Re: lynx-dev changes for Japanese (was: Re: reading SJIS docs)

[Klaus Weide]

On Thu, 13 Jan 2000, Hataguchi Takeshi wrote:

> On Wed, 12 Jan 2000, Henry Nelson wrote:
> 
> > > (1) If Japanese document's charset is specified explicitely
> > >     by the MIME tag or HTTP responses, Lynx will assume the charset 
> > >     as it.
> > > (2) Change the Japanese charset detection strategy when chaset 
> > >     isn't specified explicitely.
> > > 
> > > Please test these changes with CJK_EX and if necessary, with

Why should one *not* test these changes *without* CJK_EX?
Lynx should strive to provide good CJK support without "EXtensions"
being necessary.

For example, I am willing to test some CJK things - as far as that is
possible, without a real CJK environment (although I am even
installing some Japanese fonts and terminal emulator now) - but I am
not really willing to define the diffuse & amorphous (and undocumented
in English, except that it provides a "CJK EXtension") CJK_EX symbol.
I shouldn't have to.  I should get CJK support at least as good as
(say) in lynx 2.7.1 without it, and hopefully your recent improvements
will also benefit in this situation.

> > > USE_TH_JP_AUTO_DETECT.

> > (Kind of hate to have the "USE_TH_JP_AUTO_DETECT" define, though.
> > Couldn't it eventually become the default after a short test period?)
> 
> I thought almost all Japanese like (1)'s behavior but someone might not
> like (2)'s behavior. So I introduced the macro "USE_TH_JP_AUTO_DETECT".
> But I've tested (2)'s changes these days and found no probrem.
> So I think it's better to replace "USE_TH_JP_AUTO_DETECT" to "CJK_EX"

Please no, let's not pack even more implications into that "CJK_EX" -
which ideally should not be there at-all.

> (or remove it) now.

Removing it looks more reasonable.  Or replace it with a run-time
option or flag if that makes sense.  But I think that autodetection
may interact with the ASSUME_CHARSET stuff, once it has been sorted
out how ASSUME_CHARSET and "raw" *should* interact with CJK display
character sets properly, so it may be premature to remove the
USE_TH_JP_AUTO_DETECT now.

> If my patch is adopted, can I ask this renaming to the coordinator?
> Thanks in advance.

I suggest to keep it, for now, and have it defined by default.

   Klaus