From a2aa1a156c29658b62a691f8aa343985043b8c9c Mon Sep 17 00:00:00 2001
From: Junegunn Choi <junegunn.c@gmail.com>
Date: Mon, 27 Jan 2025 15:40:21 +0900
Subject: Allow {q} placeholders with range expressions

e.g. {q:1}, {q:2..}
---
 ADVANCED.md | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

(limited to 'ADVANCED.md')

diff --git a/ADVANCED.md b/ADVANCED.md
index a045140c..71ee15ae 100644
--- a/ADVANCED.md
+++ b/ADVANCED.md
@@ -517,18 +517,15 @@ remainder of the query is passed to fzf for secondary filtering.
 
 INITIAL_QUERY="${*:-}"
 TRANSFORMER='
-  words=($FZF_QUERY)
-
-  # If $FZF_QUERY contains multiple words, drop the first word,
-  # and trigger fzf search with the rest
-  if [[ ${#words[@]} -gt 1 ]]; then
-    echo "search:${FZF_QUERY#* }"
-
-  # Otherwise, if the query does not end with a space,
-  # restart ripgrep and reload the list
-  elif ! [[ $FZF_QUERY =~ \ $ ]]; then
-    pat=${words[0]}
-    echo "reload:sleep 0.1; rg --column --line-number --no-heading --color=always --smart-case \"$pat\" || true"
+  rg_pat={q:1}      # The first word is passed to ripgrep
+  fzf_pat={q:2..}   # The rest are passed to fzf
+  rg_pat_org={q:s1} # The first word with trailing whitespaces preserved.
+                    # We use this to avoid unnecessary reloading of ripgrep.
+
+  if [[ -n $fzf_pat ]]; then
+    echo "search:$fzf_pat"
+  elif ! [[ $rg_pat_org =~ \ $ ]]; then
+    printf "reload:sleep 0.1; rg --column --line-number --no-heading --color=always --smart-case %q || true" "$rg_pat"
   else
     echo search:
   fi
-- 
cgit v1.2.3