Providing a non-nil universal argument to `db/load-known-ssh-keys` now readds
all known SSH keys to the current SSH agent, irregardless of whether they are
already present or not.
This will allow to bind `db/org-add-link-to-other-item` with different initial
inputs to different keys to have often used selection directly at hand.
Setting the initial input is a bit tricky, though, because
`org-refile-get-location` does not allow to set it. Instead, we have to
temporarily overwrite `completing-read` with the corresponding parameter set
directly.
This is supposed to inhibit this warning:
```
Warning: Eager macro-expansion skipped due to cycle:
… => (load "db-utils.el") => (macroexpand-all (defalias 'db/sync-magit-repos-from-projectile …)) => (macroexpand (eval-when-compile …)) => (load "db-utils.el")
```
It should make calling this function also more robust.
Also updated the implementation to be more “dash-y”.
The values in the variable `org-time-stamp-formats` do not contain angle
brackets anymore. Using the function `org-time-stamp-format` is more robust
anyway and is also the official way to retrieve this format.
This is meant to mitigate an issue with table alignment when those tables are
results of source blocks, contain links in cells and are hidden (i.e., folded):
in those circumstances, the table alignment seems to be broken, because the cell
width does not take into account that links are shorter than their textual
representation.
I am not quite sure whether this is a bug in my configuration or a bug in Org
mode itself. The function to compute cell width is `org-string-width`, but
maybe this function does not get passed the correct input when tables are folded
and need to be aligned?
This still always decrypts the password, because `ssh-add` does not check
whether the key is already present in the current agent. This should propably
done separately before calling `ssh-add`.
This is to prepare computing the password only when needed, i.e., when the key
is not already included in the running agent. We are not there yet, though.
I am not sure why it hasn't been working until now, and I am not sure why it
indeed does work now, but it does, and that sufficient for me – at least for
now.
This seems to be fast enough, and since I often find myself aborting a current
backlink insertion query because I missed to include all Org mode buffers,
making it the default only seems to be reasonable.
When not provided in dynamic workload reports, `start-date` and `end-date`
should be treated as unconstrainted. However, instead dates were queried
interactively in that case, because `org-read-date` is being used to normalize
date display. Fixed that.
We must widen the current restriction when searching for the template item, not
when copying the body from it. In the latter case, `org-with-point-at` will
handle the necessary widening.
When the template associated with the item at point contains sub-items itself,
those may have been assigned ID or CUSTOM_ID properties. Copying the template
would thus duplicate these properties, violating their implicit uniqueness
constraint, so we now remove those properties from the inserted checklist
template.
When a checklist template contains headings on its own, the
`CHECKLIST_INSERTED_P` property until now would be inserted at the last heading
in this template, instead of at the heading where the template is supposed to be
inserted in the first place. Fixed this.
Displaying the current filters in the mode line alone often goes unnoticed for
me, so I need a more direct display. Let's add it to the first structural
header.
Mostly moving functions around and updating page headers. Also removed some
obsolete functions, see updated command list for `db-org` use-package
declaration.
