Recent commits to derivepassphrase.git (f81c616548e53fe088636bf9eb6fbaf7873c8c25)

Document recent changes to notes handling in the changelog

2025-02-07T19:09:54+01:00

</pre>

Optionally support printing vault service notes before the passphrase

2025-02-07T16:01:14+01:00

When deriving a passphrase, if the service has any notes, vault(1)
prints the notes after the passphrase.  However, akin to source code
comments before the code in question, service notes may actually be
better placed above the derived passphrase instead, for some types of
notes.  Support such placement via a new command-line option
`--print-notes-before`; the default, vault(1)-compatible behavior can be
explicitly selected via `--print-notes-after`.
</pre>

Document the modern editor interface

2025-02-07T15:19:22+01:00

Update the manpages to include the corrected and extended handling of
the `--notes` config option.  That is, document the
`--modern-editor-interface` and `--vault-legacy-editor-interface`
options, and the new warning messages about `--notes` requiring
`--config` and that a backup copy of the notes have been made.

Some minor adjustments to the vault(1) legacy editor interface warning
messages were made to ease inclusion in the manpage while making the
intent clear.  Additionally, derivepassphrase with the legacy interface
correctly fills in the old notes contents if any, instead of
unconditionally filling in the placeholder text.
</pre>

Reintroduce a "modern" editor interface à la git-commit or git-rebase

2025-02-06T14:27:21+01:00

Reintroduce the previous git-like editor interface as the "modern"
editor interface to the "vault" subcommand, and provide explicit options
`--modern-editor-interface` and `--vault-legacy-editor-interface`,
defaulting to the latter for compatibility.

Furthermore, for the legacy interface, introduce a backup copy of the
notes to guard against accidental data loss, because the legacy
interface offers no way to abort mid-editing, making fatal mistakes all
the more likely.
</pre>

Warn the user when `derivepassphrase vault --notes` is ineffective

2025-02-06T13:16:37+01:00

In 7b0f4e121a5e688b59abad73b8b60bacfc7d02ed, we adjusted
`derivepassphrase` to only act upon `--notes` if `--config` is also
given, for compatibility with vault(1).  However, this is incompatible
with previous versions of `derivepassphrase`, and potentially confusing
even to vault(1) users.  As such, `derivepassphrase vault` now emits
a warning if it is called with `--notes` but without `--config`.
</pre>

Use a vault(1)-compatible naive notes editing system

2025-02-05T14:44:16+01:00

We originally provided a modern git-like editor interface, with a marker
for explanatory texts and with support for aborting the edit by leaving
the file empty.  vault(1) however takes the contents of the notes file
literally.  For compatibility, we must do the same.
</pre>

Use hypothesis for notes handling tests

2025-02-05T14:40:56+01:00

Have the existing or the new notes be generated data (whichever is
appropriate).

Also handle the previously untested case of aborting the notes edit when
there are no stored notes yet.  *This* test does not use hypothesis.
</pre>

Adapt notes printing test to arbitrary notes

2025-02-05T12:07:24+01:00

</pre>

Clean up notes testing functions

2025-02-04T13:30:22+01:00

They had similar bad defaults and easily avoidable linting overrides
like the no-op test.
</pre>

Use a better test for no-op editing notes

2025-02-04T13:17:03+01:00

"no-op" editing of notes makes no sense for a service which does not
have notes yet.  Also, "no-op" editing may involve extra whitespace at
the beginning or end of the contents.  Use these two facts to construct
a better test.

Also, do not use click's convenience feature of using a `None` return to
signal that no changes have been made.  The check is heuristic in nature
(it involves filesystem timestamps), and vault(1) actually uses a more
naive system that does not map cleanly to click's system.
</pre>