summaryrefslogtreecommitdiff
path: root/gps
Commit message (Collapse)AuthorAgeFilesLines
* Somewhat more effective asciifying in the Python client.Eric S. Raymond2011-02-221-14/+14
|
* Assume any 'satellites' attribute in JSON is a JSON array.Eric S. Raymond2011-02-221-1/+2
|
* Typo fixes.Eric S. Raymond2011-02-211-1/+1
|
* Greg Troxel says CLOSE_DELAY needed to go to 0.2 on his sytem to avoid trouble.Eric S. Raymond2011-02-181-2/+3
|
* Garbage-collect the client-side *SET flags.Eric S. Raymond2011-02-031-33/+30
|
* Cosmetic fix.Eric S. Raymond2011-02-021-21/+22
|
* Timestamps in TPV and SKY go from float seconds since Unix epoch to ISO8601.Eric S. Raymond2011-01-291-3/+3
| | | | | | | | Protocol version number is bumped. Python and C test clients are known to work; interfaces of the C and Python client bindings are unchanged. Third-party client-side bindings which rely on naively copying JSON members will break (implementers have been repeatedly warned not to do this).
* Catch and display regression errors from malformed JSON.Eric S. Raymond2011-01-291-1/+9
|
* Reorganize stream argument flags, and create one for timing data.Eric S. Raymond2011-01-262-10/+16
|
* The raw_hook mwember and set_raw_hook are gone; client API cleanup is done.Eric S. Raymond2011-01-251-8/+1
|
* Reduce the amount of padding time in the regression tests.Eric S. Raymond2011-01-231-2/+2
| | | | | | | | There may have been a real (minor) problem that needed to be solved this way at one time, but it was masking a major subtle bug in gpsd's packet handling that we have since fixed. We could probably set both pads to 0 on Linux but because I remember a trouble report from NetBSD I'm being a bit conservative.
* Various small fixes to Python client mode.Beat Bolli2011-01-051-3/+5
| | | | Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
* Address Berlios tracker bug #17485: problem with isotime in gps.py.Eric S. Raymond2010-12-021-0/+1
|
* Fixing several Python identation errors.Bernd Zeimetz2010-08-291-2/+2
|
* Carry through and document the poll() -> read() change.Eric S. Raymond2010-07-143-4/+4
|
* Comment fix.Eric S. Raymond2010-06-101-1/+1
|
* Incorporate the tesults of a test by Greg Troxel (comment fix only).Eric S. Raymond2010-06-091-4/+3
|
* Comment fixes and clarifications.Eric S. Raymond2010-06-091-8/+11
|
* Fix wordo from previous edit.Greg Troxel2010-06-091-3/+3
| | | | (Comment change only.)
* Set CLOSE_DELAY to 1 second.Greg Troxel2010-06-091-2/+4
| | | | | 0.05s is not sufficient on NetBSD. (I believe 0.05s ends up being rounded to 0 when used with integer seconds.)
* Reintroduce write padding in hopes of making the test suite work under NetBSD.Eric S. Raymond2010-06-071-5/+21
|
* regress-driver fixesBeat Bolli2010-06-061-1/+1
| | | | | | | | | These changes are necessary on my 32-bit Linux to make the tests work again after the GPSD_HOME addition. The fake.py changes is needed in case GPSD_HOME is not set at all (my initial problem); the 2nd hunk exports GPSD_HOME for subprocesses of regress-driver and uses it internally.
* add GPSD_HOME environment variable support to gpsfakeJon Schlueter2010-06-051-8/+19
| | | | | | | | check the environment variable GPSD_HOME to find gpsd on path specified if not there check normal PATH env variable for gpsd added note about GPSD_HOME in gpsfake.xml but might not be in correct location changed regress-driver to use GPSD_HOME instead of adding current directory to the PATH
* Use 127.0.0.1 for loopback tests.Greg Troxel2010-06-041-1/+1
| | | | | | | | | Previously, the tests use 127.0.1.255. While it isn't clear if standards require communications to this address to work, historical behavior on BSD dating back to 4.4BSD is that 127.0.0.1 is the sole valid loopback address and the rest of 127/8 is unreachable due to a blackhole route. Therefore, avoid addresses other than 127.0.0.1 in 127/8. Causes et-332 test to pass on NetBSD.
* Speed up the test suite by killing the daemon faster.M Joonas Pihlaja2010-06-041-2/+5
| | | | | | | | The daemon is rather quick to die on command, so it's enough to keep on killing it until it's gone, rather than sending a signal to it once and then gratuitiously sleeping for a second. Signed-off-by: Eric S. Raymond <esr@thyrsus.com>
* Rip out grubby code from a previous attempt to fix the multipacket problem.Eric S. Raymond2010-06-041-11/+1
| | | | All regression tests pass.
* Following the multipacket fix, all tests now pass without write padding.Eric S. Raymond2010-06-041-10/+0
|
* Make packet sniffer pass packets with bad checksums with type BAD_PACKET.Eric S. Raymond2010-05-081-2/+0
| | | | | | | | Such packets used to be truncated to zero length with the packet type matched before the checksum was bad. This caused confusion at various points in the code that were checking for zero-length packets. All regression tests pass, both in pty and UDP modes.
* Make gpsfake detect when it pulls a 0-length packet from a logfile.Eric S. Raymond2010-05-071-0/+3
| | | | | When this happens, it indicates a bug in the packet getter. It should never do that.
* Sanity-check things that look like responses better.Eric S. Raymond2010-05-061-3/+4
|
* NUL-terminate write to control socket to avoid junk in log messages.Eric S. Raymond2010-05-051-3/+3
|
* Implement a -u switch to force the test framework to use UDP, not ptys.Eric S. Raymond2010-05-041-2/+3
| | | | | | The good news is that rgression tests all pass with this case and -u off. The bad news is that they fail with -u on. It looks like the switch forces some other bug, previously masked, into the open.
* Shutdown fails on a UDP socket.Eric S. Raymond2010-04-291-1/+1
|
* Remove obsolete FIXME.Eric S. Raymond2010-04-261-1/+0
|
* Restore interpretation of Serial: headers inb test loads.Eric S. Raymond2010-04-261-3/+3
| | | | And allow an explicit speed argument to override them.
* Add a sanity check.Eric S. Raymond2010-04-261-1/+1
|
* Hide the wire protocol a bit more effectively.Eric S. Raymond2010-04-261-0/+3
|
* Fix minor bug introduced by recent refactoring.Eric S. Raymond2010-04-261-0/+2
|
* Fix a delegation error.Eric S. Raymond2010-04-251-1/+1
|
* The gpsfake.py module can broadcacst on UDP rather than using a pty.Eric S. Raymond2010-04-251-37/+74
| | | | This is a step towards making UDP data sources work.
* Yet more Python-client refactoring. All regression tests pass.Eric S. Raymond2010-04-232-10/+9
|
* More Python client partitioning. All regression tests pass.Eric S. Raymond2010-04-222-188/+206
|
* Back off the last change to gps.fake.py to make regressions pass again.Eric S. Raymond2010-04-221-6/+29
| | | | | Matters were confused by some build-system breakage, but putting the end-of-line padding back in seems to fix things.
* Return the length of the data in read().Bernd Zeimetz2010-04-221-1/+1
| | | | | The legth of the remaining data might be 0 if the data in the buffer ends with \n.
* gps.py: Make next() work at all.Bernd Zeimetz2010-04-221-24/+4
| | | | | | The next function contained a code copy from __json_unpack, was buggy in general and did not support oldstyle connections. The current implementation should have the same function for newstyle connections.
* There are utf8 chars in the file, so let Python handle them.Bernd Zeimetz2010-04-221-0/+1
|
* Make verbose mode of gps.py actually work.Bernd Zeimetz2010-04-221-2/+8
|
* __set_device__ is define in next(), not in the class.Bernd Zeimetz2010-04-221-1/+1
|
* gps.py takes -v as option, not -w.Bernd Zeimetz2010-04-221-1/+1
|
* No more write padding in gpsfake. either.Eric S. Raymond2010-04-201-10/+0
| | | | All regression tests pass.
View Lorry Controller Status