Re: lynx-dev reverse search

[Vlad Harchev]

On Fri, 28 Jul 2000, David Combs wrote:

> On Fri, Jul 28, 2000 at 11:43:57AM +0500, Vlad Harchev wrote:
> > On Thu, 27 Jul 2000, David Combs wrote:
> > 
> > > On Thu, Jul 27, 2000 at 08:42:51AM -0400, Larry W. Virden wrote:
> > > > If someone goes into play with reverse search, how about considering 
> > > > adding
> > > > regular expressions at the same time?  The ability to search for
> > > 
> > > Doesn't gnu have a regexp library?  Maybe more than one?
> > 
> >   Yes, they have one (used by emacs). But regexp lib should support unicode 
> > to
> > be used in lynx. Only few support unicode - latest Spencer's one and the
> > slightly unfinished by Mark Leisher (and also in new perl, but it's hard to
> > extract the exact rx stuff).
> >  
> 
> Does lynx support unicode now?
> 
> The conversion from ascii to unicode must be HORRIBLE,
> screwing up everything, all those unix string-length calls,
> arrays of chars, "is this a space" -- effecting just
> about EVERY line of code.
> 
> Maybe it's been done already; but I sure couldn't find a more
> awful task for someone to have (except converting apl into
> c, maybe :-).
> 
> Who wants to write yet-another-regex program?
> 
> Seems to me that to add in an ASCII-ONLY regex-feature,
> accessible via some multiple-key combination (to turn
> "ascii-regex mode", and turn it off, too).
> 
> Probably 99% of the users use ascii (does anyone have
> ANY statistics on usage?) -- if so, the easy way out is to 
> just hook-in some existing regex-library.
> 
> ---
> 
> Not being a developer (of lynx), I don't know what
> lynx has in unicode already -- so you tell me.
> 
> ---

  Yes, lynx supports unicode. If display charset is set to utf8, then lynx
will store everything in utf8 (and will display in it) - lynx always stores
its internal representation of the document in the display charset set.
  Since utf8 is a variable-length encoding for unicode, and provided that pure
ascii string will be the same ascii string in utf8 (i.e. the same bytes
exactly) nothing screws up, and yes, it's possible to use non-unicode-aware rx
library for utf8 text (if it's ascii was encoded in utf8, then results of
the search will be correct, if it's something other than ascii in the string -
result will be wrong but nothing will crash).

>[...] 
> David
> 
> ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden
> 

 Best regards,
  -Vlad


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