[gpsd-commit-watch] [SCM] GPSD branch, master, updated. release-3.16-72-gf51c403

[Jon Schlueter]

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  f51c4031ee41087840e9f14a1315e101bbea59f1 (commit)
       via  89ea20912d19ef83fc615a5f31914b1e525e73c6 (commit)
       via  01bc4db1bda42e980fbc4ee1885e48e408939829 (commit)
       via  fdeccf80ed70820d74fcffd0b0f1443354b1338e (commit)
      from  9e8ec3a51760191fa1a776da89bffae807378092 (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 f51c4031ee41087840e9f14a1315e101bbea59f1
Author: Fred Wright <address@hidden>
Date:   Sun Feb 28 19:39:12 2016 -0800

    Parallelizes gps-regress and gps-makeregress.
    
    This change turns each of these tests into a set of individual
    single-case targets, so that they can be run in parallel.  It still
    uses regress-driver, but with a single logfile for each instance.  It
    uses the new -q option to reduce the verbosity when run this way.  The
    "herald" using gpsfake -T is turned into a separate target that's a
    dependency of the others, so that it always runs first.
    
    Because consuming a large number of PTYs would be antisocial (not to
    mention possibly not working), it uses UDP mode.  TCP mode would also
    work, but is slightly slower (albeit negligibly except on FreeBSD).
    
    Since the output format is deficient in this mode (including not
    having the summary at the end), it runs the old way whenever multiple
    jobs aren't enabled.  But the single-case tests are still generated,
    to permit running single cases (or sets of cases) by name.
    
    It might be more conservative not to parallelize gps-makeregress, but
    one can get that effect simply by not enabling multiple jobs.
    
    This change also removes an obsolete comment about clearing the SHM
    segments before the run.
    
    TESTED:
    Ran gps-regress and gps-makeregress, in both serial and parallel
    modes, on three versions of OSX, as well as Linux, FreeBSD, and
    OpenBSD.

commit 89ea20912d19ef83fc615a5f31914b1e525e73c6
Author: Fred Wright <address@hidden>
Date:   Sun Feb 28 19:39:11 2016 -0800

    Adds -q (quiet) option to regress-driver.
    
    This allows reducing the verbosity when running multiple single-case
    instances of regress-driver (e.g., for parallelism).  It is not the
    same as redirecting stderr, since the "Processing" messages also go to
    stderr, and are useful in this case.
    
    TESTED:
    Ran regress, build, and view modes, with and without -q, and verified
    expected output.
    
    Signed-off-by: Jon Schlueter <address@hidden>

commit 01bc4db1bda42e980fbc4ee1885e48e408939829
Author: Fred Wright <address@hidden>
Date:   Sun Feb 28 19:39:10 2016 -0800

    Makes fake.py's TCP and UDP port assignments parallel-compatible.
    
    The current method for assigning ports uses a counter initialized to a
    constant.  Although this works fine for multiple sessions managed by a
    single instance of fake.py, it fails miserably when running multiple
    parallel instances of fake.py.  The fix is to allow the OS to assign
    the port numbers, since it's guaranteed to pick unused ports.
    
    In the TCP case, this is simply a matter of specifying 0 as the port,
    and then extracting the actual assigned port number with
    getsockname().
    
    In the UDP case, it's more complicated since the port number being
    picked is actually for *gpsd's* end, which can't be done in a
    straightforward manner.  The workaround, which was already being used
    to pick the control-socket port for gpsd, is to bind a socket with a
    reusable address, close it, and then assume that the port will remain
    available until gpsd grabs it.  This change turns the existing code to
    do that into a function, with the socket type now being specifiable.
    
    TESTED:
    Ran all daemon tests in both TCP and UDP modes, on three versions of
    OSX as well as Linux, FreeBSD, and OpenBSD.  Used default WRITE_PAD
    values except on OSX, where it was reduced to 1ms to save time.
    
    Signed-off-by: Jon Schlueter <address@hidden>

commit fdeccf80ed70820d74fcffd0b0f1443354b1338e
Author: Fred Wright <address@hidden>
Date:   Sun Feb 28 19:39:09 2016 -0800

    Fixes gpsfake -v to use stderr instead of stdout.
    
    TESTED:
    Ran a quick gpsfake -v with >/dev/null, and observed output.
    
    Signed-off-by: Jon Schlueter <address@hidden>

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

Summary of changes:
 SConstruct     |   38 +++++++++++++++++++++++++++++---------
 gps/fake.py    |   37 ++++++++++++++++++++++---------------
 gpsfake        |    2 +-
 regress-driver |   13 ++++++++-----
 4 files changed, 60 insertions(+), 30 deletions(-)


hooks/post-receive
-- 
GPSD