Michael Kaply wrote:
>
> We have the fix for this - we are trying to get the NSPR folk to check it in.
Thanks.
> It can't be debugged,
There are good days and bad days. A -g (no -O2) build is generally okay
in the VA debuggers (meaning you get Source View), until you step into
xptcall, where they normally crash. Yet often where one (iscdebug vs.
idebug) crashes, the other doesn't, which helps if you have both and
much patience.
I think it's important to maintain the emx PM build if for no other
reason than it's free, and I'm talking not dollars and cents but open
source, right down to the tools. The one unfortunate exception is the
debugger, and there the real solution is to add OMF support to gdb.
Also, gcc just keeps getting better and better, and may well be
outperforming icc before PM's time runs out.
> There are also bugs in select() on EMX that were/are causing PSM problems as
> well.
Dimitry submitted a multithread select patch for that which wtc checked
it in on March 1st.
h~
