> On Aug. 22, 2016, 6:39 p.m., Joseph Wu wrote:
> > 3rdparty/stout/include/stout/os/windows/shell.hpp, lines 114-115
> > <https://reviews.apache.org/r/51217/diff/3/?file=1479628#file1479628line114>
> >
> >     I'm not sure I understand what this comment is saying.  Are you 
> > explaining the reason behind the `exit(...)`?

On Posix: 
After a process X calls execlp to execute the application Y, the application Y 
runs in the process originally created to run the application X. Therefore the 
parent of X can use PID/pipes from the invocation of X to control the 
application Y
On Windows: 
After a process X calls execlp to execute the application Y, the application Y 
start a new process Y. The process X terminates at this time (it will never 
return from execlp https://msdn.microsoft.com/en-us/library/1kxct8h0.aspx). The 
parent of X cannot use X process handle to control the process Y. With this 
change the parent of X can use the X process handle to wait for the termination 
of Y, as in POSIX.


- Daniel


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/51217/#review146398
-----------------------------------------------------------


On Aug. 19, 2016, 5:38 p.m., Daniel Pravat wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/51217/
> -----------------------------------------------------------
> 
> (Updated Aug. 19, 2016, 5:38 p.m.)
> 
> 
> Review request for mesos, Alex Naparu, Artem Harutyunyan, Alex Clemmer, 
> Joseph Wu, and Michael Park.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added `os::execlp' in `stout` library.
> 
> 
> Diffs
> -----
> 
>   3rdparty/stout/include/stout/os/posix/shell.hpp 
> 1d73ae5fa7182ba433eaee0cc0034af6b36d77db 
>   3rdparty/stout/include/stout/os/windows/shell.hpp 
> c6e141aba0abe2c7fe5410e867f7db47d632e765 
> 
> Diff: https://reviews.apache.org/r/51217/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Daniel Pravat
> 
>

Reply via email to