file(1) no longer recognises XML content

[Gavin Panella]

Binary package hint: file

1) The release of Ubuntu you are using, via 'lsb_release -rd' or
   System -> About Ubuntu.

Intrepid

2) The version of the package you are using, via 'apt-cache policy
   packagename' or by checking in Synaptic.

4.24-4

3) What you expected to happen

That `file --mime-type some.xml` (where some.xml is any file
containing valid XML) should return either text/xml, application/xml,
or something more specific (like application/atom+xml for example).

4) What happened instead

More often than not, just text/plain is returned. Sometimes, like for
/usr/share/mime/text/x-haskell.xml, text/x-pascal is returned, or for
/usr/share/xml/docbook/schema/dtd/catalog.xml, text/html is reported,
both of which are wrong.

I ran:

  locate *.xml | xargs file --brief --mime-type |
    egrep '(application|text)/xml'

and got no output at all. The `locate` command found about 28000
files.

[Brian Murray (brian-murray)]

I've confirmed this using file version 4.24-4 on Intrepid.

[Brian Murray (brian-murray)]

Additionally using file version 4.21-3ubuntu1 on Hardy I saw this:

bdmurray@bizarro:~$ file -i /usr/share/mediatomb/web/cm.xml
/usr/share/mediatomb/web/cm.xml: text/xml

While on Intrepid:

13:04:03 - flash:[~] file -i /usr/share/mediatomb/web/cm.xml
/usr/share/mediatomb/web/cm.xml: text/plain charset=us-ascii

So this seems to be a regression.

[Gavin Panella (allenap)]

This is fixed in Debian, package version 4.26-1.

[Colin Watson (cjwatson)]

Confirmed fixed. The diff from 4.24-4 to 4.26-1 includes this chunk:

@@ -25,14 +31,23 @@
 # Extensible markup language (XML), a subset of SGML
 # from Marc Prud'hommeaux (<email address hidden>)
 0 search/1/cb \<?xml XML document text
+!:mime application/xml
 0 string \<?xml\ version\ " XML
+!:mime application/xml
 0 string \<?xml\ version=" XML
+!:mime application/xml
+>15 search/1 >\0 %.3s document text
+>>23 search/1 \<xsl:stylesheet (XSL stylesheet)
+>>24 search/1 \<xsl:stylesheet (XSL stylesheet)
 0 string \<?xml\ version=' XML
+!:mime application/xml
 >15 search/1 >\0 %.3s document text
 >>23 search/1 \<xsl:stylesheet (XSL stylesheet)
 >>24 search/1 \<xsl:stylesheet (XSL stylesheet)
 0 search/1/b \<?xml XML document text
+!:mime application/xml
 0 search/1/b \<?XML broken XML document text
+!:mime application/xml

 # SGML, mostly from rph@sq