before stopping the scanner.

We changed conf['paths']['X'] to use conf.getpath('paths', 'X'),
so that paths with `~` get expanded, but cleanup was forgotten.
Also remove extra path check in main.
And run cleanup as part of the integration tests.

Bugfix v0.7.0.

Exit the scanner with error stoping threads first when there is not
any functional destination, since the destinations can not be recovered.
After merging #28897, `stop_threads` can be used (instroduced in #28869).

Add methods to store consecutive destination failures and retrieve
the destinations that are still functional.
Since destinations can fail because of Tor circuits, it's not count
individual failures but consecutives one.
Also exit with error if there are no functional destinations left.
The maximum number of consecuitve failures is set to 10, but it
may need to be changed depending on the percentage of circuits and
requests that fail.

in every measurement.
This removes the need for an extra lock for every measurement
It should also not be depending on a time interval, but on the
number of failures detected.
Not counting number of failures since it would need to modify the
destination or list of at runtime. It should be done in a future
refactor.

Fixes bug #28897. Bugfix v0.3.0

and create function to set options that can fail because they are
not supported by some Tor versions at runtime.

Fixes bug 28692. Bugfix v0.4.0

The log message might not correspond to the reason.
Only the first one that fails will be logged, but not refactoring
here.
Add an integration test to check a runtime option fails.

to a function.
Replace comment by docstring.
Add an integration test to check the option is set.

after finishing the first loop.
As noted in bug28933_01.

The bandwidth file is generated only with relays that have some
successful measurements, but we forgot to load all results
(including the relays with failed measurements) after we started
to include the number of failure measurements in each relay
bandwidth line.

Closes #29568. Bugfix v0.4.0 (line introduced in v0.1.0).

that will always have timeout in the get and head methods.

Override get and head methods to always have a timeout.

Merge: 8dbe8905.
Coded introduced in a550d731 and
5dae9b99.

Closes: #29149. Bugfix v0.1.0.

when the cli argument `log-level` is provided.
Before, it was only setting the minimal log level, but it was not
changing the stdout level.

Closes: #29199. Bugfix v0.6.0.

Instead, return the reason why the circuit could not be built,
as other errors do.

There is no any case in our code where path is an integer.
Remove asserts, check only the lenght of the circuit and check it
in a more logical way.

It is observed that when the circuit can not be built the first
time, consecutive attemps will also fail.
The relay will be measured anyway in the next iteration.
This will also speed up the scanner.
There is no need to remove the 3 attemps in Destination.
_perform_usability_test, since the method will be removed in #28897.

Fixes bug #29295. Bugfix v0.1.0.

that depends on the HTTP request timeout.

since the relay might start being measured some time after
apply_async is called if there're not available threads.

also split main function into an extra main_loop function to be
able to stop the threads after they have started.
Also check end event in the mean loop and before starting to
measure a new relay.

Fixes bug #28869. Bugfix v0.1.0.

so that the objects that manage the threads can be stop at any
time.

when worker thread finish.

instead of assigning it to the ResultDump object.

After several iterations, the number of relays that are prioritized
is high, this log prints so many lines.

result.freshness_reduction_factor are hard-coded values, no need
to assert.

Not all the relays in the network are returned by this method, but
a fraction of them. Create an argument to allow to change this
default behaviour to return all the relays in the network.