Re: [PATCH] Fix READLINE_ARGUMENT with "M-1 M-2" and with `vi-arg-digit'

On 6/27/21 8:12 AM, Koichi Murase wrote: Description: Now "bind -x" can be used with numeric arguments when the numeric arguments are specified by e.g. "M-1 2 3". However, when the numeric arguments are specified by e.g. "M-1 M-2 M-3", it still fails with the error message `bash_exe

Re: [PATCH] Fix READLINE_ARGUMENT with "M-1 M-2" and with `vi-arg-digit'

2021年6月29日(火) 4:33 Chet Ramey : > > On 6/27/21 8:12 AM, Koichi Murase wrote: > > Description: > > > >Now "bind -x" can be used with numeric arguments when the numeric > >arguments are specified by e.g. "M-1 2 3". However, when the > >numeric arguments are specified by e.g. "M-1 M-2 M-3

Re: [PATCH] Fix READLINE_ARGUMENT with "M-1 M-2" and with `vi-arg-digit'

> > Thanks for the report. Wow, do people use M-1 M-2 to specify multiple- > > digit arguments? > > [...] I use the latter, and in that case, it is intuitive > to input multiple-digit numbers with the Meta key kept to be pressed. I forgot to mention that the terminal will send "M-1 M-2" in this ca

Re: [PATCH] fix super-linear complexity of ${v^} and ${v//A}

On 6/27/21 2:13 AM, Koichi Murase wrote: Description: The time complexities of ${v/...} and ${v^...} are expected to be linear with respect to the length of the variable contents. However, the time complexity of ${v/...} is quadratic if there is no replacement when the length is larg

shell-backward-kill-word needs to behave more like werase

The default werase setting erases the series of white spaces before the caret, then it erases the preceding word (without touching the spaces before the word). I expected that shell-backward-kill-word would do the same with the difference that the entire quoted argument would be erased. However,