[Pharo-project] Transcript rant

Stéphane Ducasse stephane.ducasse at inria.fr
Tue Apr 19 09:14:44 CEST 2011


I know all that :) I favor changes. 
Now theadSafeTranscript was thread safe.
How can I get the old transcript because I need to get fast for debugging an ugly bug.

> From the release notes of CUIS:
> "New Transcript
>  Fast (Calls return quickly)
>  Fast (Updates are immediate)
>  Thread safe
>  Can log to file in addition to Display
>  Display output is optional (for headless images)
>  Can work on a bare bones image (Doesn't need UI, dependency, events,
> processes)
>  Optional morphic window for seamless integration in Morphic
>  UI framework agnostic. Can be used to debug Morphic, even with the
> morphic window open, without crashing the image!
>  No longer a text editor. Display is read only."
> 
> The idea of the current Transcript are, not to use a standart widget
> for decoupling from Morphic, in case of errors the transcript still
> remains functional.
> 
> I propose adding a new transcript morph, that has the previous
> functionality ( cut and paste ), but keep the new transcript model (
> as in the ISSUE 3948).
> 
> Fernando
> 
> On Sun, Apr 17, 2011 at 6:59 PM, Stéphane Ducasse
> <stephane.ducasse at inria.fr> wrote:
>> How can I get the old transcript?
>> 
>> Stef
>> 
>> On Apr 7, 2011, at 9:58 AM, Fernando Olivero wrote:
>> 
>>> I propose some modifications to the Transcript as a (singleton) class,
>>> and revert to the use of TranscriptModel.
>>> 
>>> See the latest post on 3948:  Transcript and ThreadedTranscript needs
>>> to be merged.
>>> 
>>> Fernando
>>> 
>>> On Wed, Apr 6, 2011 at 10:39 PM, Igor Stasenko <siguctua at gmail.com> wrote:
>>>> On 6 April 2011 20:24,  <csrabak at bol.com.br> wrote:
>>>>> Em 06/04/2011 09:07, Igor Stasenko <siguctua at gmail.com> escreveu:
>>>>>>> On 6 April 2011 13:54, Fernando Olivero <fernando.olivero at usi.ch>
>>>>>    wrote:
>>>>> 
>>>>>>> Hi Stef, i will as an initial step take Igor's suggestion and make
>>>>>>> Transcript a global wich  points to an instance of TranscriptModel
>>>>>>> ( the new name i propose for the current class taken from CUIS).
>>>>>>> I still  believe that composition  is better then  inheritance for
>>>>>>> handling different behaviors of the Transcript, so i argue against
>>>>>>> making it a subclass of WriteStream and making it polymorphic with
>>>>>>> a write stream instead.
>>>>>>> 
>>>>>>  That's totally  not my concern. As  long as thing  which sits there
>>>>>>  responds to writestream protocol , i don't care about rest.
>>>>>> 
>>>>> 
>>>>> Would it make sense then to lift the methods from {write}stream and
>>>>> put then in a Trait so the composition is made through reuse of code?
>>>>> 
>>>>> My 0.0199999....
>>>> 
>>>> hehe.. yes, but if to follow this road every time you need to do small
>>>> improvement,
>>>> then at some day you will discover that you rewriting whole system
>>>> from scratch :)
>>>> 
>>>>> 
>>>>> --
>>>>> Cesar Rabak
>>>> 
>>>> 
>>>> --
>>>> Best regards,
>>>> Igor Stasenko AKA sig.
>>>> 
>>>> 
>>> 
>> 
>> 
> 




More information about the Pharo-project mailing list