: pmm
Time: 5 August 2008, 11:05:02 pm
UUID: dcf7a200-2493-4fbd-8a9d-349d8b1dc5b9
Ancestors: Seaside-Squeak-Tests-pmm.6

- moved WACoded to Seaside-Core

Name: Scriptaculous-Tests-pmm.47
Author: pmm
Time: 5 August 2008, 11:05:23 pm
UUID: 39049980-e33f-40dd-99a5-402b432f4b8f
Ancestors: Scriptaculous-Tests-obi.46

- moved WACoded to Seaside-Core

Name: RSS-Tests-pmm.3
Author: pmm
Time: 5 August 2008, 11:05:48 pm
UUID: 943438b1-271b-4fc5-b6a9-e4eefe8aff74
Ancestors: RSS-Tests-pmm.2

- moved WACoded to Seaside-Core

Original comment by [email protected] on 5 Aug 2008 at 9:09

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

This change has introduced a wrong way dependency from Seaside-Squeak-Core to 
Seaside-Core (since WAGenericSqueakCodec and WAUtf8SqueakCodec subclass WACodec.

Original comment by [email protected] on 29 Aug 2008 at 1:53

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

I moved the continuation to its own package. The new load order is now supposed 
to be:

1. Seaside-Squeak-Continuation
2. Seaside-Core
3. Seaside-Squeak-core

That should work withhout any undeclareds or circular dependencies. 

Original comment by [email protected] on 30 Aug 2008 at 9:51

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Seems fixed. John, can you verify and reopen if necessary?

Original comment by [email protected] on 30 Aug 2008 at 5:39

• Changed state: Fixed
• Added labels: ****
• Removed labels: ****