Re: smart-insert-last-word bug in zsh 5.0.7

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Jun 12, 11:16am, Vincent Lefevre wrote:
} Subject: smart-insert-last-word bug in zsh 5.0.7
}
} ypig% autoload -U smart-insert-last-word
} ypig% zle -N insert-last-word smart-insert-last-word
} ypig% echo abc def
} abc def
} 
}   echo [Esc].[Left][Backspace][Ctrl-e] [Esc].
} 
} The second [Esc]. should have given:
} 
}   echo df def[ ]
} 
} like without smart-insert-last-word.

So the problem here is that smart-insert-last-word is using the cursor
position as a clue for whether it should start the search over, or look
farther up the history.  This is still true in 5.0.8.

When you do

  [Esc].[Left][Backspace][Ctrl-e][Space]

you change the line, but then return the cursor to its previous offset.
smart-insert-last-word thinks this means it should pick up where it left
off, because neither the history number nor the cursor position have
changed.

Now that we have UNDO_CHANGE_NO, that's probably a better way to track
the state, but maybe someone can find a different problem with that?


diff --git a/Functions/Zle/smart-insert-last-word b/Functions/Zle/smart-insert-last-word
index 27b0849..5288c1d 100644
--- a/Functions/Zle/smart-insert-last-word
+++ b/Functions/Zle/smart-insert-last-word
@@ -47,13 +47,13 @@ setopt extendedglob nohistignoredups
 zle auto-suffix-retain
 
 # Not strictly necessary:
-# (($+_ilw_hist)) || integer -g _ilw_hist _ilw_count _ilw_cursor _ilw_lcursor
+# (($+_ilw_hist)) || integer -g _ilw_hist _ilw_count _ilw_changeno _ilw_lcursor
 
 integer cursor=$CURSOR lcursor=$CURSOR
 local lastcmd pattern numeric=$NUMERIC
 
 # Save state for repeated calls
-if (( HISTNO == _ilw_hist && cursor == _ilw_cursor )); then
+if (( HISTNO == _ilw_hist && UNDO_CHANGE_NO == _ilw_changeno )); then
     NUMERIC=$[_ilw_count+1]
     lcursor=$_ilw_lcursor
 else
@@ -116,4 +116,4 @@ fi
 (( NUMERIC > $#lastcmd )) && return 1
 
 LBUFFER[lcursor+1,cursor+1]=$lastcmd[-NUMERIC]
-_ilw_cursor=$CURSOR
+_ilw_changeno=$UNDO_CHANGE_NO