Re: lynx-dev strange interaction with gettext/anonymous/gz

[Nelson Henry Eric]

> Well, I haven't even looked at the dev.n code yet; but since there aren't
> many alternatives AFAIK, and gettext seems to be the GNU way, in principle
> I'm all for it.  I just prefer to look at other things now, and leave the

Echos my sentiments.

> If the problem was only that it was too restrictive with respect to length,
> that could be easily changed by bumping up that number, but I somehow
> doubt that that was the only problem.

I tried moving the number to twice what you had, but it still wasn't
enough to catch all the problem SUBJECTS.  In truth, the only way I
could get it to work was to block all your checks on the strings and
force the whole thing to a "valid" (or was that invalid?) string.

It seemed to me that what was happening was that the string was getting
truncated somewhere, with the result that occasionally a multibyte
character would get split in a place that would result in a non-ascii
character that messed up what came after it.  At the time, I thought
it was because the function seemed to analyze the =?iso-nnnn part only
one time, whereas in fact it can occur multiple times.  It also seemed
like it was trying to cut out the =?iso-nnnn part, when it could just
stay there.  But really my memory has faded a lot.

> Are you asking for the other HTmmdecode() because it also treated 
> non-Japanese strings, or for some other reason?

Just thought it was a very noble idea for Lynx to try to handle
any language, not just CJK.  I'm sure you've got your own priorities,
so fiddle with it only if you want.  I just wanted to let you know
that I would continue to check that it was working or not with CJK.
(My level of understanding has not increased from a year ago, so
take that tongue-in-cheek.)

> How does the current Lynx show that Subject line?

Fine (Lynx also did the posting):

                        fj.test, Articles 105995-106024
                                       
   ([1]Earlier articles...)
   
   Articles in fj.test
     * [3]"$B%F%9%H$G$9!#(J" - OwlTKHS
     * [4]"$B;n83Cf(J" - OwlTKHS
     * [5]"Re: test test" - "LOVE CAT"
[...]
       ignore @@NCM" - Usenet exProwler
     * [12]"$B%F%9%H$G$"$j$^$9$k(J" - "skintone"   <== this one's funny -HN
     * [13]"none" - Anonymous
[...]
     * [29]"EFNET NoCeM Report ncmreport.fj+japan+tnn-sbi.19981117-0415.1
       ignore @@NCM" - Usenet exProwler
     * [30]"$B<+%5!<%P7PM3$GB>address@hidden<%k<u?.(J" -  <== this is it
       address@hidden
     _________________________________________________________________
   
__Henry