From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4895 invoked by alias); 25 Oct 2014 08:17:35 -0000 Mailing-List: contact zsh-users-help@zsh.org; run by ezmlm Precedence: bulk X-No-Archive: yes List-Id: Zsh Users List List-Post: List-Help: X-Seq: 19295 Received: (qmail 13514 invoked from network); 25 Oct 2014 08:17:33 -0000 X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on f.primenet.com.au X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham version=3.3.2 Date: Sat, 25 Oct 2014 10:17:32 +0200 From: Eric Smith To: zsh-users@zsh.org Subject: Re: spell check on the command line Message-ID: References: <141024083614.ZM20933@torch.brasslantern.com> <141025002453.ZM22210@torch.brasslantern.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <141025002453.ZM22210@torch.brasslantern.com> User-Agent: Mutt/1.5.21 (2010-09-15) Looks great Bart, thank you! I cannot find the application "suggest" anywhere, what do you suggest? Eric Bart Schaefer wrote on Sat-25-Oct 14 9:24AM > On Oct 24, 11:19pm, Eric Smith wrote: > } > } ~> setopt CORRECTALL > } ~> touch this_settting_doess_not_yet_check_the_spelling_of_these_tags_in_a_filename.txt > > Oh, so you want a regular spelling dictionary applied in some way, to > substrings within each command-line argument. Or perhaps only to > substrings within file names, but since it's not possible to tell in > general which arguments are file names (a shortcoming correctall has > already), it's effectively the same problem. > > Let's ignore the complications of shell keywords like "elif", complex > commands like loops, and multi-line buffers. Your basic choices are: > > - override the accept-line widget; or > - create (or add to if already using one) a zle-line-finish widget; or > - attempt to handle it all in prexec. > > The simplest one is zle-line-finish so I'll do a quick example of that > here. If you do it in preexec it's more difficult to implement the > (a)bort and (e)dit cases. > > I'll assume that anything matching [[:punct:]] is taken as a word break, > and also that a program or function "suggest" exists that will spit out > pairs of wrong words and their replacments (unlike "spell" which emits > only the wrong words). > > zle-line-finish() { > local spelt misspelt nyae > print -R ${BUFFER//[[:punct:][:digit:]]/$'\n'} | suggest | > while read misspelt spelt > do > read -k 1 nyae$'?\n'"correct '$misspelt' to '$spelt' [nyae]? " > case ${nyae:l} in > (a) zle send-break;; > (e) zle push-line; return;; > (y) BUFFER=${BUFFER/$misspelt/$spelt};; > esac > [[ -n $nyae ]] && print -nR $nyae > done > [[ -n $nyae ]] && { print; zle redisplay } > } > zle -N zle-line-finish > > Use at your own risk. -- Eric Smith