[squeak-dev] Compiler detox

Nicolas Cellier nicolas.cellier.aka.nice at gmail.com
Fri Sep 20 20:20:31 UTC 2013


Indeed, with newCompiler trick, there is normally no much need to subclass
Compiler, thus no need to override compilerClass/evaluatorClass.
Overriding parserClass is enough for playing with different syntax,
encoder, etc...

It's a bit less clear with Decompiler which is not in a healthy state.
I mean that measured in neuron power, maintenance is really costly.
Keep a good pile of sugar not to far from your computer if you want to dive
into it :)


2013/9/20 Eliot Miranda <eliot.miranda at gmail.com>

> Hi Tim,
>
>
> On Thu, Sep 19, 2013 at 7:27 PM, tim Rowledge <tim at rowledge.org> wrote:
>
>>
>> On 19-09-2013, at 6:20 PM, Nicolas Cellier <
>> nicolas.cellier.aka.nice at gmail.com> wrote:
>>
>> > There are more clean ups coming soon, when i'll figure out the right
>> order to not break the update stream:
>>
>> I'm not familiar enough with how the Compiler is structured these days
>> but thank you for working on tidying it up.
>> Suggestion; rather than specialised messages try a subclass of Compiler
>> that deals with interactive use and another for batch use. Unless that is
>> what it does already, in which case ignore me ;-)
>>
>
> One of the nice improvements is to ditch compilerClass, parserClass et al
> in favour of newCompiler, newParser, which can then allow exotic
> initialization etc.
> Compilation cues provide a much nicer way of providing input context to
> the compiler.
> The BytecodeEncoder backend now insulates the compiler from bytecode
> encoding and hence supports multiple bytecode sets (in use here at Cadence
> and needing to be more generally released when time allows).
>
> I think these are the big improvements over recent years.
>
>
>>
>> tim
>> --
>> tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
>> Error 13: Illegal brain function. Process terminated.
>>
>>
>>
>>
>
>
> --
> best,
> Eliot
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20130920/9c9c2c3d/attachment.htm


More information about the Squeak-dev mailing list