Re: Emacs aborts when trying to use replace-string with "bad" character

[Tim Van Holder]

On Tue, 15 Feb 2005 01:18:39 -0500, Richard Stallman <address@hidden> wrote:
> Does it work well with this change?
> 
> *** minibuf.c   12 Dec 2004 13:24:44 -0500      1.278
> --- minibuf.c   14 Feb 2005 10:22:51 -0500
> ***************
> ***************
> *** 598,603 ****
> --- 602,611 ----
>     minibuffer = get_minibuffer (minibuf_level);
>     Fset_buffer (minibuffer);
> 
> +   /* If appropriate, copy enable-multibyte-characters into the minibuffer.  
> */
> +   if (inherit_input_method)
> +     current_buffer->enable_multibyte_characters = enable_multibyte;
> +
>     /* The current buffer's default directory is usually the right thing
>        for our minibuffer here.  However, if you're typing a command at
>        a minibuffer-only frame when minibuf_level is zero, then buf IS
> ***************
> *** 669,678 ****
> 
>     minibuf_prompt_width = (int) current_column (); /* iftc */
> 
> -   /* If appropriate, copy enable-multibyte-characters into the minibuffer.  
> */
> -   if (inherit_input_method)
> -     current_buffer->enable_multibyte_characters = enable_multibyte;
> -
>     /* Put in the initial input.  */
>     if (!NILP (initial))
>       {
> --- 677,682 ----
> 

Yes, CVS emacs of 10 minutes ago, with this change applied, does not
abort anymore.
Also, apologies for the direct reply - seems GMail doesn't have a "reply all".