Over the years, I have looked at hdb, ptkdb, Eclipse EPIC and a few other variations. I realize that all code is not the same and what you work on may or may not be helped by what I describe. The problems were almost always identified when debugging serially. I found that fairly quickly after adopting these practices, that I had very few problems with the code with running across multiple threads/processes. In these cases, I would tweak my code slightly to run two serial processes where I could debug these types of issues. In these cases, I was either using a database or network client/server. This worked for me except when troubleshooting inter-process communications. I was already a pretty heavy user of Log::Log4Perl, so I used DEBUG and TRACE levels where needed when running in parallel to help identify any bugs there. I added a command line switch to control whether to run serially or parallel. I created an if construct that would run everything in a single threaded serial manner to allow me to step through and debug effectively. This behavior led me to change my coding patterns. Its probably been a couple of years since the last time I tried it. This can lead to a mess depending on the number of threads/processes. I have not found a way to use the perl debugger for threaded/forked apps outside of a Linux X Windows environment where the debugger will create separate xterms for each thread/process.

I do have some experience that may help you as you look. I sorry, but I don't have an answer for you.

Pearl ide free online