Re: [RFC] Fix PosixProcess by porting VMProcess from Classpath...

[Casey Marshall <rsdio at metastatic dot org>]

>>>>> "David" == David Daney <ddaney@avtrex.com> writes:

David> It is said that:

David> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11801

David> Could be fixed by using VMProcess from Classpath instead of the
David> current PosixProcess.

David> There are several other problems that I have been experiencing
David> related to not reaping terminated or failed Processes that
David> would also be fixed.

David> Q1: Does this seem like a good idea?.

I think so. It will bring libgcj closer to Classpath, making both
easier to maintain.

David> Q2: Is someone else currently working on fixing this.

I used natPosixProcess.cc to make a CNI implementation of this class's
native methods, to try it out.

I'll attach what I did, but the POSIX part is trivial to implement.

David> Q3: Would it be acceptable to only fix this for PosixProcess
David> and assume that Win32 et al work and don't need fixing.

AFAIK there are only real implementations for POSIX and Win32, and a
stubbed-out one for eCos. Replacing ConcreteProcess would just mean
breaking Win32, which isn't too bad.

-- 
Casey Marshall || rsdio@metastatic.org

Attachment: natVMProcessPosix.cc
Description: Text document