[Lynx-dev] site newly returns 503 unavailable for lynx

[Karl Berry]

A few days ago, my local public library system (which is running koha)
started returning the http error "503 service unavailable" with lynx and
links to any search url, for example:
https://coastlinelibrarynetwork.org/cgi-bin/koha/opac-search.pl?advsearch=1&idx=ti%2Cphr&q=these+precious+days&op=and&idx=au%2Cwrdl&q=ann+patchett&do=Search

The normal result would be a page saying "No results found" (expected,
since this book hasn't been published yet). It's the same 503 error for
searches on existing items.

It works fine with graphical browsers (firefox, chrome, seamonkey), and
wget returns the page source, which is full of JavaScript. The site has
worked fine with lynx for many years.  The library home page
(https://coastlinelibrarynetwork.org) does come up with lynx, but not
the searches. 

I've tried different user-agent strings and assorted other options, but
nothing seems to make a difference. My guess is that something in the
JavaScript is aborting with the 503.

Any ideas on how to proceed? Is there a plausible way to get JavaScript
in lynx? I couldn't find anything that looked feasible to me on the
web. The librarians and/or their tech support (a company named Bywater)
would help me if they could, but they know nothing about lynx. I've
failed to find anything on the web.

This is under CentOS7.9 with the current version of lynx
(2.8.9rel.1 (08 Jul 2018)), which I compiled from the original source,
as well as the old distro version (2.8.8dev.15 (18 Nov 2012)).

I haven't tried a headless firefox/chrome/whatever yet.
I guess that will be my next effort.

Any help appreciated. --thanks, karl.