jarthianur / virtualflightradar-backend Goto Github PK

Add ability to read gps data from gpsd.

From what I gathered so far the steps to get the FLARM Radar working would be:

Please correct me if wrong !

• Working rtl-sdr installation
• Install and run ogn-rf and ogn-decode
• Run local aprsc-server
• Make ogn-decode report to localhost instead of glidern{1;2}.glidernet.org
  (How do you accomplish that? DNS spoofing?)
• Build and run VFR-B. Connect VFRB to local aprsc-server
• Connect XCSoar to VFR-B

Some more questions:
Will the new ogn binaries run without connection to the internet?
What about OGN-DDB whitelist ? Will ogn-decode report targets that are not listed in the DDB?

I hope I can get your system to run here. That would be neat. :-)

Describe the bug

The require function in bootstrap.sh is used to check arguments given to a function or script.
Like require 1 2, but this does not work as expected, because positional args are not passed as env vars.
require 3 eg. shows the bug.

This is not severe, as require is only used as safeguard, but should be corrected anyway.

Is there also a function to use real GPS data stream instead of creating GGA and RMC for a static position ?

In order to evaluate a probability of arrival, check whether an aircrafts estimated flight path crosses base position.

This issue tracks the release status for version 4.

TODO:

• Fix all existing tests
• Change all processors to use libfmt instead of old C functions
• Verify this to be working for moving station - in flight
• Verify new concurrent container to be threadsafe
• Verify new concurrent access to aircrafts does not break round based processing logic
• Verify in-place and pointer based string processing to be safe
• Audit code
  • Double check everything to be safe and secure
  • Rename and refactor where necessary
  • Check code style consistency
  • Write code-docs
• Rewrite docs and readme to be up-to-date and more clear and concise
• Update versions and changelogs

Use previous position reports to evaluate an aircrafts possible flightpath.

Values not given in report (adsb).

• ground speed
• climb ratio
• turn ratio
• heading

Subject

enhancement, bug

Summary

The priority handling for multiple feed merges is unstable for more than two feeds of same type.
This is actually no bug, but can cause unwanted behavior, when concurrent feeds of different priorities
want to update data. Hence improve the priority handling.

Expected behaviour

Concurrent feeds should handle write attempts in a proper way.

Actual behaviour

Having more than two concurrent feeds can cause unintended data updates.

Severity

Severe

Document and verify steps for cross compilation from Linux to Raspi.

Describe the bug

Building on fedora linux fails, because it is considering yum as pkg manager and does not find all packages.

To Reproduce

• be on fedora linux
• run.sh build

Expected behavior

The install script should detect dnf for fedora and respective packages.

System

• OS: Fedora
• Version: *
• Platform: *

Additional notes
Add any other notes about the problem here.

Subject

enhancement, build

Summary

Enhance the build process with spline. First use this in travisCI.

Expected behaviour

Build this project in a simple pipeline.

Actual behaviour

Currently this project is built sequentially in the travis file.

Severity

Nice to have, but not essentially.

Describe the bug

When having a sensor client, the connection is not closed and reopened after timeout for no data received.

To Reproduce
Steps to reproduce the behavior:

1. use a atm/wind feed
2. let source send no data for the time configured as timeout

Expected behavior

The connection should be closed and reopened correctly.

System

• OS: *
• Version: *
• Platform: *

Additional notes

This could also affect the keep-alive beacon for aprsc clients.

Using the driver from sensorics as interface to read pressure, temperature and humidity from small local sensor (bme280).

Subject

enhancement

Summary

Integrate BetterCodeHub into the CI process.

Expected behaviour

Gain report for code quality.

Actual behaviour

No code quality integration.

Severity

Nice to have

Change hard-coded input feeds to generically ones.
List feeds in config like "feeds=sbs,ogn,gps...".
Thus parse config section wise.
Spawn input threads according to the feeds list.

Along to more unit tests, especially for Feeds, add regression tests to the build process.

Is your feature request related to a problem? Please describe.
To make server handling more flexible, it would be nice to configure the maximum allowed connections via config file instead of hard-coded at compile time.

Describe the solution you'd like
Add a config entry for max server connections and implement this dynamic approach.

Describe alternatives you've considered

Additional notes

Is your feature request related to a problem? Please describe.
When connecting a feed, or reconnecting after lost connection, a fixed delay (eg 2 min) is actually not desirable.

Describe the solution you'd like
A back-off strategy would allow (re-)connects with an increasing but limited delay.
For example, the first attempt fails, so a second one is done after 2 sec. If that also fails, the next attempt follows after another 15 sec, and so on. It would then saturate at 2 min or whatever.

Describe alternatives you've considered

Additional notes

OGN closes connection after 30 minutes. To fix (hopefully) send a beacon every few minutes.
In addition reconnect if read gets uncritical errors (eof,...) for every client. Limit reconnect trials.

Let feed depend on other feed.
A feed may be configured to sleep until a certain feed dies.
In the config:

[feed1]
...
standby=feed2

Hey,

I was researching on how to build a standalone FLARM Radar using rtl-sdr and stumbled upon your project.
What is the correct command to build your source code? I could not find any of the make-cmake-style scripts I am used to. I (naively) tried gcc VFRB.cpp which resulted in lots of errors.

Subject

bug

Summary

Netcode seems to be not exceptionsafe.

Expected behaviour

VFRB should exit clean and without exceptions.

Actual behaviour

Program is terminated with uncaught exception.

Steps to reproduce

Run VFRB and cancel via <CRTL+C>. If nothing happens, do again.

Severity

severe

Note

[INFO] Sat Aug 4 09:13:47 2018:: (VFRB) caught signal to shutdown ...
[INFO] Sat Aug 4 09:13:47 2018:: (AprscClient) disconnect from aprs.glidernet.org:14580
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) timeout: Operation canceled
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) read: End of file
[ERROR] Sat Aug 4 09:13:47 2018:: (AprscClient) failed to read message
[INFO] Sat Aug 4 09:13:47 2018:: (SensorClient) disconnect from 10.0.0.30:10110
[INFO] Sat Aug 4 09:13:47 2018:: (SbsClient) disconnect from localhost:30003
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) read: End of file
[ERROR] Sat Aug 4 09:13:47 2018:: (SbsClient) failed to read message
[INFO] Sat Aug 4 09:13:47 2018:: (AprscClient) disconnect from localhost:14580
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) timeout: Operation canceled
[INFO] Sat Aug 4 09:13:47 2018:: (GpsdClient) disconnect from localhost:2947
[INFO] Sat Aug 4 09:13:47 2018:: (GpsdClient) stopped watch
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) read: Connection reset by peer
[ERROR] Sat Aug 4 09:13:47 2018:: (GpsdClient) failed to read message
[INFO] Sat Aug 4 09:13:47 2018:: (SensorClient) disconnect from localhost:7997
[DEBUG] Sat Aug 4 09:13:47 2018:: (Client) read: End of file
[ERROR] Sat Aug 4 09:13:47 2018:: (SensorClient) failed to read message
[INFO] Sat Aug 4 09:13:47 2018:: (Server) stopping all connections ...
[DEBUG] Sat Aug 4 09:13:47 2018:: (Server) accept: Operation canceled
terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injectorboost::system::system_error >'
what(): remote_endpoint: Bad file descriptor
Abgebrochen