I was expecting our filter code to work in a way it didn't. I
thought that saying that DependencyViolation applied to "*" would
hit all of the files -- but actually, "*" wasn't implemented. I had
to say "*.c" and "*.h"
* Move the shellcheck script from the Makefile to its own script file
* Reformat the shellcheck script so it's easier to read and modify
* Call the shellcheck script from the pre-commit hook
Fixes bug 30967; not in any released version of Tor.
Most shell arguments should be quoted to avoid mistakes.
But since all branch names are hard-coded, or supplied by the script user,
we don't need to be too concerned about command injection.
Quoting all shell arguments would take a major refactor.
(Probably using arrays.)
Part of 31314.
Skip test branches that are the same as remote maint/release/master
branches.
Add a TOR_PUSH_SAME and -s argument to git-push-all.sh to change this
default.
Part of 31314.
TOR_GIT_PUSH provides the git push command and default arguments.
Also fix handling of git-push-all.sh script arguments and arguments that
are passed through to $TOR_GIT_PUSH, using a "--" argument as a separator.
Fix on 29879.
Add a -t <test-branch-prefix> argument to git-merge-forward.sh and
git-push-all.sh, which makes these scripts create, merge forward, and
push test branches.
Add a -r <remote-name> argument to git-push-all.sh, so the script can
push test branches to a personal remote.
Closes ticket 31314.
(I've chosen to suppress some instances rather than 'fix' them,
since the fix would require arrays or major refactoring.)
Fixes bug 31519; bug not in any released Tor.
The pre-push hook already calls the pre-commit hook, which calls
practracker.
Also update the script comments to avoid similar issues in future.
Fixes bug 31462; bugfix on 0.4.1.1-alpha.
Add a TOR_PUSH_DELAY variable to git-push-all.sh, which makes the script
push master and maint branches with a delay between each branch. These
delays trigger the CI jobs in a set order, which should show the most
likely failures first.
Also:
* make pushes atomic by default, and
* make the script pass any command-line arguments to git push.
Closes ticket 29879.
A .may_includes file can be "advisory", which means that some
violations of the rules are expected. We will track these
violations with practracker, not as automatic errors.
Our topological sort code really deserves a function of its own.
Additionally, don't print from inside the topological sort code:
instead, return a result, and let the caller print it.
I'll want to make this block into a series of functions in a
subsequent commit, but I'm doing this separately to get the
indentation change out of the way.
This branch will end up with making checkIncludes.py an integrated
part of practracker, for ticket 31176.
Padding circuits were regular cells that got closed before their padding
machine could finish. This means that they can still receive regular cells from
their past life, but they have no way or reason to answer them anymore. Hence
let's ignore them before they even get to the proper subsystems.
I'm using 500 as a file size limit, and 15 as an include limit.
This affects comparatively few files, but I think they are the worst
ones.
Closes ticket 31175.
This test runs practracker with a set of 0 thresholds, to make sure
that it enumerates all its values right. It tries running with an
empty exceptions file, and with an exceptions file that covers
_some_ of the data, and it makes sure that the outputs are as expected.
Now that there is only one toplevel place where we print problems,
we can redirect just that one print to a file when we are
regenerating the exceptions.txt file. Previously we redirected
sys.stdout, which is naughty, and forced us to send warnings (and
warnings alone) to stderr.
These flags let you suppress the message about the number of
problems and warnings, and let you control the thresholds above
which something counts as a problem.
I need this for testing.
Instead of having "consider" functions that have to call a global
ProblemVault, we can now generate all the metrics for the code
separately from the decision about what to do for them.
I'm about to refactor the code into a set of iterators that yield
*all* the metrics for the code, and then add a filter on top of that
to return the problems.
Previously, when TestingTorNetwork was set, we would manually adjust
the initvalue members of a bunch of other config_var_t, and then
re-run the early parts or parsing the options.
Now we treat the initvalue fields as immutable, but instead assign
to them in options_init(), as early as possible. Rather than
re-running the early parts of options, we just re-call the
options_init_from_string() function.
This patch de-kludges some of our code pretty handily. I think it
could later handle authorities and fallbacks, but for now I think we
should leave those alone.
When an exception is present, we can now violate the limit by a little
bit and only produce a warning. The strict flag overrides this
behavior.
I've given file sizes a 2% tolerances and function sizes/include
counts a 10% tolerance.
Part of 30752
Instead of excluding directories at the last minute if they happen
to appear in our filenames, we exclude them early, before recursing
into all their subdirectories.
Part of 29746.
The practracker_tests.py unit test file called a function by its old
name.
Also, practracker counted functions as starting one line after the
function name, and ending with the closing brace. Now they start
with the open brace and end with the closing brace.
Coverity doesn't understand that if begin_cell_parse() returns 0 and
sets is_begindir to 0, its address field will always be set.
Fixes bug 30126; bugfix on 0.2.4.7-alpha; Fixes CID 1447296.
