summaryrefslogtreecommitdiff
path: root/TODO
Commit message (Collapse)AuthorAgeFilesLines
* The gpsmon bug was probably a result of bogus packets from the lexer.Eric S. Raymond2010-05-101-7/+0
|
* TODO polishing. All regression tests pass.Eric S. Raymond2010-04-271-3/+3
|
* Remove the IRC transcript about regression-test failures.Eric S. Raymond2010-04-191-17/+0
| | | | | I've established that this is purely a bug in the termination condition of the main loop inside gpsfake, and am working on that.
* Record sanooj's insight.Eric S. Raymond2010-04-181-0/+17
|
* Add Baylink's todo item.Eric S. Raymond2010-04-161-0/+5
|
* Add notes on ports to be finished.Eric S. Raymond2010-04-161-0/+10
|
* Update the TODO list.Eric S. Raymond2010-04-151-21/+31
|
* First cut at JSON output for TNT Revolution device.Eric S. Raymond2010-04-101-9/+0
| | | | | Not yet documented, but now we at least have a regression test. All regression tests pass.
* Simple dumb fix for gpsmon spin bug. All regression tests pass.Eric S. Raymond2010-04-071-13/+0
|
* Fix the select()-spin bug on disconnected devices.Eric S. Raymond2010-04-071-31/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Addresses Berlios tracker bug #14638: 100% cpu when Bluetooth gps device vanishes. Also a couple of reports on the mailing lists. The underlying problem here was that: * Disconnecting a USB device causes reads from it to begin returning 0, in effect an end-of-file condition. * select(2) sets an active bit not on "I/O is ready" but on "read would not block" -- including the end-of-file condition. * Therefore, select() will spin any time its fd set includes a disconnected device. The fix is to check for a zero-length read explicitly and always take the device out of the active set when that happens. We were actually doing that here, but the test was defective in two ways: 1. The check for a zero return from gpsd_poll(), indicating I/O error or zero-length read, needed to be *before* the check for full paccket rather than after. This effectively disabled it. 2. There was a conditional arm in the gpsd_poll() code that made it ignore zero-length reads for up to a full cycle. I think this may have been a fossil from long ago when I experimented with non-blocking reads on devices. The fix for the gpsmon spin bug is probably quite similar. That's up next.
* Add all our info on the spin bug to TODO, and instrument all opens/closes.Eric S. Raymond2010-04-061-0/+44
|
* Add report of time reference instability on loss of fix.Eric S. Raymond2010-04-011-0/+10
|
* Endianess issue fixed.Eric S. Raymond2010-03-291-7/+0
|
* Suppress JSON reporting of fields the fix quality won't support.Eric S. Raymond2010-03-271-17/+4
| | | | | | | | This should suppress a bug reported by Greg Troxel where, due to numeric instability in geodetic fix computations, wecwere getting cross-archirecture differences in the low-order digits of altitude for invalid fixes on SiRF-II chips. This required that ten regression test be rebuilt; I eyeballed the diffs to make sure we didn't lose actual data.
* We now understand NMEA status fields better. Document appropriately.Eric S. Raymond2010-03-221-36/+0
|
* Vanish away Id and Rev $ keywords, git won't expand them to anything useful.Eric S. Raymond2010-03-131-2/+0
|
* Emulating the Gypsy signals no longer makes sense. Eric S. Raymond2010-03-111-7/+0
| | | | We'll let geoclue take care of that.
* All the IPv6 todo items are done.Eric S. Raymond2010-03-051-34/+0
|
* The numeric instability in earth_distance() has been fixed.Eric S. Raymond2010-03-011-11/+0
|
* More bug diagnosis.Eric S. Raymond2009-12-221-2/+3
|
* Typo fix and some bug diagnosis.Eric S. Raymond2009-12-221-11/+13
|
* Various AIS corrections turned up by working on xgps's AIS display.Eric S. Raymond2009-12-201-5/+0
|
* Remove fulfilled items.Eric S. Raymond2009-12-171-5/+0
|
* Add todo item with link to GeoClue...Greg Troxel2009-12-151-0/+10
| | | | | ...and the notion of considering how the dbus interfaces should relate, mostly to have the URL handy for gpsd hackers.
* Update the TODO.Eric S. Raymond2009-12-151-49/+14
|
* Replace the old Motif-based xgps with a new pygtk-based xgps (formerly xgps2).Eric S. Raymond2009-12-151-6/+0
| | | | | Replaces 1200-odd lines of difficult-to-maintain C code with 400-odd-lines of Python.
* Remove a TODO item, since NTRIP Is reported working again.Eric S. Raymond2009-12-141-8/+0
|
* DGPSIP and NTRIP device opens were being done in a bad way.Eric S. Raymond2009-12-071-0/+8
| | | | | | Code could not have produced good results. Move the opening code to somewhere more sane - it might work now, we need to find some way to test.
* Ship a DEVICE notification on every driver changeEric S. Raymond2009-12-051-6/+0
|
* Update TODO.Eric S. Raymond2009-12-051-0/+2
|
* Fix description of bu303-nofix, and add details. This is not aboutGreg Troxel2009-12-051-2/+13
| | | | NetBSD - the test fails for me on Linux/i386.
* Update the list of known regression-test problems.Eric S. Raymond2009-12-051-3/+13
|
* Add a to-do item for 2.91.Eric S. Raymond2009-12-041-0/+6
|
* This commit resolves the time overflows in the rgm3800 and mkt3301.Eric S. Raymond2009-12-041-11/+0
| | | | | | | | | | | These were causing port problems on systems with 32-bit time. It turns out the problem was with the assumption that these devices always deliver a valid time in $GPRMC. They don't when the navigation warning bit (second field 'V') is on! The NMEA driver now knows. Also, the code now contains a sanity check - it will log a complaint if it sees a date moere than a year in the future. This invariably indicates some driver-level problem with time extraction.
* Add "GPSD and Code Excellence" essay to white papers.Eric S. Raymond2009-12-041-10/+20
| | | | | | Update the future roadmap (our schedule slipped by 5 days). Update the TODO file with driver bugs we probably won't be able to fix before 2.91.
* Instrumentation for the client library.Eric S. Raymond2009-11-181-1/+1
| | | | | | #define LIBGPS_DEBUG to verbosely dump unpacked data to stderr; this can be used to check that the client library is delivering everything it should.
* cgps is converted to use new protocol.Eric S. Raymond2009-11-171-4/+0
|
* Force -X11 inclusion in X_LIBS.Eric S. Raymond2009-11-061-1/+0
|
* gpsprof is fully working with the new protocol.Eric S. Raymond2009-09-291-2/+0
| | | | All regression test pass. Codebase splints clean.
* Better description of new Python API. Eric S. Raymond2009-09-241-7/+2
| | | | | query() is really gone; add a resulting to-do item. Codebase splints clean, all regression tests pass.
* gps_query() is removed.Eric S. Raymond2009-09-231-3/+2
|
* New protocol can now ship profiling information, like old $.Eric S. Raymond2009-09-231-2/+0
|
* Fully support AIS type 22 and 23 messages. All regression tests pass.Eric S. Raymond2009-09-221-2/+0
| | | | Codebase splints clean.
* Separate WATCH_NMEA from WATCH_RAW in the visible API.Eric S. Raymond2009-09-211-0/+2
| | | | All regression tests pass. Codebase splints clean.
* List known driver-level issues.Eric S. Raymond2009-09-211-0/+5
|
* Old-protocol dependency removed from gpsmon.c.Eric S. Raymond2009-09-191-2/+0
|
* Remove a fulfilled TODO item, add some warning comments,Eric S. Raymond2009-09-171-5/+0
| | | | and fix a minor inconsistency.
* On second thought, we ought to support two client bindings... Eric S. Raymond2009-09-111-5/+1
| | | | ...so they are a sanity check on each other.
* J command and switch are gone.Eric S. Raymond2009-09-111-6/+0
| | | | | | | | | Now that we have reliable end-of-cycle detection in NMEA, we can always clear the fixbuffer at start of cycle and accumulate data until we transmit at end of cycle. Accordingly, there is no longer any bneed for users to choose between jittery-but-correct and buffered-but-laggy reports. Accordingly, the daemon 'J' command is no longer necessary. All the client side options and commands connected with it can go away too.
* Improve synthetic test load for AIS unit test. Eric S. Raymond2009-09-071-4/+11
| | | | | Update TODO, because AIS issues are cleaned up. All regression tests pass.
View Lorry Controller Status