[gambit-list] Multi-thread debugging

Derek Peschel dpeschel at eskimo.com
Fri Apr 25 18:52:21 EDT 2008


On Fri, Apr 25, 2008 at 10:51:37AM -0400, Marc Feeley wrote:
>                                               Note that each thread  
> has its own REPL but there is only one console, so the accesses to the  
> console are serialized.

>                                                                 It is  
> sometimes useful to force this switch from one REPL to the other by  
> entering a dummy expression.

Hopefully we can continue this discussion here, or on the Wiki, to flesh
out the documentation.

Aside from the quitting and exception-handling behaviors described in the
manual, does the primordial thread have any special properties?

Do REPL level changes affect the thread state (the things you describe --
number of threads, current or future ownership of console, etc.) in any way?
I mean the level changes themselves, not the errors or interrupts that cause
them.

Using the Java REPL and a program I wrote to test exception handling,
sometimes I've seen ^C interrupts start a debugger in the thread I want,
and sometimes I've seen them switch threads or possibly start a new thread.
What might affect ^C's behavior like this?  I can send you some test cases
but it will be a few days (I'm out of town).

-- Derek



More information about the Gambit-list mailing list