GNU bug report logs - #73499
31.0.50; sql-interactive-mode problem with package caps-lock-mode

Previous Next

Full log

View this message in rfc822 format

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stephen Berman <stephen.berman <at> gmx.net>, Stefan Monnier <monnier <at> iro.umontreal.ca>, Stefan Kangas <stefankangas <at> gmail.com>, Andrea Corallo <acorallo <at> gnu.org>
Cc: 73499 <at> debbugs.gnu.org
Subject: bug#73499: 31.0.50; sql-interactive-mode problem with package caps-lock-mode
Date: Sat, 28 Sep 2024 14:14:42 +0300

> Date: Thu, 26 Sep 2024 14:44:12 +0200
> From:  Stephen Berman via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs <at> gnu.org>
> 
> 0. Install the GNU ELPA package caps-lock.
> 1. emacs -Q
> 2. M-x package-initialize
> 3. Start an SQL inferior process, e.g. with `M-x sql-sqlite RET test RET'.
> 4. In the *SQL: SQLite* buffer type `M-x caps-lock-mode' and at sqlite>
>    prompt type "select * from".
> => The typed text appears as this: "SELECT * FRoM", with a lowercase "o"
>    instead of an uppercase "O".
> 
> This is because sql-interactive-mode-map binds both "o" and "O" to the
> command `sql-magic-go', but caps-lock-mode changes case only if the last
> key event either invoked `self-insert-command' or
> `isearch-printing-char' or if the command it invoked remaps
> self-insert-command, and since neither is the case here, no case change
> occurs.  (You can of course enter "O" by typing S-o, but caps-lock-mode
> is supposed to obviate the need to do that.)  This problem also happens
> with PostgreSQL and presumably other SQL interpreters supported by
> sql.el (but I have only tested SQLite and PostgreSQL).
> 
> I have found two ways to fix this problem.  One is simply to add
> `sql-magic-go' to the list `caps-lock-commands'.  In caps-lock.el this
> is a defvar with the value '(self-insert-command isearch-printing-char).
> Instead of just changing this value in the source code, I think it would
> be better to make the variable a defcustom, since many packages bind
> letter keys to commands other than `self-insert-command' but may
> nevertheless want to change the case of these letters when
> caps-lock-mode is enabled.  So it would be up to the user to add
> `sql-magic-go' to the value.  Here's a patch that allows this (I don't
> have a local clone of the GNU ELPA repo, so it's just a plain diff):
> 
> 
> The alternative fix is to change the implemention of `sql-magic-go' so
> it doesn't interfere with caps-lock-mode; this would absolve the user
> from the responsability of customizing which commands trigger case
> change, but requires more invasive changes to sql.el (at least I haven't
> come up with a simpler change), specifically, (i) removing the bindings
> of "o" and "O", (ii) making `sql-magic-go' a function instead of a
> command, which checks if the last keyboard input was either of these
> letters and if so, does what `sql-magic-go' is supposed to do, and (iii)
> adding `sql-magic-go' to `post-self-insert-hook' if `sql-electric-stuff'
> is customized to give the string "go" special handling.  Here's the
> patch:

Stefan, Andrea, Stefan: any opinions on which way is better?

Thanks.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.