Andrew Tridgell
dee1bf3da3
Replay: show smaller parameter changes
2015-07-07 07:56:17 +10:00
Andrew Tridgell
bdddfae57f
Replay: don't write out duplicate FMT messages
...
this makes replay output the same as input for when input is a replay
log
2015-07-04 22:43:54 +10:00
Andrew Tridgell
93800fb3a1
Replay: automatically find message types, and improve rate detection
2015-07-04 15:14:14 +10:00
Andrew Tridgell
b20318aaeb
Replay: remap msgids on pass-through to avoid conflicts
...
if msg IDs have changed since the log was produced they need to be
remapped
2015-07-03 09:32:05 +10:00
Andrew Tridgell
bd27fed241
Replay: fixed missing import
2015-07-02 17:14:42 +10:00
Andrew Tridgell
ef36d74a1c
Replay: added generation of -checked logs
2015-07-02 17:05:09 +10:00
Andrew Tridgell
c74d39558c
Replay: initial version of CheckLogs.py
2015-07-02 17:05:09 +10:00
Andrew Tridgell
4f9ccf727a
Replay: cope with bad NSats field from old logs
2015-07-02 17:05:09 +10:00
Andrew Tridgell
a5abb7c698
Replay: use tabs in replay_results.txt
2015-07-02 14:28:44 +10:00
Andrew Tridgell
f79ae4b749
Replay: add to replay_results.txt when run with --check
2015-07-02 14:28:44 +10:00
Andrew Tridgell
b93101b1bf
Replay: reduced verbosity
2015-07-01 21:11:46 +10:00
Andrew Tridgell
2c0d6c0406
Replay: fixed yaw error calculation
2015-07-01 20:44:08 +10:00
Andrew Tridgell
d07b93e661
Replay: fixed writing of CHEK messages
2015-07-01 20:44:08 +10:00
Peter Barker
8bbeedb158
Replay: flush dataflash log on floating point exception
2015-06-30 16:19:22 +10:00
Andrew Tridgell
34689cf949
autotest: cope with mavproxy logging msg changes
2015-06-30 14:06:17 +10:00
Andrew Tridgell
2291a8994e
Replay: fixed build
2015-06-29 15:59:53 +10:00
Peter Barker
2266f118e9
Replay: DataFlash frontend/backend split
2015-06-26 16:02:50 +10:00
Andrew Tridgell
f2756ecb22
Replay: added checking of solution
2015-06-26 15:55:19 +10:00
Andrew Tridgell
9016a3d3db
Replay: added --chek-generate option
...
adds CHEK messages based on current estimates
2015-06-26 15:55:18 +10:00
Andrew Tridgell
42023eb17a
autotest: check for .gitmodules before ldir check
...
another attempt to get build_binaries going
2015-06-26 11:32:15 +10:00
Peter Barker
d4a34b3da1
SITL: abort sim_vehicle.sh if build fails a second time
...
Previously a build could fail and we would end up running the old code
2015-06-25 10:16:24 +10:00
Andrew Tridgell
48828a7834
autotest: more changes to cope with older builds without submodules
2015-06-24 09:36:21 +10:00
Andrew Tridgell
fd17a69cd1
autotest: skip binary builds without submodules
...
the build_binaries is trying to build for navio and pxf without
submodules
2015-06-23 21:43:44 +10:00
Andrew Tridgell
63895cbe75
autotest: build pxf and navio binaries for firmware.diydrones.com
2015-06-23 14:27:49 +10:00
Peter Barker
54efa3d727
Replay: split off a ReplayVehicle class
...
Like "Copter copter" and "Plane plane": "ReplayVehicle replayvehicle"
2015-06-23 09:49:13 +10:00
Andrew Tridgell
7e23912d17
autotest: added -b option for build target to sim_vehicle.sh
2015-06-19 16:39:45 +10:00
Grant Morphett
0645474b3b
autotest: Fixed the LOG_BITMASK for Rover in SITL.
2015-06-18 17:54:19 +10:00
Andrew Tridgell
9e36e2c562
Replay: fixed handling of dual-GPS HDop field
2015-06-18 09:28:59 +10:00
Andy Piper
a8df968441
Fix vagrant build issues with submodules.
2015-06-18 08:47:32 +10:00
Andrew Tridgell
47834759e5
Replay: fixed build warnings
2015-06-17 12:46:55 +10:00
Andrew Tridgell
31f7525a61
Replay: added --no-imt flag
2015-06-17 12:46:55 +10:00
Lucas De Marchi
ac09a61238
Travis: split build by target instead of vehicle
...
This way we can group together the builds that are fast so we pay the
"setup price" just once. Let the long PX4 build on its own VM because
it's the one that takes most of the time building NuttX.
By grouping the projects by target it's also easier to split the build
in more VMs if we want to speed up the build.
2015-06-17 09:26:04 +10:00
Lucas De Marchi
a435269839
Travis: make most targets run with -j2
...
As documented in Travis the VMs have 1.5 cores
(http://docs.travis-ci.com/user/speeding-up-the-build/#Parallelizing-your-build-on-one-VM ).
So let's make it only -j2.
The exception is PX4 build that runs without any -j setting since NuttX
fails to build if we do that.
2015-06-17 09:26:04 +10:00
Lucas De Marchi
985764fb1c
Travis: build APMrover2 and AntennaTracker for apm2
2015-06-17 09:26:03 +10:00
Lucas De Marchi
5c5bdc2c7b
Travis: use associative arrays
...
This makes the script be more declarative and easier to extend.
2015-06-17 09:26:03 +10:00
Lucas De Marchi
ae40aed2a2
Travis: split build using vehicle types
...
Travis build is timing out because it's taking too much time. Let's
split the build so 1) they can run concurrently and 2) we don't timeout.
We still lose some time due to multiple checkouts, but at least we don't
fail the build due to timeout.
2015-06-17 09:26:03 +10:00
Lucas De Marchi
a124fbc4c2
Travis: don't build same configuration multiple times
...
In ArduPlane "make all" is the same as "make sitl" and both sitl and
linux are already covered by the loop below. Don't build the same
configuration multiple times.
2015-06-17 09:26:03 +10:00
Andrew Tridgell
43d6060ba7
autotest: removed "Library Parameters" heading in autodocs
2015-06-16 10:01:03 +10:00
Andrew Tridgell
c7cb980498
Replay: support IMT messages
2015-06-16 09:53:55 +10:00
Andrew Tridgell
46926765c4
autotest: support general models
2015-06-15 09:20:56 +10:00
Andrew Tridgell
22d5d2e251
scripts: fixed build_binaries for submodules
...
no longer have PX4_ROOT
2015-06-15 06:54:02 +10:00
Peter Barker
71bc48189e
Replay: remove dangling declarations from LogReader.h
...
These seem to have been left in after MsgHandler was factored out
2015-06-14 17:25:08 +10:00
Peter Barker
f996bf4307
Replay: move LR_MsgHandlers into their own files
2015-06-14 17:17:42 +10:00
Peter Barker
7eb588e9ff
Replay: use a subclasss of DataFlashFileReader to determine log rate
2015-06-14 16:47:52 +10:00
Peter Barker
d41b21cd19
Replay: split off a base class for reading dataflash logs
2015-06-14 16:47:52 +10:00
Peter Barker
0b5e645075
Replay: grab reference to timestamp rather than timestamp
...
PARM messages would not have been updating the clock correctly
Also, random stack corruption...
2015-06-14 16:34:01 +10:00
Staroselskii Georgii
c0dceb667d
Travis: reduced number of threads used for a build
2015-06-11 16:27:51 +10:00
Peter Barker
125042e1db
Replay: use GetOptLong for command line parsing
2015-06-10 10:14:15 +10:00
Peter Barker
f4db241bf6
Replay: avoid dataflash buffer filling up on many pass-through messages
2015-06-09 21:52:46 +10:00
Peter Barker
5fb46b0e11
Replay: fix dataflash overflow on timestamp-poor (older) logfiles
2015-06-09 21:52:45 +10:00
