Recent commits to derivepassphrase.git (2b1ca491019b2366b079155ad2a40353814c21e2)

Merge topic branch 'usage-fixes' into master

2025-01-07T23:42:07+01:00

* t/usage-fixes:
  Incorporate revised wording from the manpage into the README
</pre>

Incorporate revised wording from the manpage into the README

2025-01-07T23:36:49+01:00

The revised wording spells out the strength guarantee of the derivation
scheme and the purpose of the configuration file more explicitly.

We intentionally keep the introductory sentences (one-line description
and first sentence of the first explanatory paragraph) different,
because they have different contexts to work with; in particular, the
manpage contains a synopsis section and can refer to the SERVICE metavar
in its one-line description, whereas the README cannot, and should also
mention SSH keys early.
</pre>

Fix omissions of `tomli` (Python 3.9 and 3.10) and PyPy in the README

2025-01-07T23:27:39+01:00

</pre>

Add remaining re-linting changes (part 2 of 2)

2025-01-07T22:44:39+01:00

The remaining re-linting changes pertain to new linting codes:

  * TC006/runtime-cast-value (formerly TCH006) suggests that cast calls
    use stringified type declarations.  Implemented.

  * B033/duplicate-value requires sets not to contain duplicate values.
    This is used in the `_types.js_truthiness` function to both test
    against and to document all of the falsy values except NaN.
    Ignored.

  * RUF043/pytest-raises-ambiguous-pattern requires patterns in the
    `match` argument of `pytest.raises` to be either raw strings, or
    passed through `re.escape`.  Implemented.

  * RUF046/unnecessary-cast-to-int forbids explicitly casting the result
    of `math.ceil` to int.  I believe `mypy`/`typeshed` originally had
    problems with this, but this no longer appears to be the case.
    Implemented.
</pre>

Update ruff to v0.8.x, reformat and re-lint (part 1 of 2)

2025-01-07T22:44:39+01:00

We update `ruff` to v0.8.x and reformat the sources.  Most
conspicuously, `assert` statements typically no longer break the
condition to test, and rather push the error message to the next line
instead.  Also, some implicit string concatenations are undone if they
*do* fit onto a single line.

We also re-lint the sources, and commit part of the linting fixes.
Specifically, omissions in the docstring (raised exceptions, return
values) are now marked as linting errors on the docstring itself,
instead of at the omitted object (`raise` or `return` statement).
Alternatively, we add the missing info to the docstring.  We also commit
linting fixes regarding improper usage of underscores in variable names
(leading underscore instead of trailing underscore on temporary
variables).  Finally, we correct some improper calls to the logging
system (root logger instead of application logger).
</pre>

Format everything with ruff (v0.6.9)

2025-01-07T22:44:39+01:00

</pre>

Merge topic branch 'i18n-fixes' into master

2025-01-07T22:44:15+01:00

* t/i18n-fixes:
  Add more translator's comments
  Fix missing context in CANNOT_UNDERSTAND_AGENT translatable string
  Fix phrasing of "Cannot update %s settings" message
</pre>

Add more translator's comments

2025-01-07T19:54:00+01:00

... for diagnostic labels, introductory paragraphs of help texts, and
error messages of the form "user aborted operation".
</pre>

Fix missing context in CANNOT_UNDERSTAND_AGENT translatable string

2025-01-07T19:52:31+01:00

</pre>

Fix phrasing of "Cannot update %s settings" message

2025-01-07T19:46:51+01:00

The "Cannot update the global/service settings without actual
settings" messages contained "global" and "service" as (English) format
string parameters, and the actual translatable string contained no
indication what this placeholder actually meant.  This is of course
a horrible way to request a translation for those two messages.  We now
include translatable strings for each of the two possible values of the
metavar ("global settings" and "service-specific settings"), and proper
translation advice for the metavar value and the original translatable
strings.

The manpages have been updated as well, to reflect the slight change in
phrasing of the message.
</pre>