When failing to communicate with the svn server at initialization time, a visit status
ends up in "not_found" status.

Other failures to communicate ends up in a visit status marked as "failed".

Related to T3030

With the latest swh.loader.core 0.17, a partial without without snapshot is failed.

Related to T3030

This code hasn't used this module for years (last import removed in
25e415b7).

Related to T1532
Related to T1410
Related to D3965

Fixes build [1]

[1] https://jenkins.softwareheritage.org/job/DLDSVN/job/tests/820/console
[2] https://jenkins.softwareheritage.org/job/DENV/job/tests/lastBuild/console

Related to T2610

Related to T2610

flake8 hook has been removed from https://github.com/pre-commit/pre-commit-hooks
so now use the one from https://gitlab.com/pycqa/flake8

Due to an invalid isinstance type check, the loader ended up with error when a
directory was removed in a subversion commit.

Related to T645

pytest wastes a lot of time in .hypothesis and .git; this commit excludes them.

... with other loaders.

Prior to this commit, the pre-snapshot existing code to resolve the starting
point of a visit was convoluted and hard to follow.

This simplifies it to actually use the snapshot to start a new visit as other
loaders do.

Related to T2105

This fixes [1]

[1] https://jenkins.softwareheritage.org/view/Debian%20packages/job/debian/job/packages/job/DLDSVN/job/gbp-buildpackage/44/console

Related to T645

Related to T2484

Related to T2494

Coverage remained the same as this part is already tested by other scenario.

branch names and targets are expected to be bytes.

with extra_headers now being an attribute of the Revision entity.

Related to T2481

This fixes ci build failure [1]

[1] https://jenkins.softwareheritage.org/job/DLDSVN/job/tests/718

This aligns once more with other loaders ("data" test resources folder). It's
not particulary interesting but this also cleanup 2 archives which contained
themselves for no good reason...

[1] https://forge.softwareheritage.org/D3397#inline-23568

Related to T2462

It was unexpectedly "eventftul" so far [1]. Most probably since [2] [3] or
something near.

[1] https://forge.softwareheritage.org/D3399#inline-23571

[2] https://forge.softwareheritage.org/D2887

[3] https://forge.softwareheritage.org/D2889

Related to T2462

This actually uses a real modified repository.

Prior to this, this faked the alteration by modifying the input of the loader
which is a hack.

Related to T2462

Related to T2462

This actually fix an issue with the snapshot which in some cases were not of
the right type.

Related to T2462

Related to T2462

Related to T2462

Related to T2462

Related to T2462

Related to T2462

This adds both check for the scenario about repository with or without changes.

Related to T2462

Related to T2462

For this, switch to using a pg storage to share the state amongst visit.

Related to T2462

Related to T2462

This starts to align the existing svn tests with the core/package ones. For
now, utilities functions are in this repository close to the tests. They will
most possibly be refactored out in swh-loader-core at some point.

Related to T2462