Hi Bruce, Chet On Thu, 11 Dec 2025 at 18:27, Bruce Jerrick <[email protected]> wrote: > >> It would be nice if there were an option to the 'read' builtin to echo >> dots rather that nothing at all > > I thought I'd take a swing at this.
Try this patch: https://github.com/kurahaupo/bash/tree/patch-echosubst This defines a new bind option "echo-substitution", which is consulted when read is passed both the '-s' (echo-suppression) and the '-e' (use readline) options, like: bind "set echo-substitution $SUBSTITUTION" read -es VAR where SUBSTITUTION can be: 1. “:none” (or empty) - the current behaviour, showing nothing; 2. “:random” (randomly chosen ASCII chars); 3. “:auto” - selects “•” when in a UTF-8 locale, or “*” otherwise; 4. one multi-byte or single-byte character, either literally or as U+xxxx; 5. any string of printable ASCII characters (not starting with a colon). Examples: The existing behaviour is available (and is the default): $ bind "set echo-substitution :none" $ read -esp 'Quiet> ' VAR *# now type "Hello World"* Quiet> $ echo "$VAR" Hello World This is how I interpreted the original feature request: $ bind "set echo-substitution ." $ read -esp 'PlainDots> ' VAR *# now type "Hello World"* PlainDots> ........... $ echo "$VAR" Hello World In hindsight perhaps it meant this: $ bind "set echo-substitution U+2027" $ read -esp 'SmallDots> ' VAR *# now type "Hello World"* SmallDots> ‧‧‧‧‧‧‧‧‧‧‧ $ echo "$VAR" Hello World Or this: $ bind "set echo-substitution U+2022" $ read -esp 'BigDots> ' VAR *# now type "Hello World"* BigDots> ••••••••••• $ echo "$VAR" Hello World Old-school folk might prefer this: $ bind "set echo-substitution U+002a" $ read -esp 'Stars> ' VAR *# now type "Hello World"* Stars> *********** $ echo "$VAR" Hello World More fanciful options include: $ bind 'set echo-substitution U+ff0a' $ read -esp 'FatStars> ' VAR *# now type "Hello World"* FatStars> ＊＊＊＊＊＊＊＊＊＊＊ echo "$VAR" Hello World $ bind "set echo-substitution YabbaDabbaDoo" $ read -esp 'Fred> ' VAR *# now type "Hello World"* Fred> YabbaDabbaD $ echo "$VAR" Hello World $ bind $'set echo-substitution .\u2027\u2022 *\uff0a' $ read -esp 'Hybrid> ' VAR *# now type "Hello World"* Hybrid> .‧• *＊.‧• * $ echo "$VAR" Hello World $ bind "set echo-substitution :random" $ read -esp 'Gibberish> ' VAR *# now type "Hello World"* Gibberish> 9H($Yto`~<. $ echo "$VAR" Hello World For now, multi-byte characters are disabled unless in a UTF-8 locale. Some attempts are made to prevent using non-printable characters as substitutes, but this isn't fully implemented. (I note that readline already seems to recognise double-width characters, so it must already be linked with libunicode or have some equivalent mechanism, so this might not actually be much more work.) -Martin PS: while testing this patch, I discovered that UTF-8 locale detection is only done when readline starts up. This means that “LC_ALL=C bind 'set echo-substitution :auto'” won't necessarily pick “*”. PPS: it's not just "read"; it works in any readline instance whenever "stty -echo" is in effect, including the shell prompt.
