Hi Igor, Le 28/07/2015 18:35, Igor Stasenko a écrit :
On 27 July 2015 at 18:31, Thierry Goubier <[email protected] <mailto:[email protected]>> wrote: Le 27/07/2015 15:42, Igor Stasenko a écrit : Nothing here is about things to fix, but rather about how impossible to debug event handling in system that runs and relies on very same events.. Esay Igor, just use a code event tracer as for example Jejak, and all you describe below won't happen. Or an event logger, or metalinks, or anything. I don't understand. You have the tools, why don't you use them? Perhaps because i unaware of them? And while i agree that using robust tools would help, i don't see how tools can help with "unkilling" an image which you just killed by own hand because of mistake you made :)
That one is hard to recover from. We can only trace (and use an external logging tool), or remote exception/debugging capacity.
About Jejak, i only found this: http://lists.gforge.inria.fr/pipermail/pharo-project/2012-July/067679.html
The current version is hosted on github, at http://github.com/ThierryGoubier/Jejak
Where i can read bout it? It sounds very useful indeed.
There is no real paper on it; I can send you the technical report I wrote on it in 2007 at UBO, but it focuses more on how it was implemented (especially optimised) for VisualWorks.
Thierry
