The test case in EPSCanvasTest (introduced today with MorphicExtras-hjh.210.mcz)

shows what happens if you move

     FlattenEncoder subclass: #Canvas

to
    Object subclass: Canvas

See comment there.

--Hannes

On 9/30/17, H. Hirzel <hannes.hirzel at gmail.com> wrote:
> Tobias, thank you for referring to Marcel Weiher as the author of the
> Postscript support in Squeak.
> This helped me to find the following email with comments about the
> Encoder hierarchy
>
> <citation>
> Marcel Weiher
> <marcel.weiher at gmail.com>	Wed, Feb 27, 2013 at 9:14 PM
> To: The general-purpose Squeak developers list
> <squeak-dev at lists.squeakfoundation.org>
>
> On Feb 25, 2013, at 20:22 , Nicolas Cellier
> <nicolas.cellier.aka.nice at gmail.com> wrote:
>>
>> Great, that confirms my own analyze.
>> But the parallel hierarchy PrintEncoder StoreEncoder and
>> PropertyListEncoder are presumably not used, and were just experiments
>> right?
>
> Pretty sure, yes.
>
> </citation>
>
>
> So it seems that
>
> NullEncoder
>     FlattenEncoder
>         ByteEncoder
>             PrintableEncoder
>
> are  a means to structure the code for the subclass PostscriptEncoder
> and thus to help to understand how the code works (see attached screen
> shot).
>
> NullEncoder
>     FlattenEncoder
>         ByteEncoder
>             PrintableEncoder
>                 PostscriptEncoder
>
>
> Note that Canvas and thus then PostscriptCanvas is a subclass of
> FlattenEncoder.
> As Nicolas Cellier points out only PostscriptCanvas and subclasses use
> the NullEncoder and FlattenEncoder protocols. For example
>     aMorph asEPS
> calls
>     ^ EPSCanvas morphAsPostscript: self rotated: false.
> which at the end does
>     ^target contents.
> which is implemented in NullEncoder.
>
> The question thus is how much does PostScriptCanvas rely on the
> NullEncoder and FlattenEncoder protocols?
>
> It might be quite easily possible  to reimplement the necessary
> methods in PostScriptEncoder and then make Canvas a subclass of
> Object.
>
> This will allow Canvas and subclasses to be put into a separate
> category 'Graphics-Canvas' in the base system.
>
> PostScript support would then be something added on top of it.
>
>
> To summarize:
> --------------------
>
> The refactoring idea brought up for discussion is thus
>
> 1. Push the necessary NullEncoder and FlattenEncoder instance variable
> 'target' and the necessary methods down to PostscriptCanvas
>
> 2. Make Canvas a subclass of object thus separating then Encoder
> hierarchy from the Canvas hierarchy.
>
> 3. Put Canvas and some subclasses into a category 'Graphics-Canvas'
>
> Regards
>
> Hannes Hirzel
>
>
> --Hannes
>
> On 9/30/17, Tobias Pape <Das.Linux at gmx.de> wrote:
>>
>>> On 30.09.2017, at 01:42, David T. Lewis <lewis at mail.msen.com> wrote:
>>>
>>> According to the update stream history at
>>> http://files.squeak.org/history,
>>> NullEncoder and friends were added in  1466EncodingFilters-MPW.cs by
>>> MPW:
>>>
>>>   Object subclass: #NullEncoder
>>>   NullEncoder subclass: #FlattenEncoder
>>>   FlattenEncoder subclass: #ByteEncoder
>>>   ByteEncoder subclass: #PrintableEncoder
>>>   PrintableEncoder subclass: #PrintEncoder
>>>   PrintableEncoder subclass: #PropertyListEncoder
>>>   PrintableEncoder subclass: #StoreEncoder
>>>
>>> According to SqueakMap, MPW is Mark McCahill (mccahill at duke.edu).
>>
>> No, thats Marcel Weiher. He did a quite a lot Squeak/Postscript Stuff.
>> I CC'ed him.
>>
>> Marcel, can you comment on the Encoder Hierarchie?
>> (Full thread here)
>>
>> Best regards
>> 	-Tobias
>>
>>
>>>
>>> According to Wikipedia, Mark McCahill at Duke is this guy:
>>> https://en.wikipedia.org/wiki/Mark_P._McCahill
>>>
>>> According to the method time stamps, these classes date back to January
>>> 1901, so probably actually 2001.
>>>
>>> Dave
>>
>>
>>
>>
>