Emmanuel Bourg dixit: >I stumbled on the same issue when playing with Maven pom files and it >seems to be a feature of xmlstarlet: > > http://xmlstar.sourceforge.net/doc/UG/ch05s01.html > >If the document has a default namespace it has to be specified >explicitly with the -N option and then prefixed to the name of the >elements in the XPath expression:
Interesting, but… I don’t *know* ahead of the time if this is true, if the files I work on are, say, HTML pages I scraped from some remote website and, if necessary, ran through either tidy or html5tidy to make something resembling XML out of them for the sole procedure of using xmlstarlet sel. >xmlstarlet sel -N xhtml=http://www.w3.org/1999/xhtml -T -t -c >'//xhtml:h3[@id="baz"]' -n test2.xml > >This is rather annoying and the default namespace should probably be >ignored, but at least there is a workaround. I “just” want to get the values of some tags (and I do not wish to “parse” XML using regex, as that is Just Not Done™), I don’t care about namespaces and whatnot, so I consider this an anti‐ feature. But thanks for the info! bye, //mirabilos -- Thorsten Glaser (2. Vorstandsvorsitzender) Teckids e.V. – Erkunden, Entdecken, Erfinden. https://www.teckids.org/
