You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Alex Clemmer <cl...@gmail.com> on 2016/04/24 01:40:56 UTC
Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/
-----------------------------------------------------------
Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
Repository: mesos
Description
-------
Libprocess: Implemented `subprocess_windows.cpp`.
Diffs
-----
3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
Diff: https://reviews.apache.org/r/46608/diff/
Testing
-------
Thanks,
Alex Clemmer
Re: Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
Posted by Michael Park <mp...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/#review131235
-----------------------------------------------------------
3rdparty/libprocess/src/subprocess_windows.cpp (lines 305 - 352)
<https://reviews.apache.org/r/46608/#comment195141>
(1) According to the "Security Remarks" section of https://msdn.microsoft.com/en-us/library/windows/desktop/ms682425(v=vs.85).aspx, it's recommended that we set the `lpApplicationName` parameter, rather than embedding the module name inside the `lpCommandLine` parameter since we need to make sure that the modulen name in `lpCommandLine` needs to be wrapped in quotes in case of spaces in the name. We don't seem to be doing this correctly.
(2) `argumentsEscaped` is `new`ed, but not `delete`d, and according to the documentation for `CreateProcess`, it does not call `delete` on `lpCommandLine`. It looks like we're leaking?
(3) Are we required to escape the quotes like we're doing here for `argumentsEscaped`...? I don't see anything like this in the documentation.
What I would prefer to do here I think is to take a `argv` by value, then do:
```cpp
string program = argv[0];
argv.erase(argv.begin());
string args = strings::join(" ", argv);
... = CreateProcess(
program.data(),
args.data(),
...);
```
If we wanted to avoid using `lpApplicationName` for whatever reason, how about:
```cpp
string program = "\"" + argv[0] + "\"";
argv.erase(argv.begin());
string args = strings::join(" ", argv);
string cmd = strings::join(" ", program, args);
... = CreateProcess(
NULL,
cmd.data(),
...);
```
- Michael Park
On April 23, 2016, 11:41 p.m., Alex Clemmer wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/46608/
> -----------------------------------------------------------
>
> (Updated April 23, 2016, 11:41 p.m.)
>
>
> Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
>
>
> Bugs: MESOS-3637
> https://issues.apache.org/jira/browse/MESOS-3637
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Libprocess: Implemented `subprocess_windows.cpp`.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/46608/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Alex Clemmer
>
>
Re: Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
Posted by Michael Park <mp...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/#review130974
-----------------------------------------------------------
3rdparty/libprocess/src/subprocess_windows.cpp (lines 534 - 539)
<https://reviews.apache.org/r/46608/#comment194941>
A few comments here.
(1) I think we can simply this to:
```
string env;
foreach (const string& key, const string& value, environment.get()) {
env += key + '=' + value + '\0';
}
```
then pass `env.data()` to `createChildProcess`.
(2) Currently, if the string is larger than 32767 bytes, `createProcessEnvironment` returns `NULL`. We subsequently pass that result to `createChildProcess`, which result in the following behavior:
> lpEnvironment [in, optional]
A pointer to the environment block for the new process. If this parameter is NULL, the new process uses the environment of the calling process.
Is this what we want? I think we can definitely incorporate whatever behavior we need if we go beyond the 32767 bytes, but I'm not sure exactly what behavior is expected here.
- Michael Park
On April 23, 2016, 11:41 p.m., Alex Clemmer wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/46608/
> -----------------------------------------------------------
>
> (Updated April 23, 2016, 11:41 p.m.)
>
>
> Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
>
>
> Bugs: MESOS-3637
> https://issues.apache.org/jira/browse/MESOS-3637
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Libprocess: Implemented `subprocess_windows.cpp`.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/46608/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Alex Clemmer
>
>
Re: Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
Posted by Michael Park <mp...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/#review130682
-----------------------------------------------------------
3rdparty/libprocess/src/subprocess_windows.cpp (line 77)
<https://reviews.apache.org/r/46608/#comment194535>
Indent 2 spaces.
3rdparty/libprocess/src/subprocess_windows.cpp (line 79)
<https://reviews.apache.org/r/46608/#comment194537>
`s/foreach(/foreach (/`
3rdparty/libprocess/src/subprocess_windows.cpp (line 153)
<https://reviews.apache.org/r/46608/#comment194534>
Shouldn't we just `return duplicate;` here? Falling through and "leveraging" the `return handle` on the next line seems like a bad pattern to set.
- Michael Park
On April 23, 2016, 11:41 p.m., Alex Clemmer wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/46608/
> -----------------------------------------------------------
>
> (Updated April 23, 2016, 11:41 p.m.)
>
>
> Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
>
>
> Bugs: MESOS-3637
> https://issues.apache.org/jira/browse/MESOS-3637
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Libprocess: Implemented `subprocess_windows.cpp`.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/46608/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Alex Clemmer
>
>
Re: Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
Posted by Daniel Pravat <dp...@outlook.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/#review130852
-----------------------------------------------------------
3rdparty/libprocess/src/subprocess_windows.cpp (line 348)
<https://reviews.apache.org/r/46608/#comment194756>
This is not requered in the current codebase.
3rdparty/libprocess/src/subprocess_windows.cpp (line 606)
<https://reviews.apache.org/r/46608/#comment194759>
Since the process is not in suspended there is no need to resume here.
3rdparty/libprocess/src/subprocess_windows.cpp (line 607)
<https://reviews.apache.org/r/46608/#comment194760>
The handle for the thread can be closed in the createChildProcess()
- Daniel Pravat
On April 23, 2016, 11:41 p.m., Alex Clemmer wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/46608/
> -----------------------------------------------------------
>
> (Updated April 23, 2016, 11:41 p.m.)
>
>
> Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
>
>
> Bugs: MESOS-3637
> https://issues.apache.org/jira/browse/MESOS-3637
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Libprocess: Implemented `subprocess_windows.cpp`.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/46608/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Alex Clemmer
>
>
Re: Review Request 46608: Libprocess: Implemented
`subprocess_windows.cpp`.
Posted by Alex Clemmer <cl...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/46608/
-----------------------------------------------------------
(Updated April 23, 2016, 11:41 p.m.)
Review request for mesos, Alex Naparu, Daniel Pravat, Artem Harutyunyan, Joris Van Remoortere, Michael Park, M Lawindi, and Yi Sun.
Bugs: MESOS-3637
https://issues.apache.org/jira/browse/MESOS-3637
Repository: mesos
Description
-------
Libprocess: Implemented `subprocess_windows.cpp`.
Diffs
-----
3rdparty/libprocess/src/subprocess_windows.cpp PRE-CREATION
Diff: https://reviews.apache.org/r/46608/diff/
Testing
-------
Thanks,
Alex Clemmer