On Thu, May 7, 2015 at 8:48 AM, Nicolas Cellier <
nicolas.cellier.aka.nice at gmail.com> wrote:

> Does that mean that current Spur update process is not robust to a further
> change of Character class definition? (like adding a class var for example)
>

The Monticello package patcher is robust for class var and pool dictionary
changes.  I think the image bootstrap is too.  But checking that it is is
timeconsuming.  That's one major motivation behind the feature freeze.  I
want to stop wasting cyclers on the bootstrap and progress to the full
release where we're free to modify definitions as we see fit.


>
> 2015-05-07 15:56 GMT+02:00 Levente Uzonyi <leves at elte.hu>:
>
>> Before Spur Character has always had an instance variable named value.
>> Characters are immediate in Spur, so the Character class has no instance
>> variables there. If you browse Collections.spur-ul.584.mcz, you'll find the
>> following definition:
>>
>> Magnitude immediateSubclass: #Character
>>         instanceVariableNames: ''
>>         classVariableNames: 'CharacterTable ClassificationTable
>> DigitValues LetterBits LowercaseBit UppercaseBit'
>>         poolDictionaries: ''
>>         category: 'Collections-Strings'
>>
>>
>> Levente
>>
>> On Thu, 7 May 2015, karl ramberg wrote:
>>
>>  Then the Character changes must be changedInstance variable value was
>>> added in Collections.spur-ul.584.mcz it seems like
>>>
>>> On Thu, May 7, 2015 at 2:03 PM, Tobias Pape <Das.Linux at gmx.de> wrote:
>>>
>>>       On 07.05.2015, at 14:01, karl ramberg <karlramberg at gmail.com>
>>> wrote:
>>>
>>>       > I tried loading these changes manually but my image just crashes
>>> hard.
>>>       >
>>>       > If I try to add instance variable 'value' to Character I just
>>> get a debugger and no option to continue.
>>>       >
>>>       >
>>>
>>>       This may be because Character cannot have instance variables in
>>> Spur, IIRC…
>>>
>>>
>>>
>>>
>>
>>
>>
>
>
>
>


-- 
best,
Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20150507/e14cec0b/attachment.htm