I once did a VW program that logged the messages received by the 
instances of a class. It was easy to code; the challenge was filtering. 
I abandoned the program because its long logs didn't tell me anything 
new. (Logging even a few seconds of /all /messages is prohibitive, there 
are too many of them).

My principle was simple. AFAICR, the code went something like this:

 1. All messages handled by a class pass through its /methodDict/. Make
    a copy of this dict, say /methodDictCopy/
 2. Automatically create a new set of methods for the original
    /methodDict /according to their number of arguments:
    aClass>>perform: aSymbol
    /or /aClass>>perform: aSymbol with: firstObject etc.
    In this method,
          Log message to a file, if wanted.
          Look up the requested method in /methodDictCopy.
    /Execute it.

As I said above, the scheme worked, but it was never used in practice 
because I didn't find it useful. Its weakness is that it only logs 
messages received by instances of one class. An advantage is that it 
does not require any changes to the VM. What's needed is a log of the 
messages flowing between selected instances of many classes. The above 
scheme could be extended to do that, but I thought the description 
belonged to the design stage. May be it could be a reverse engineering tool?

I used the OORAM role modeling tool to move the description of the 
message flows from the execution to the design stage. This tool helped 
the programmer design the answer to three questions: What are the 
objects? How are they interconnected? What do they do? With these three 
answers, the implementation with classes was straight-forward.

I later moved the description from the design stage to the coding stage 
with the /DCI programming paradigm/.

Today, I code the flow of messages explicitly. I have also extended the 
Squeak universe of objects with all objects available through an IoT. 
Alan's object orientation is now literally realized in practice: “/…its 
semantics are a bit like having thousands and thousands of computers all 
hooked together by a very fast network./” But that is another story.

  TRee

On 2020-07-10 02:13, tim Rowledge wrote:
> I'm looking for some good message logging stuff - something a little more flexible that `Transcript show:` level - and so far the best I can see is a rather old package on squeaksource named 'Logging' originally by Keith Hodges but with quite a few updates by Mat Fulmer.
>
> I wonder if anyone is using it in a current image? It seems to rely on some process specific variable code that *looks* like has been incorporated into the trunk. I don't want to spend too much time peering at it is somebody else already knows whether it is unusable now or better yet still 'load and go'
>
> Or indeed, it somebody has a pointer to sometihng newer/better/shinier?
>
> tim
> --
> tim Rowledge;tim at rowledge.org;http://www.rowledge.org/tim
> Ubi dubium ibi libertas
>
>

-- 

/The essence of object orientation is that objects collaborateto achieve 
a goal. /
Trygve Reenskaug mailto: trygver at ifi.uio.no <mailto:%20trygver at ifi.uio.no>
Morgedalsvn. 5A http://folk.uio.no/trygver/
N-0378 Oslo http://fullOO.info
Norway                     Tel: (+47) 468 58 625

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20200710/ef54f278/attachment.html>