Re: lynx-dev More on the lynx segmenation fault I was getting

[Leonid Pauzner]

>> The problem is that before this release, lynx accepted character names of
>> ISO Latin 1.  It not only doesn't, but crashes on Solaris 2.5 and 6 when
>> it finds that info in one's lynxrc or lynx.cfg file - which of course
>> it will find when one tries to test or use it with the previous defaults.
> ok - LP had a comment about that (but I think we should handle it more
> gracefully)

I thought that incorrect value will fall back to iso-8859-1,
but it was not happen:-(
Plan to revise the code before dev26 (probably).
>From the other hand, it is easy to add a special check
for old charset names while reading .lynxrc/lynx.cfg,
but the new .lynxrc will be incompatible with earlier version anyway.
I was thinking about introducing a new variable name,
like "display_character_set" instead of present "character_set",
is it better?


>> I started looking at fixing things, but the problem is pretty pervasive in
>> the LYCharSet.c file, where most of the functions think that the character
>> set is going to be a valid number, while the code that looked up the char
>> sets returns a negative one.

thanks for report.