Christian Brabandt wrote:

> I found another problem that was introduced with 7.3.251.
> 
> vim -u NONE -U NONE -N -i NONE -c "call setline('.', 
> ['','foobar',''])|$"
> 
> Now the cursor is on the last line. Execute this command:
> dv?bar
> 
> First, notice, that the buffer now looks like this:
> 
> -----
> foo
> -----
> 
> e.g. the first line was wrongly deleted. Now if you undo the previous 
> command (press u), you see this:
> 
> ----
> foo
> foobar
> 
> ----
> 
> So the undo did not work correctly.
> 
> I did a hg bisect to find the patch, that introduced this regression and 
> according to my testings, it was introduced with patch 7.3.251:
> 
> ,----
> | Änderung:        2957:fe6ad3fd8532
> | Marke:           v7-3-251
> | Nutzer:          Bram Moolenaar <[email protected]>
> | Datum:           Fri Jul 15 17:51:34 2011 +0200
> | Dateien:         src/normal.c src/ops.c src/version.c
> | Beschreibung:
> | updated for version 7.3.251
> | Problem:    "gH<Del>" deletes the current line, except when it's the last
> |             line.
> | Solution:   Set the "include" flag to indicate the last line is to be
> |             deleted.
> `----
> 
> I am not sure if there are actually two problems (wrong undo and 
> wrongly joined lines) or if this is just the same problem. However, the 
> following patch fixes the wrongly joining lines for me and the wrong 
> undo does not occur anymore:
> 
> diff --git a/src/ops.c b/src/ops.c
> --- a/src/ops.c
> +++ b/src/ops.c
> @@ -1961,7 +1961,8 @@
>                 /* Special case: gH<Del> deletes the last line. */
>                 del_lines(1L, FALSE);
>                 curwin->w_cursor = curpos;      /* restore curwin->w_cursor */
> -               if (curwin->w_cursor.lnum > 1)
> +               if (curwin->w_cursor.lnum > 1 &&
> +                       curwin->w_cursor.lnum > curbuf->b_ml.ml_line_count)
>                     --curwin->w_cursor.lnum;
>             }
>             else

Thanks for finding out what is wrong and the suggested patch.
I'll check later if this is the best solution.

And to prevent further regressions a test would be nice.

-- 
Veni, Vidi, VW -- I came, I saw, I drove around in a little car.

 /// Bram Moolenaar -- [email protected] -- http://www.Moolenaar.net   \\\
///        sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org        ///
 \\\            help me help AIDS victims -- http://ICCF-Holland.org    ///

-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

Raspunde prin e-mail lui