* fname for generic file paths
* tor_fname for paths relative to the top-level tor directory
* src_fname for paths relative to tor's src directory
With prefixes as required to disambiguate different paths of the same
type.
Part of 32962.
* distinguish between paths relative to the top-level tor directory,
and paths relative to tor's src directory
* canonicalise paths before using them
* check that the script is run from the top-level tor directory
* check that the file is being created in tor's src directory
Part of 32962.
"Ours" merge, because ticket 32883 replaces this code with calls
to tor_run_main(). (Which then calls the pubsub functions.)
Therefore, this PR is obsolete in master, but it still needs to be
merged, to avoid conflicts during any future backport.
And canonicalise file paths, so that output is consistent, even in
out-of-tree builds.
Capturing stderr output produces better diagnostics.
Diagnostics for 32705.
We need a test branch for the earliest maint branch (maint-0.3.5),
but we can't merge forward into it, because there is no previous
active maint branch.
Also, edit a git-push-all comment that still said "029".
Bugfix on 32772; bug not in any released version of tor.
The variable lenght cells are handled directly by
channel_tls_handle_var_cell() from an OR connection reading its inbuf. The
channel var cell handler (agnostic) was never used.
Closes#32892
Signed-off-by: David Goulet <dgoulet@torproject.org>
We hit this assertion with bug 32868, but I'm stymied figuring out
how we wound up with a routerstatus like this. This patch is a
diagnostic to attempt to figure out what is going on, and to avoid a
crash in the meantime.
Now that these tests are skipped on windows _and_ android, we should
name the macro that skips them after the functionality that the
tests depend on.
Closes last part of 32825.
When practracker is disabled, its output will be empty. We don't
want that happening during our tests.
Fixes bug 32705; bugfix on 0.4.2.1-alpha, when test_practracker.sh
was introduced.
When looking up an entry in the rend_cache, stop asserting that it exists but
rather confirm it exists and if not, return that no entry was found.
The reason for that is because the hs_circ_cleanup_on_free() function (which
can end up looking at the rend_cache) can be called from the
circuit_free_all() function that is called _after_ the rend cache is cleaned
up in tor_free_all().
We could fix the free all ordering but then it will just hide a future bug.
Instead, handle a missing rend_cache as a valid use case as in while we are in
the cleanup process.
As Tor becomes more modular, it is getting more and more difficult to ensure
subsystem callstack ordering thus this fix aims at making the HSv2 subsystem
more robust at being called while tor is pretty much in any kind of state.
Fixes#32847.
Signed-off-by: David Goulet <dgoulet@torproject.org>
We never used them very much, and although they had potential to
clarify some of our tests, they also made some of the logic harder
for people to follow. Clang-format can't make head or tail of them,
so the time has come to say goodbye to them.
