Re: [Nano-devel] Re: nano 1.3.5 - cut & paste

[David Lawrence Ramsey]

--- Rocco Corsi <address@hidden> wrote:
>On Fri, 2004-11-26 at 21:09 +0100, Jean-Philippe Gu?rard wrote:
>> I've moved to nano 1.3.5 and I encountered the following issue: when 
>> I cut and paste text from another X11 application or console, the 
>> pasted text in nano stops after the first non-ASCII character.
>> 
>> For example, if I cut: "Mon texte en ASCII ?tendu", when pasting, I 
>> will get: "Mon texte en ASCII ?".
>
>I've reproduced the problem on Fedora Core 3.  Both with Gnome Terminal 
>and Xterm.
>
>For me the accented e is not displayed.

I've looked into it, and the problem is apparently caused by my improper 
handling of sequences that contain more than one valid non-ASCII 
character.  I would normally have a patch for this against 1.3.5, but 
the minimum set of changes needed to fix this cause nano to lock up when 
pasting some text containing valid non-ASCII characters, and I'm stumped 
as to why.  I'm almost done rewriting parts of the input routines to 
handle things properly.  Apparently it's not a *real* text editor unless 
you have to rewrite the input routines every three releases :)

It'll take some more time to complete this, though.  In the meantime, 
the non-working non-ASCII parsing code has been ripped out of CVS, so 
you can try pulling current CVS and using that in the meantime.  Sorry 
for the inconvenience.

If you go ahead with current CVS, there are a few other major changes 
you should know about:

* the cutting/uncutting routines have been rewritten to use the 
partitioning code, so the code is much simpler; note that nano no longer 
duplicates Pico's buggy behavior of adding a magicline unconditionally 
if uncutting leaves the cursor on the original magicline; it now only 
adds a magicline when it's actually needed

* the routines to back up and restore unjustified paragraphs have been 
rewritten to use the partitioning code, so resizing/suspending after a 
justify works properly now

* DB's overhaul of the multibuffer code to make it use a doubly-linked 
list for efficiency and otherwise clean it up has been added