bug in text directive

[Anniepoo]

This is a bit of wikipedia scraper code. It grabs the country name as the innerhtml of the a tag.

:- module(datascoop, [test/1]).


:- use_module(library(http/http_open)).
:- use_module(library(sgml)).
:- use_module(library(xpath)).
:- use_module(library(sgml_write)).

test_url('https://en.wikipedia.org/wiki/Visa_requirements_for_Yemeni_citizens').

test(row(
         WIKIREF,
         Country

     )) :-
    test_url(URL),
    load_html(URL, DOM, []),
    !,
    xpath(DOM, //div(h2(index(I), span(@id(string)="Visa_requirements_map"))), Div),
    succ(I, TI),
    xpath(Div, table(@class='sortable wikitable', index(TI)), Table),
    findall(Head, xpath(Table, /self(tbody/tr/th(text(atom)=Head)), _),
           [ 'Country',
             'Visa requirement',
             'Allowed stay',
             'Notes (excluding departure fees)'
           ]),
    !,
    xpath(Table, tbody/tr, Row),
    xpath(Row, td(index(1))/a(@href), WIKIREF),
% if you change this to text from normalize_space it binds Country to "\240\Afghanistan"
% changing to normalize_space shows "Afghanistan".
    xpath(Row, td(index(1), normalize_space), Country).

[JanWielemaker]

Why is this a bug? \240\ is a non-breaking space that you should get if you ask for text and not if you normalize spaces (remove from both ends, use a single space for each internal sequence).