https://git.schokokeks.org/derivepassphrase.git/tree/ef40e2a2624047f1b02e87716f17d8569dc94bfb 2024-09-01T11:17:19+02:00 tag:gitlist.org,2012:commit/ef40e2a2624047f1b02e87716f17d8569dc94bfb 2024-09-01T11:17:19+02:00 Marco Ricci m@the13thletter.info <pre>* t/clean-up-and-update-tool-configuration: Upgrade ruff to 0.6, and redo configuration and rule selection Undo single-child hatch environment inheritance Specify project dependencies more clearly and explicitly Sort sections in pyproject.toml &lt;/pre&gt; tag:gitlist.org,2012:commit/aacd09bdcbdb01df7cb819396727d2427636b144 2024-09-01T11:08:47+02:00 Marco Ricci m@the13thletter.info <pre>* t/issue9-allow-all-unicode-passphrases: Allow all textual strings, but warn on unnormalized ones GitHub: Closes #9. &lt;/pre&gt; tag:gitlist.org,2012:commit/b4d8439fa4207b665ad8ea2217f21f807f603734 2024-09-01T11:06:43+02:00 Marco Ricci m@the13thletter.info <pre>* t/issue1-vault-config-exporter: Fix formatting and linting issues in exporter tests Add Changelog entry for the vault exporter Rename vault v0.2/v0.3 classes, and fix API weirdnesses and test coverage Add more tests of the storeroom format exporter's internals Test exporter data loading functionality more robustly Move exporter command-line interface into a separate module Add preliminary tests for the exporter Add command-line interface to the exporter Move vault key and path detection logic into the exporter module Add prototype for "vault v0.2/v0.3"-type data export Support exports from outside the current directory Add dependency on cryptography for the exporter Add an actual storeroom exporter, not just a reader Add docstrings and better variable names on storeroom exporter Add example for "storeroom"-type data export Add prototype for "storeroom"-type data export GitHub: Closes #1. &lt;/pre&gt; tag:gitlist.org,2012:commit/fb6913b190365893f3d5e83225a496c9061c6e54 2024-09-01T11:05:14+02:00 Marco Ricci m@the13thletter.info <pre>&lt;/pre&gt; tag:gitlist.org,2012:commit/e109f3ac7c5c09b89816405e3ebc9dd195685024 2024-09-01T10:22:13+02:00 Marco Ricci m@the13thletter.info <pre>hatch currently still defaults to ruff 0.4.5, which includes some now-deprecated rules and lacks some others. Furthermore, hatch's default configuration for ruff is very opinionated and not explained anywhere. Upgrading the ruleset to 0.6 still incurs the overhead of having to look at every single new ruff rule, determining whether it fits the hatch configuration, and allowing it. This is a similar amount of effort to enabling all of ruffs rules and deciding, based on the flagged violations, whether a rule or rule category is worth keeping. This has now been implemented: almost all rules are now selected, and the exceptions are documented in the project settings. (ruff actually includes documentation on linting rules that tend to get violated if ruff is also used for reformatting.) hatch is configured to use ruff 0.6 and this user configuration, ignoring its own default ruff configuration. Some further ruff settings are necessary, e.g. adapting the default copyright text template to match hatch's copyright text template, as is one new mypy setting (the "ignore-without-code" error). Note: the new ruleset has not actually been applied to the codebase yet, because the effect would be spread out across the codebase, and we expect multiple topic branches to be merged soon. &lt;/pre&gt; tag:gitlist.org,2012:commit/94068756e896574a397b4717388b02ed661a22a9 2024-09-01T00:32:49+02:00 Marco Ricci m@the13thletter.info <pre>Remove the "check if Unicode string has unambiguous byte encoding" misfeature, which is based on an incorrect understanding of Unicode normalization forms. Also remove the `AmbiguousByteRepresentationError` class. Now we allow all textual strings as master passphrases. When storing a (master) passphrase to the configuration or when interactively generating a passphrase, we emit a warning if we detect that the master passphrase is not normalized (with the configured normalization form, stored in `.global.unicode_normalization_form`). &lt;/pre&gt; tag:gitlist.org,2012:commit/5022fff219bf1bc30988d3070e564ad058f0d5c4 2024-08-31T22:09:41+02:00 Marco Ricci m@the13thletter.info <pre>&lt;/pre&gt; tag:gitlist.org,2012:commit/0169d5fbbb766fef119548e8b07cd0e2bcd43ac6 2024-08-31T21:26:49+02:00 Marco Ricci m@the13thletter.info <pre>The `Reader` classes are now named `VaultNativeConfigParser`, because `Reader` is a very non-descript name. The interface is now somewhat more pythonic, and uses fewer internal branches. Further tests of the internals have been added, so the module has 100% test coverage. &lt;/pre&gt; tag:gitlist.org,2012:commit/349f20ebdbc4c3bd4e4fc9e8454dc10b5d081396 2024-08-31T21:24:29+02:00 Marco Ricci m@the13thletter.info <pre>Further tests are necessary to achieve 100% coverage, but these concern the detection of format errors, so the test files are rather hard to generate. &lt;/pre&gt; tag:gitlist.org,2012:commit/fa6125de4455fb187db8888e510400e9e9820b3b 2024-08-31T21:04:32+02:00 Marco Ricci m@the13thletter.info <pre>Move the `_load_data` function, which deals with determining the correct vault storage format (with automatic rollover) into the top level, where the function can be tested in isolation, and stubbed out if necessary. Also, use the environments with disabled cryptography support to test for correct failure behavior of the exporter. Use `importlib.util.find_spec` instead of an `import` statement to check for importability of a module, which does not require handling of `ModuleNotFoundError` if restricted to top-level modules/packages. &lt;/pre&gt;