Re: Exception handling coroutine fix needed

On Jul 21, 2011, at 10:07 PM, Jan-Paul Bultmann wrote:

> 
> Hey,
> 
> The fix to this is rather simple, I have one implemented on the Io level in my fork. A rather naive approach that I am afraid is not the most efficient.
> Problem is, it won't handle all cases, since this would require it to be done on C level, which I am still getting familiar with.
> If you want I can write a fix for that, but it might take me a day.
> 
> But even if this is fixed, there is still a second problem.
> Coroutine setMessageDebugging does not propagate down to it's children. So the debugger won't help you there. (Not a problem if you don't use it though :))
> I am afraid the naive approach of checking the value recursively upwards though the Coroutine graph won't cut it since the mechanism relies on replacing the message sending function
I have to correct myself on this one. Due to a hack awaiting decision on how to resolve it since 2007 the debugging perform function will always be called.
There is a boolean controlling wether to go into debug mode or not. :D

Cheers Jan
> This is something I would consider out of my league.
> 
> Both problems are probably easy to solve for the old hands here :)
> 
> A bounty system is something I would really encourage, I tried it myself for the documentation :) but no one replied or even commented :(
> 
> Oh and btw do you have something more debuggerish than vmWillSendMessage :) , would be quite interested as I am currently working on a graphical debugger :)
> 
> Cheers Jan
> 
> On Jul 21, 2011, at 2:00 PM, petri rokka wrote:
> 
>>  
>> Hello,
>> 
>> I would really need a fix to the problem I have with debugging code that has exception handling.
>> 
>> Problem is that  try-block will create a new coroutine for itself so io-interpreter fails to do the right thing if i try to pause and resume coroutine that has the try-block.
>> I can't pause/resume the code that is in the try block by trying to pause/resume its parent. 
>> 
>> Is it possible speed up this fix by paying somehow and how much time/money would it require? Is there a bounty system for io-language stuff?
>> 
>> Kind regards,
>> 
>> Petri
> 
> 
>