Re: LYNX-DEV 2.8 bug in handling > ?

[T.E.Dickey]

> > something like that - the error recovery is different between 2.7.2 (Fote)
> > and 2.8pre (Klaus).  Klaus' approach insists that things be nested
> > properly, otherwise the whole stack gets popped (an oversimplification, but
> > the effect is that small errors are very noticable).  Fote's approach tries
> 
> But isn't Fote's approach available in 2.8pre, too?  I thought BOTH
> parsers functioned in the development tree.  If that is not the case,
> then the "keystroke_help.html" file in lynx_help needs to be changed:
>       ^V          SWITCH_DTD    switch between two ways of parsing HTML.
> 
> When I toggle with ^V, I get a message about "tag soup" vs "sgml".  I
> hope more than the message is working!
That's something different: I don't know all of the details, but this
one adds qualifying information to the tags, and they behave differently.

The case that I know better has to do with the logic used to unstack tags.

In post-2.8, I intend to study that part first, since it seems to be the
most trouble-prone (I have some ideas for how it should be done differently,
and, as I said, perhaps Fote's code is a much-reduced form of that).

The SWITCH_DTD code I'll have to study also (initially, though, I'm
concerned that the program doesn't work properly when I turn it off
altogether - which is a little different from using the ^V toggle - or at
least the example that I ran didn't).  There's a portability problem with
the table entries anyway (I'll fix that, and thereby gain a little more
insight into the code).

> __Henry
> 


-- 
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey