Re: [Lynx-dev] lynx2.8.7dev.12

[Thomas Dickey]

On Sun, Jan 18, 2009 at 08:04:34PM +0000, Thorsten Glaser wrote:
> Thomas Dickey dixit:
> 
> >2008-12-31 (2.8.7dev.12)
> 
> Updated.
> 
> >* fix an off-by-one difference between UTF-8/non-UTF-8 comparisons in
> >  HText_appendCharacter() which would leave an extra character wrapped in 
> > some
> >  cases, e.g., in the options menu when field-numbering was turned on -TD
> 
> Looks like this (80x25 uxterm) now:

...yes (UTF-8/non-UTF-8 display the same now)

I think you're referring to this line though:

>   Keypad mode                      : [11][(3)__Links and form fields are 
> numbere
> d]

This seems to be expected behavior, notwithstanding my fix
(the strings aren't truncated, so some will wrap anyway).

>   Emacs keys                       : [12][(1)__OFF]
>   VI keys                          : [13][(1)__OFF]

By the way, a few lines further would show _this_ odd feature:

  Keyboard layout                  : [15][(1)__ROT13'd keyboard layout__________
..]                                                                             
^^ (the code really does this intentionally, when the string
    does not extend as far as the field definition - some of the
    other strings in this selecton do use the full width of the
    field)

Text fields in the options menu differ from the selection-fields.
The latter have a fixed width which depends on their contents,
and can wrap, since turning on numbering changes the nice estimate
that formatted them into ~40 columns.

> Submit to LYNXOPTIONS:/
> 
> >* modify havevisible() and UCTransToUni() functions to process UTF-8 input,
> >  fixing a problem saving bookmarks whose title is UTF-8 (prompted by patch 
> > by
> >  Takao Fujiwara) -TD
> 
> Bookmark files in UTF-8 are now broken:
> 
>      * [16]Putty (SSH f�r Win32, verwende ich selbst)                         
>                                    
>      * [43]smallcode � Blog Archive � Code page for Win32 console programs 
> (and more stuff: Self-extracting      
>      * [492]Center for Information Technology Policy � Lest We Remember: Cold 
> Boot Attacks on Encryption         

This part is a bug report - no, I didn't notice this.
 
> Furthermore, I had to add YACC=yacc to the configuration environment
> to prevent it from picking up GNU Bison, which is ⓐ not suitable when
> compiling for the base system, and ⓑ broken:
> 
> y.tab.c:322:6: "YYENABLE_NLS" is not defined
> y.tab.c:762:6: "YYLTYPE_IS_TRIVIAL" is not defined

that's expected behavior (a bias on the part of the autoconf
developers from long ago).  But if your file timestamps are
set to reflect the original timestamps, then "make" shouldn't
have anything to build.
 
> This is for parsdate.y.

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net