| Commit message (Collapse) | Author | Age |
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
Having an error dialog appear at application startup seems better than
simply failing to present the device dialog later on.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
| |
| |
| |
| |
| | |
It's not a public library, so hide it away in a subdirectory.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
| |
| |
| |
| |
| |
| | |
The libaltos Makefile needs to find the relevant jni.h file, which is
found in the JVM_INCLUDE path, all nicely located by the configure.ac script.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|/
|
|
| |
which assumes there's only one matching .. probably ok on Debian?
|
| |
|
|
|
|
|
|
|
|
| |
the fat .jar file is used in non-native builds to run from a directory
containing all of the freetts jar files along with the altosui jar
file. We don't want this on a real install where freetts is installed separately.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
This ensures that the JAVAROOT directory gets created by adding it to
the variable used to set the CLASSPATH environment value.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
This appears to matter to Vista. Dunno why.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
The implicit ordering doesn't appear reliable.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
Instead of looking for the first state change record, use the Flight
record to get the boost tick.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
|
| |
The very first record in the eeprom is the flight number, but it is
time-stamped with the 'boost' time, and so it gets sorted until much
later, delaying the return of data until the rocket enters boost
mode. This drops all of the nice pad GPS and state date on the floor.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
| |
This ensures that compiling only a few changed files will
locate the old .class files instead of using a stale .jar file.
|
|
|
|
|
|
|
|
|
|
|
| |
When flight status changes rapidly, the queue of voice data can get
quite long. This change does two things -- first, it remembers when
the altitude reporting happens due to flight events and delays the
periodic reporting until a suitable time after that, second it ensures
that the voice data has all been delivered before generating a new
altitude report.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
drain() blocks until all pending phrases have been processed, allowing
the UI code to avoid pending data that will end up stale by the time
it is emitted.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
This makes the first altitude report time consistently 10 seconds
after launch, instead of some random time depending on when the rocket
launched relative to the time the device connection was made.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
Normally, the replay process makes one final report after the file has
been parsed. However, if the reading process is interrupted to display
something else, this report is just annoying, so don't make it.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
Reported by voice, it's useful to see these on the display as well.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
While reading eeprom files, the GPS record is reconstructed each time
the system sees the first GPS log item (the time field), but as the
date isn't repeated, we need to copy it from the old GPS data record.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
These were in place while validing the GPS data reconstruction code
that handles eeprom files missing the first GPS date line due to the
record overwriting bug in old firmware versions.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
When an eeprom file contains an invalid line, just return
AO_LOG_INVALID instead of throwing an exception. This allows us to
replay and parse files with extraneous serial communication.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
Again, use 'cp -p' instead of 'cp -a' to get files into the archive
instead of links. Also, make the shell script 'altosui' instead of
'altosui-fat'.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
Must contain 'altosui.jar' instead of altosui-fat.jar.
Also, was using 'cp -a' instead of 'cp -p' which made files
represented by symlinks not end up in the archive.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
Was shipping altosui.jar instead of altosui-fat.jar
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
| |
Instead of using git revision counts for version numbers, use explicit
versions numbers configured in the configure.ac file. Expose published
files with version numbers.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
|
| |
|
|
|
|
|
|
|
| |
This switches from hand-written Makefiles to automake with libtool
for these parts of the system.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
Use the altus-metrum icon for an application icon and a windows start
menu/desktop icon.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
altus-metrum.ico: Windows ICO file
altus-metrum-16x16.jpg: 16x16 jpg image for java
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
And forgot to change the dependency in the Makefile
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
|
|
|
|
|
|
|
| |
nsis happens to be packaged in debian, and it appears to build usable
installers, which is all very cool.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
| |
|
|
|
|
|
|
|
| |
This catches the USB device being unplugged and makes sure the
user sees an error dialog in this case.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
Without flushing the configuration commands to the serial device, it
never sees them as the telemetry input thread doesn't flush.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
| |
This reverts commit ba65e4aeb952a1cf49a77f1e24e235508fcea71f.
Testing the old code
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Opening another serial device involves shutting down the display
thread (to reset its state) and spawning another one. Shutting down
the display thread normally closes the serial device as a part of the
process, and if this isn't done before the new serial device is
opened, then the new serial device ends up getting closed too.
Interrupting the display thread and waiting for it to stop before
opening the new serial device solves the problem.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
|
|
|
| |
If the monitoring thread is active, then setting the radio channel can
sometimes cause the monitoring thread to get stuck. I'm not entirely
sure why though. For now, work around the issue by making sure
monitoring is off, and the monitoring thread has stopped, before
changing the radio channel.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
Telemetry CRC errors can signal problems with TeleMetrum or TeleDongle
units, so report them in the UI.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
|
|
|
| |
Without the paths, the OS X zip file doesn't create a usable
application structure.
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
| |
Signed-off-by: Keith Packard <keithp@keithp.com>
|