[gpsd-commit-watch] [SCM] GPSD branch, master, updated. release-3.16-105-gad68497

[Eric S. Raymond]

This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GPSD".

The branch, master has been updated
       via  ad684978ffad23c8bee753eb90a9fa67dd07682d (commit)
       via  f3f0ba3f60f891c0e90aac15a7e928b67b0ece97 (commit)
       via  42dc44b6c27bb50e999c6b289c42c119f3340dc0 (commit)
       via  45d873ea3a6e2122900d2b785b7cd131a7759569 (commit)
      from  1829ba9c5106340b83be4a6dc5d69b37421bc98c (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit ad684978ffad23c8bee753eb90a9fa67dd07682d
Author: Fred Wright <address@hidden>
Date:   Tue Mar 15 16:52:12 2016 -0700

    Switches parallel regression tests from UDP to TCP.
    
    Although UDP is a bit faster than TCP (albeit negligibly so on most
    platforms), the method for assigning "fake data" ports is less robust
    for UDP than for TCP.  This leads to occasional "can't connect" errors
    with parallel regression tests.  The fix is just to use TCP instead.
    
    TESTED:
    Successfully ran parallel "scons build-all check" on OSX, Linux,
    FreeBSD, OpenBSD, and NetBSD, both with and without coveraging.

commit f3f0ba3f60f891c0e90aac15a7e928b67b0ece97
Author: Fred Wright <address@hidden>
Date:   Tue Mar 15 16:52:11 2016 -0700

    Removes build-time dependency on 'python2'.
    
    Since some systems lack a 'python2' command, the switch to 'python2'
    in the shebang lines can cause failures.  This fix gets around it by
    explicitly invoking $PYTHON for all Python tools (except valgrind.py,
    which already gets the versioned shebang line).  It also exports
    $PYTHON to the environment, so that regress-driver can do the same
    instead of defaulting to 'python2'.
    
    If scons becomes Python3-compatible before GPSD does, or if GPSD
    becomes pickier about Python versions than scons, some additional code
    will be needed to find a suitable Python version, but for now just
    using the same Python as scons is adequate (and is already what was
    being done for some programs, just not all).
    
    This does *not* address the issue of the *installed* Python programs
    not working without 'python2'.  That's left for a separate change.
    
    TESTED:
    Verified that "scons build-all check" now works without a 'python2' in
    the command path, both with and without coveraging.

commit 42dc44b6c27bb50e999c6b289c42c119f3340dc0
Author: Fred Wright <address@hidden>
Date:   Tue Mar 15 16:52:10 2016 -0700

    Improves coverage testing of fake.py.
    
    This adds three "regression" tests which run passthrough.log in all
    three transport modes, for more complete coverage in fake.py.  The
    target for all three is 'gpsfake-tests'.
    
    Although the original motivation for 'test-nondaemon' was to exclude
    slow-running tests, and the gpsfake tests are fairly quick, it would
    be misleading to classify them as "nondaemon".  So a new 'test-quick'
    target is added which includes these but not the full set of
    regression tests.
    
    TESTED:
    Ran all relevant test targets, with and without coveraging.

commit 45d873ea3a6e2122900d2b785b7cd131a7759569
Author: Fred Wright <address@hidden>
Date:   Tue Mar 15 16:52:09 2016 -0700

    Adds support for Python coveraging.
    
    This extends the "coveraging" option to apply coveraging to any Python
    programs executed during the build.  It adds a new option
    "python_coverage" to allow configuring the command to be used for
    applying coveraging to Python programs, with a default value of
    "coverage run" (which is typically appropriate, provided that the
    coverage package is installed).  Setting this to the null string
    disables Python coveraging.
    
    If "python_coverage" has its default value and parallel builds are
    enabled, "--parallel-mode" is automatically appended to the coverage
    command.  If a non-default coverage command is used, it's the user's
    responsibility to handle parallelism appropriately.
    
    Note that this does not apply coveraging to generated Python code.
    
    This also adds the coverage result files to .gitignore, including the
    C coverage files which weren't there previously.
    
    TESTED:
    Ran "scons check" as well as sample cases of regress-driver -b and -v,
    both with and without coveraging.  Also verified that it works with the
    full path to the tool supplied as the python_coverage option.

-----------------------------------------------------------------------

Summary of changes:
 .gitignore     |    5 +++++
 SConstruct     |   45 ++++++++++++++++++++++++++++++++++++++++-----
 build.txt      |   16 ++++++++++++----
 regress-driver |    6 +++---
 4 files changed, 60 insertions(+), 12 deletions(-)


hooks/post-receive
-- 
GPSD