Re: [Lynx-dev] problems using google

[Ian Collier]

On Fri, Nov 15, 2019 at 08:21:23AM +0100, Dick Sterkenburg wrote:
> what is the meanig of the ctrl-v option, why does it bring the links back?

I do not know the full details about what this does, but the help in
lynx.cfg for this option states:

 If TAGSOUP is set, Lynx uses the "Tag Soup DTD" rather than "SortaSGML".
 The two approaches differ by the style of error detection and recovery.
 Tag Soup DTD allows for improperly nested tags; SortaSGML is stricter.

The issue is that Lynx thinks the tags in Google search results are
improperly nested, and therefore from the above description "Tag Soup"
is going to display them better than the strict mode.

In fact, they are not improperly nested according to current standards.
The issue is that the <A> tag which links the results contains two <DIV>
tags containing, respectively, the title of the link and its URL.

In HTML prior to 5, <A> elements were not allowed to contain block
elements such as <DIV>, and Lynx was written to this standard.
However, in HTML 5, the content model of the <A> tag is "transparent"
which means it is allowed to contain whatever the parent element may
contain.  In most cases, this means that <A> tags are now allowed
to contain <DIV> tags.

So this is something that the Lynx developer could consider fixing,
though I'm not sure how involved that would be.

imc