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

Wed Jul 14 19:06:00 GMT 2004

Mohan Embar wrote:
>Hi People,
>>>>>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.
>>>>>>As far as I can tell, the two problems in the Bugzilla report
>don't affect Win32. (I haven't rigorously tested this, though,
>just desk-checked the code and consulted the MSDN documentation.)
>What's more, the Win32 implementation doesn't require the
>additional dedicated process-spawning thread.
>>I can understand the desire to consolidate the codebase, though....
>>I think we'll be able to keep Process as an abstract base class, and use 
Win32Process as is (perhaps getting rid of the ConcreteProcess class 
name hack, though). Then, its just a matter of having Runtime.exec() 
return the right one depending on which platform its running on. One way 
to do that would be a _Jv_platform_runtime_exec() call in posix.h/win32.h.
Regards
Bryce