[ANN] WideString findSubstring patch attempt

I put some workaround at http://bugs.squeak.org/view.php?id=6366
Maybe tests can be completed, but that's a start.
Any volunteer to cross check ?


This also makes the test from Philippe give these results:

'#:\*' match: (String with: (Unicode value: 50976) with: (Unicode
value: 47532) with: (Unicode value: 47484)). -> false

'*' match: (String with: (Unicode value: 50976) with: (Unicode
value: 47532) with: (Unicode value: 47484)). -> true



Beware: (Character value: 47484) and (Unicode value: 47484) are not the
same because of the charSet encoded in leadingChar...

In curent multilingual implementation, (Character value: 47484) is not
really usable...

Nicolas

2007/7/28, nicolas cellier <[hidden email]>:
Well it gives me exactly the same I would get when running a web application.

Philippe

El 7/27/07 8:11 PM, "nicolas cellier" <[hidden email]> escribió:

Nicolas:
Thanks for your work, but #('WideStringTest>>#testSubstrings') still fail in
3.10 with your fix

Edgar J. De Cleene a écrit :
>
> Nicolas:
> Thanks for your work, but #('WideStringTest>>#testSubstrings') still fail in
> 3.10 with your fix
>
>

Hi Edgar,
It seems you have been confused by the names...

#testSubstrings was designed for another bug, substrings.
As indicated in comment it is related to
http://bugs.squeak.org/view.php?id=6367

It seems you have loaded my test but not my patch.
So it leads to a failure in 3.10 7137.

Note that Andreas provided an alternate solution later at
http://bugs.squeak.org/view.php?id=6573

I tried to open a discussion about respective interest of both
solutions, but Andreas seems too busy.

So, up to you to check and choose one...

--------------------------------

Concerning http://bugs.squeak.org/view.php?id=6366, i loaded the test
(#testFindSubstring) then the patch in 3.10 7137, and the WideString
tests pass.

Nicolas

nicolas cellier a écrit :
Edgar,

To follow up, I re-checked the tests provided by johnmci at
http://bugs.squeak.org/view.php?id=5331

Though they are not complete (these WideString involve only characters
with charCode<256..), they are stressing latest 3.10.

However, all proposed WideString-*-Test.*.cs pass if you apply following
patches:

http://bugs.squeak.org/view.php?id=5331 beginsWithWideStringPatch.1.cs
http://bugs.squeak.org/view.php?id=6367 WideString-substrings-Test.1.cs
http://bugs.squeak.org/view.php?id=6366 
Collection-WideString-findSubstring-Patch.1.cs
http://bugs.squeak.org/view.php?id=3574 
Collection-String-IndexOf-Patch.3.cs

Last one is already in 3.10 7137

cheers

Nicolas

El 7/28/07 6:43 PM, "nicolas cellier" <[hidden email]> escribió:

Yes , I load both...
 
>> Note that Andreas provided an alternate solution later at
>> http://bugs.squeak.org/view.php?id=6573
>>
>> I tried to open a discussion about respective interest of both
>> solutions, but Andreas seems too busy.

Yes , I see on Mantis before and wait both you reach some agreement.

Maybe now Andreas see I take risk and put his needed code with less burro
cracia ... (Semaphore thing)


>> So, up to you to check and choose one...
>>
>> --------------------------------
>>
>> Concerning http://bugs.squeak.org/view.php?id=6366, i loaded the test
>> (#testFindSubstring) then the patch in 3.10 7137, and the WideString
>> tests pass.
>>
>> Nicolas

Tomorrow I recheck all.

Very thanks for work !

Edgar