GitLab

Heartbeat coverage improvement

See merge request !101

Merge branch 'heartbeatUTImprovement' of https://gitlab.torproject.org/milad-laly/sbws into heartbeatUTImprovement

This should increase the coverage to 100% and should be passing the tests/commands when running tox

Changed variables to PEP 8 standard

I didn't see the link https://tpo.pages.torproject.net/network-health/sbws/contributing.html#code-style
changed the variables

line length was too long

line 64 was too long

improvements and being inline with pep8

As juga suggested in the commits, I've done.
I tried to figure out another way instead of manually defining the value but couldn't figure it out

As juga suggested in the commits, I've done.
I tried to figure out another way instead of manually defining the value but couldn't figure it out

line 64 was too long

I didn't see the link https://tpo.pages.torproject.net/network-health/sbws/contributing.html#code-style
changed the variables

This should increase the coverage to 100% and should be passing the tests/commands when running tox

fix: doc: make CDN more optional

Closes #40101

See merge request !100

Closes: #40101.

because bullseye is the new Debian stable, has by default python3.9
amd tests using buster dependencies fail.

Closes #40099.

fix: Update canonical repository links

Closes #40002

See merge request !97

Closes: #40002.

Closes: #40090

stop exiting when there's a possible exception that makes sbws stalled
and instead just dump the stack. Additionaly, call pdb on sigint.

Since tor version 0.4.7.0-alpha-dev with #40337 patch, chutney relays
notice bandwidth changes.

to more appropriate names, after switching to concurrent. futures.

away from multiprocessing, because it looks like we hit python bug
22393, in which the pool hangs forever when a worker process dies.
We don't know the reason why a worker process might due, maybe oom.
See https://stackoverflow.com/questions/65115092/occasional-deadlock-in-multiprocessing-pool,

We also run into several other issues in the past with multiprocessing.
Concurrent.futures has a simpler API and is more modern.

Closes #40092.

because when the observed bandwidth is higher than the consensus (for
example when the relay is new or was some time down), it's limited by
the previous consensus, not allowing it to grow.

Since the size of the data to download depends also on the consensus
weight, this results on lower measured bandwidth too.

Closes #40091.

instead of version, so that it's more clear which version is being installed.

Closes: #40089

since the default tor with deb.tpo repository is master

since master is the default and add test for master.

- Increase the time waiting for the last measurements queued, to avoid
  canceling unfinished measurements and gc maybe not releasing thread
  variables
- Use the already declared global pool instead of passing it by args
- Log more information when the last measuremetns timeout

Closes: #40087

To pass tox tests.
This seems to have changed in black from version 20.8b1 to 21.4b2.

also replace the links to Read the Docs to pages.torproject.net
and add redirect to it.

Closes: #40085.

Update changelog, updated automatically, manually fixed typos and
and new lines.

Closes #40004.

and put in the same entry commit subject and body removing new lines.

Closes: #40004

Get rid of the last documentation in markdown and remove extra
dependency!

Cause of tpo/tpa/team#40221.