New failure in httpd units tests (xml related)

Rainer Jung <[email protected]> Tue, 6 May 2025 15:09:05 +0200
Newsgroups	gmane.comp.apache.devel
Message-ID	<[email protected]>
Hi there,

I observed a new failure in httpd units tests, namely in t/apache/pr64339.t:

# Failed test 12 in t/apache/pr64339.t at line 45 fail #4

# testing : content test for /doc.isohtml
# expected: '<html><body><p>fÃ³Ã³
# </p></body></html>'
# received: '<html><body>fÃ³Ã³
# </body></html>'
not ok 12

Note, that "<p>" and "</p>" are missing from the result. The test class 
contains the following test definition:

['/doc.isohtml', "text/html;charset=utf-8", 
"<html><body><p>fóó\n</p></body></html>" ],

The request gets send to "/modules/xml2enc/front" and the requested file 
only contains the three UTF characters similar to "foo". The conversion 
to HTML seems to be done by mod_xml2enc and/or mod_proxy_html. The test 
config contains:

#
# t/apache/pr64339.t test configuration
#
<IfModule mod_xml2enc.c>
    <IfModule mod_proxy_html.c>
       <IfModule mod_proxy.c>
          Alias /modules/xml2enc/back @SERVERROOT@/htdocs/modules/xml2enc
          <Location /modules/xml2enc/back>
              AddType application/foo+xml fooxml
              AddType application/notreallyxml notxml
              AddType application/xml xml
              AddType text/html isohtml
              AddCharset ISO-8859-1 .isohtml
              AddCharset UTF-8 .xml
              AddCharset UTF-8 .fooxml
          </Location>
          <Location /modules/xml2enc/front>
              ProxyHTMLEnable on
              # mod_proxy_html needs some configuration.
              ProxyHTMLURLMap / /blah
              ProxyHTMLLinks a href
              ProxyPass http://@SERVERNAME@:@PORT@/modules/xml2enc/back
          </Location>
       </IfModule>
    </IfModule>
</IfModule>


The failure happens for 2.4.63 builds, although I didn't notice it 
during the release vote. It is due to my update of libxml2 from 2.13.5 
to 2.14.2. I can reproduce the change with the accompanying xmllint:


.../2.14.2/xmllint --html doc.isohtml

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" 
"http://www.w3.org/TR/REC-html40/loose.dtd">
<html><body>f&oacute;&oacute;
</body></html>

.../2.13.5/xmllint --html doc.isohtml

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" 
"http://www.w3.org/TR/REC-html40/loose.dtd">
<html><body><p>f&oacute;&oacute;
</p></body></html>

We might need to make the test more tolerant wrt. the generated 
paragraph tags or use a different type of input file.

Best regards,

Rainer