| Commit message (Collapse) | Author | Age |
| ... | |
| | |
| |
| |
| |
| |
| |
| |
| | |
* ADD 16384-bit public half of offline certification-only key.
* ADD 4096-bit public halves of online signing+encryption keys.
* ADD gnupghome/gpg.conf file with settings (both for the fake testing key
and the real keys).
* ADD gnupghome/* to the .gitignore.
|
| | |
| |
| |
| |
| |
| | |
* ADD GPG keys for testing purposes. They are expired, passwordless, marked
with the UID "TESTING (INSECURE!) <bridges@bridges.torproject.org>". The
(INSECURE!) part tells GnuPG to not waste real entropy when using this key.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
* REMOVE RUN_IN_DIR
* CHANGE EXTRA_INFO_FILES to a list with the correct filenames.
* CHANGE HTTPS options to bind to 127.0.0.1:6789 by default (for testing).
* ADD missing HTTP_N_BRIDGES_PER_ANSWER option.
* CHANGE to enabling bridge fingerprints by default.
* CHANGE to GPG signing emails by default.
|
| | |
| |
| |
| | |
* FIX also the markdown formatting of indented blocks in some places.
|
| | |
| |
| |
| |
| | |
We usually have 'cached-extrainfo' and 'cached-extrainfo.new', not
'extra-infos'.
|
| |\ \ |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
These aren't sanitised -- they are completely faked. There has been so much
confusion over what these things look like...I really just want something that
I can point to. I (or anyone else) should update the metrics dataformat pages,
as well as perhaps provide better clarification in tor-spec.git for the lines
whose format is unclear.
* ADD docs/DESCRIPTORS.md.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
* ADD better descriptor generator, it makes descriptors according to
dir-spec.txt now. They are even signed with OpenSSL generated RSA keys, and
have all the embedded document hashes and fingerprints of the correct keys
used for signing and everything.
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The main entry point and server reload function are refactored to recieve the
parsed options from the commandline script, scripts/bridgedb.
It is necessary to pass *args to `_reloadFn()` (as well as its original
stubbed lambda function), because the arguments are technically unknown to the
scope of the `reload()` function embedded in `bridgedb.Main.startup()` when it
is evaluated at compile time (they are not known until later when
`bridgedb.Main.run()` is called via scripts/bridgedb).
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
These were the result of using indexless format string syntax, which was
introduced in Python-3.1 (and then it was backported to some versions of 2.7,
but not 2.6, apparently).
* FIXES a couple print statement syntax in lib/bridgedb/Tests.py causing test
failure.
The failing test was:
- testDistWithFilterBlockedCountriesAdvanced (bridgedb.Tests.IPBridgeDistTests)
|
| | | |
| | |
| | |
| | | |
in CI.
|
| | | |
| | |
| | |
| | |
| | |
| | | |
* ADD script for generating a self-signed SSL certificate for testing.
* CHANGE dependency installs in .travis.yml to ignore suggestions and
recommendations.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
* ADD lib/bridgedb/runner.py for running trial/unittest tests.
* ADD scripts/gen_bridge_descriptors.
* Fix lib/bridgedb/opt.py by coercing a commandline parameter to a string;
it's sometimes None or an int.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
And Don't run trial-based tests until #9199's tests are merged.
(cherry picked from commit 31cff62c026a94edd114a5adff2848899569c5f6)
Created a new branch in order to drop commit
826f872e075a639250d565a8d93b1c405256f82d.
Signed-off-by: Isis Lovecruft <isis@torproject.org>
|
| |/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Three unittests were failing in lib/bridgedb/Tests.py due to changes made for
users). They were:
- testDistWithFilterBoth (bridgedb.Tests.IPBridgeDistTests)
- testDistWithFilterIP4 (bridgedb.Tests.IPBridgeDistTests)
- testDistWithFilterIP6 (bridgedb.Tests.IPBridgeDistTests)
These were failing due to chopping the first 7 bytes off the the returned
config lines, because they expected to see the "bridge " prefix and it's not
there anymore.
|
| | | |
|
| | | |
|
| | |
| |
| |
| | |
(cherry picked from commit e2a7f9319310a3c23a0c110963071cb1f4592c76)
|
| | |
| |
| |
| |
| | |
* ADD .pylintrc to tweak pylint a bit; otherwise it warns about some rather
ridiculous things.
|
| | | |
|
| |\ \ |
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | | |
* FIXES all the distutils hellspawn bugs arising from the disturbed grave of
distribute. Everything is purely setuptools now.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* CHANGE ``message_extractors`` to use the ``pkgpath`` variable. The less
places in the code with arbitrary hardcoded paths which may change at some
point, the better.
* ADD an XXX note about the glob pattern ``public/**`` in
``message_extractors``.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* CHANGE ``exclude_package_data`` to specify the package explicitly. This
seems to fix a bug where .po files were still being installed.
* ADD an exclusion for all .pot files.
* FIXES a bug where lib/bridgedb/i18n/templates/bridgedb.pot was being
installed by setuptools.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The ``get_data_files()`` function now takes a boolean ``filesonly`` parameter,
if True, it returns a list suitable for the ``package_data`` parameter of
``setuptools.setup()``, otherwise it returns a list which is compatible with
``distutils.command.install_data.install_data.data_files``.
We can now remove distutils entirely from setup.py, but I decided to make this
function compatible with both distutils and setuptools because Python
packaging has been in turmoil for a while now. (Hopefully, it's done, but I'm
not holding my breath.)
* FIXES a problem with some .mo files not getting installed (if they didn't
already have a directory for their LOCALE in build/).
* FIXES a bug where all .po files were being installed, even though they were
added to the ``exclude_package_data`` parameter of ``setuptools.setup()``.
* FIXES (along with commit ef784165002086cbc5b8131edcde815326a63009) a bug
where the robots.txt file was not being packaged correctly.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* ADD ``glob`` dependency to setup.py.
* ADD ``get_template_files()`` function to setup.py, which simply globs for
filenames which match various directory patterns, in order to obtain files
which should be installed for BridgeDB's HTTPS distributor.
* FIXES a bug where robots.txt was not being installed as a resource file.
|
| | | |
| | |
| | |
| | | |
Might as well, while we're at it.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* ADD function ``get_supported_langs()`` to setup.py, which determines which
completed, compiled translations files are present in our source tree, and
stores three things:
1. The LOCALE code for the translation.
2. The path of the directory to install the files to, relative to the
current ``sys.prefix``. (``sys.prefix`` is where Python files are
going to be installed, and it is set by the OS's interpreter first,
and then occaisionally changed according to various factors, like
whether or not we are working in a virtualenv, if the package is being
installed from a DCVS/egg/tarball/zipfile, if the package is being
installed via pip or via a package manager, etc.)
3. A list of files for that LOCALE which should be installed. In our
case, this should only ever be one <LOCALE>/LC_MESSAGES/bridgedb.mo
file per LOCALE.
These three things allow us to tell both setuptools (and the underlying
distutils mechanisms) how/where to install everything.
The next thing that ``get_supported_langs()`` does is take the list of
collected LOCALEs which are to be installed, and rewrite a line in the
following new module:
* ADD a new module, lib/bridgedb/_langs.py, which has one variable,
``supported``, which is a list of two-letter country codes of languages
installed. When "python setup.py install" is run (or any setup.py command,
for that matter) the ``supported`` variable in lib/bridgedb/_langs.py is
rewritten to be the current list of languages for which we have completed,
compiled translation files.
* ADD function ``get_langs()`` to lib/bridgedb/_langs.py. This simply fetches
the current value for ``supported``.
* ADD a package attribute, ``bridgedb.__langs__``, which uses ``get_langs()``
to store, in the installed version of BridgeDB, the list of supported
languages which were also installed. :)
* ADD an entry for ``lib/bridgedb/_langs.py`` into .gitignore, so that a
locally rewritten file is never committed.
|
| |/ / |
|
| |\ \ |
|
| |/ / |
|
| |\ \ |
|
| | | |
| | |
| | |
| | |
| | | |
* REMOVE one EOL whitespace, a blankline with extra whitespace.
* FIX a tab before a space.
|
| |/ / |
|
| | |
| |
| |
| |
| | |
* CHANGE Makefile directives 'install' and 'force-install' to recompile the
completed translations before actually installing the package.
|