mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Schwartzmeyer <>
Subject Review Request 66892: Added `SubprocessTest.PipeLargeOutput`.
Date Tue, 01 May 2018 21:24:19 GMT

This is an automatically generated e-mail. To reply, visit:

Review request for mesos, Akash Gupta, John Kordich, Joseph Wu, and radhika jandhyala.

Repository: mesos


This new test checks exercises our `Subprocess::PIPE()` logic more
thoroughly by specifically piping enough data such that a single
memory page is insufficient to hold it. This is especially important
on Windows because the OS-allocated buffer is the size of one memory
page. On Windows, it also tests that the expected end-of-file
condition, `ERROR_BROKEN_PIPE`, is set.

We also fix another use of `os::WindowsFD` in place of `int_fd`, the
public name of the type; added an important note about why we hold
onto the process's handle on Windows; pass the file descriptor structs
by const-ref instead of value; and finally check the return value
`os::close()` in `createChildProcess`.


  3rdparty/libprocess/src/subprocess.cpp 898326360d6b4f0a50d6ef3f7c86141d0aa70438 
  3rdparty/libprocess/src/subprocess_windows.hpp 4cc5675621f0846229e9358d70412f44138e7904

  3rdparty/libprocess/src/tests/subprocess_tests.cpp be99bd6a88b2abcabed8d4d16763c2e29f38eb14



`ctest -V` on Linux and Windows (also running `make check` with Autotools).


Andrew Schwartzmeyer

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message