[BUG][FIX]StrikeFontSet>>maxWidth - array out of bounds

Peter van Rooijen peter at vanrooijen.com
Thu Dec 23 21:29:14 UTC 2004


Peter van Rooijen wrote:

[hehe, previous message sent to squeak-dev courtesy of Mozilla Thunderbird]

Cheers everyone, happy holidays and a great new year!

Peter

> Trygve Reenskaug wrote:
> 
>> At 17:56 21.12.2004, Ned Konz wrote:
>>
>>> On Monday 20 December 2004 12:03 pm, Trygve Reenskaug wrote
>>> >     StrikeFontSet>>maxWidth
>>> >          ^ (fontArray at: (2 min: fontArray size)) maxWidth.
>>>
>>> I posted a fix for this on January 8th. I didn't bother to put it into
>>> Connectors 1.9 because I was coming out with Connectors 2.
>>>
>>> What problems did you have using Connectors 2?
>>
>>
>>
>>
>> The short answer: BabySRE doesn't work with Connectors 2.
>>
>> The long answer is: Of course, it didn't. BabyUML is a hodgepodge of 
>> Jacaranda, Connectors and my own stuff. Jacaranda mainly by copying 
>> code before maltreating it. Connectors the same + subclassing. And my 
>> maltreatment is surely in conflic with your improvements.
>>
>> Why not use Connectors and Jacaranda as they are? I am a toolmaker and 
>> need rapid prototyping of new ideas to discover what works and what 
>> doesen't. I am too stupid to understand the static and dynamic 
>> architecture of Connectors. I see that it is a very smart and powerful 
>> piece of code, and I have probably ruined it with my clumsy 
>> modifications. As it is, I expect I have to write my own diagramming 
>> code from scratch because I MUST have full control for my experiments.
> 
> 
> Trygve,
> 
> Just to let you know I agree 100% with your desire to have full control 
> of the code you base your work on. In my own work I have made a point of 
> it for quite some time, not to depend on someone else's code if I can 
> help it.
> 
> This may be considered heresy and bad form (which is part of why I'm not 
> writing this to the list ;-)), but it's the only way to gain reliability 
> in the absence of guaranteed capabilities (which are simply not 
> something that is currently available).
> 
> Good luck and I look forward to checking out your BabyUML some time.
> 
> Cheers,
> 
> Peter
> 
>> More than ever, I feel the need for what I try to do with BabyUML: 
>> Code that reveals more about itself than the current classes and methods.
>>
>> Cheers
>> --Trygve
>>
>>
> 
> 
> 
> 




More information about the Squeak-dev mailing list