| Age | Commit message (Collapse) | Author |
|---|
|
|
|
This makes sure the assumption that the string only ever contains a
single character holds a priori (or otherwise we'll run into a failure
at eval time). Not strictly necessary, but a small code improvement?
|
|
Using a plain space in emacs key binding syntax can be problematic
(especially making users unable to type "s" if the mode map prefix
is "s-"). Therefore we must make sure to convert the strings we have in
bqn-symbols.el into proper emacs key syntax. Luckily we can use
single-key-description to do exactly that.
|
|
This is a small refactor removing sort-of-dead code. This of course
removes the possibility to define multiple keys for a certain BQN
character, but I'm not sure this is an actual use case we need to
preserve.
|
|
bqn--make-bqn-mode-map depended on the value of bqn-mode-map-prefix, but
bqn-mode-map-prefix's set function also used bqn--make-bqn-mode-map. To
break this cycle which made emacs' warnings act up a bit, we remove
bqn--make-bqn-mode-map and defvar bqn--mode-map to nil initially,
trusting that it'll be correctly re-defined by the initialization of
bqn-mode-map-prefix via its set function.
I verified that this works correctly both in the case of the initial
value (although I swapped out the value for "M-" beforehand since in my
setup Super doesn't seem to work at all) and redefining it using
costumize-variable.
|
|
A docstring wider than 80 characters causes a warning when compiling
which isn't too much trouble to clean up.
|
|
When byte compiling bqn--make-key-command-sym wouldn't be defined when
it's used in make-insert-functions directly below. This can be fixed by
forcing the defun to be evaluated at compile-time using eval-and-compile
while still making sure it is also available when loaded normally.
|
|
|
|
|
|
|
|
|
