[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev Multibite characters for Lynx 2.8.3.rel1

From: Klaus Weide
Subject: Re: lynx-dev Multibite characters for Lynx 2.8.3.rel1
Date: Thu, 18 May 2000 13:15:08 -0500 (CDT)

On Thu, 18 May 2000, Takuya ASADA wrote:

> Hello, Lynx developers.
> I tried 2.8.3.rel1 with ncurses-5.0, and I found it can't display
> Japanese page correctly. I've discussed with Mr. Hataguchi, and
> we found a "clearok(curscr, TRUE)" was commented out in GridText.c
> at line 2289. And it was commented:
>          *  Is this repainting necessary??  Let's try without.

That was me...  I understood why repainting was necessary for UTF-8,
but did not understand it for CJK character sets.

Sorry.  I did a bit of testing then, but didn't find any obvious
disadvantages.  (I don't undersstand Japanese at all,so I just went by
graphical appearance of some test files.)

Some people had reported problems with the full refresh present, and they
seemed to go away when the refresh was removed.  But that was when we had
'lynx_force_repaint();' instead of 'clearok(curscr, TRUE);'.

This change was in the development code since 1999-10-21 (2.8.3dev.13).  
I am rather surprised that nobody reported this as a problem in all the
time since then.

> You know, curses libraly tries to make tha cost of redrawing screen
> lesser, then, it replaces only the 1st or 2nd bite of multibite 
> characters (such as Japanese). This is a reason for doing reflesh
> here. Of cource, this is not best solution, but it works well.

Yes, I understand now why it it necessary.

I would still like to know whether it is also necessary if lynx is
 - with a regualar Slang lib?
 - with a Slang lib compiled with SLANG_HAS_KANJI_SUPPORT?

> So, please remove comment out in next release.

--- /tmp/lynx2-8-3/src/GridText.c       Thu May 18 13:00:07 2000
+++ /tmp/lynx2-8-3/src/   Thu May 18 13:00:56 2000
@@ -2284,9 +2284,11 @@
         *  For non-multibyte curses.
-        *  Is this repainting necessary??  Let's try without.
+        *  Full repainting is necessary, otherwise only part of a multibyte
+        *  character sequence might be written because of curses output
+        *  optimizations.
-       /*clearok(curscr, TRUE);*/
+       clearok(curscr, TRUE);

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]