[Bug 4891] Strange error for number enclosed in colons followed by XHTML text entity

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4891

           Summary: Strange error for number enclosed in colons followed by
                    XHTML text entity
           Product: Validator
           Version: 0.8.0
          Platform: All
               URL: http://www.fourmilab.ch/entrenous/validation_2007-07-
                    29.html
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: check
        AssignedTo: dave.null@w3.org
        ReportedBy: kelvin@fourmilab.ch
         QAContact: www-validator-cvs@w3.org


Validating the cited URL with the 0.8.0 online Markup Validator results in the
following error message:

    Line 12345, Column 0: — .

Not only is the XHTML valid (as far as I can see), but this is only a 16 line
document!  The key line is the following body text (line 12) within a <p>
container:

    :12345:&mdash;

The line number reported is the number within the colons--you can change it and
line number in the error report will change to the new value.  I have made the
following observations based on experiments:

   1.  Changing the text within the colons to an alphabetic
       sequence eliminates the error message.
   2.  Changing the text entity from &mdash; to a numeric
       entity (for example &#8810) or an entity in the
       US-ASCII range (for example &quot;) eliminates the
       error message, but changing it to a character in the
       ISO-8859 0xA0-0xFF range (for example &eacute;) produces
       the error message.
   3.  Adding arbitrary alphanumeric text between the second
       colon and the text entity does not affect the error
       message, nor does text including markup (for example
       "<b>text</b>").
   4.  The error message is produced for the XHTML 1.0 Transitional
       DTD as well as the Strict.

Received on Sunday, 29 July 2007 13:44:55 UTC