apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject RE: Win32 pid_t semantics
Date Tue, 19 Dec 2000 22:55:51 GMT


> -----Original Message-----
> From: William A. Rowe, Jr. [mailto:wrowe@rowe-clan.net]
> Sent: Tuesday, December 19, 2000 4:55 PM
> To: 'Bill Tutt'; 'rassilon@list.org'
> Subject: RE: Win32 pid_t semantics
> 
> 
> Bill,
> 
>   Yes, and this has been a pain in the rear since a certain someone on
> http argued that there is only a simple int, and it was unfair to poor
> unix folk that something as simple as a process id must be a 
> transparent
> type.
> 
>   The proper way, and I should have screamed louder at the time, is to
> cache the handle -plus- the true pid.  I agree it is stupid 
> for us to use
> the HANDLE hProc over the id (given a single choice)... if 
> you want to 
> work up all the ramifications of using the idProc instead 
> please do so 
> and offer up a patch.  I don't see going back to a 
> transparent type here, 
> so if you can make sense of it, great :-)
> 
> Bill
> 
> > -----Original Message-----
> > From: Bill Tutt [mailto:rassilon@list.org]
> > Sent: Tuesday, December 19, 2000 4:33 PM
> > To: dev@apr.apache.org
> > Subject: Win32 pid_t semantics
> > 
> > 
> > Does anybody see the need for someone to create some patches to
> > threadproc/win32/proc.c to use real PIDs as opposed to the 
> per/process
> > HANDLEs?
> > 
> > It really isn't that difficul to go from a PID and a process object
> > handle. (OpenProcess)
> > 
> > Please, cc me in any replies, I'm not on the list atm.
> > 
> > Thanks,
> > Bill
> > 
> > 
> > 
> 

Mime
View raw message