Wolfgang Jährling <[EMAIL PROTECTED]> writes: > I think you're right. Instead of programming around Rubys failures, > Ruby should be fixed. Unless it is fixed, we will limit ourself to > single threaded translators. Since it really is an important issue > for Ruby in general, it certainly will be fixed sooner or later.
The easiest way to fix it is to have the Big Interpreter Lock. This is the approach taken by Python and Pike. The idea is to have a mutex that the interpreter keeps locked all the time while it is interpreting the byte code instructions, manipulating object state, etc. It only releases the lock while doing blocking system calls like read, connect and select. This way, the only code that needs to be aware of threading issues is the code that decides to release the lock. /Niels _______________________________________________ Help-hurd mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/help-hurd
