On Thu, 9 Jul 2026 at 02:50, Chet Ramey <[email protected]> wrote: > The issue is that one of readline's basic assumptions is being violated.
Why does readline still need to make that assumption? There are at least two approaches that can remove that assumption, with many specific implementations suggested over the years. A fix could be built into readline, or shipped with Bash part of a "recommended" `/etc/skel/.bashrc`;(*1) then the complaints go away for all future Bash releases.(*2) Even a fix that only works for ANSI-compatible terminals would be better than none. On 7/7/26 11:59 PM, Martin D Kealey wrote: > > I don't think it's fair to give up and say “user error” when they're > likely not the author of the command in question. > > No one is saying it's "user error." Saying it, maybe not. Strongly implying it, most definitely. We're currently saying it's the user's responsibility to do "something" different ‒ either press enter when they don't like the broken prompt, or install a script they did not possess until they got a reply to their bug report. Shifting blame and implying an error of omission (when we should reasonably *expect* their ignorance) is gaslighting. Let's have more empathy for our novice users. -Martin (*1: I suggest shipping "one file per tweak", named something like /usr/share/bash/bashrc.d/100_prompt_after_incomplete_output.) (*2: Contrast that with the current approach of telling each individual user to apply a magic incantation, and we're setting them up for failure when their environment changes.)
