bug#59082: 28.2; Undocumented `intern-soft` feature with shorthands symbols

[João Távora]

We can/should of course document the decision that they aren't supported in rubbed in the manual though (if we make that decision)...

João

On Sat, Nov 12, 2022, 16:46 João Távora <joaotavora@gmail.com> wrote:

Hmm, not sure we should be using shorthands in docstrings, which aren't read by the Lisp reader, but by something else... This is what is at stake, right?

Maybe if good reasonable semantics can be found, but not sure.

On Sat, Nov 12, 2022, 16:20 Thierry Volpiatto <thievol@posteo.net> wrote:

João Távora <joaotavora@gmail.com> writes:

> This is a feature of Lisp in general and the correct way to go from
> strings to symbols.

Great, thanks to confirm.

> Curiously, I was pleasantly surprised that much code of key symbol
> processing facilities was already using this indirection and
> shorthands automatically worked in those facilities because of that.

`substitute-command-keys` at least doesn't handle this.

Say you have the file foo.el with
`read-symbol-shorthands` == (("f-" . "foo-")) and containing definitions
like `f-dothis`, `f-dothat` etc... and a var `f-help-string` containing string like
"\<f-map>\[f-dothis]: Description", if you define a function `f-help`
containing (substitute-command-keys f-help-string) it will fail
complaining `f-dothis` is void.

--
Thierry