patch applied (packages/HaXml): In Xtract patterns, downcase HTML only for tags+attrnames, not attrvals.

Malcolm Wallace Malcolm.Wallace at cs.york.ac.uk
Wed Jan 31 13:26:03 EST 2007


Wed Jan 31 10:15:04 PST 2007  Malcolm.Wallace at cs.york.ac.uk
  * In Xtract patterns, downcase HTML only for tags+attrnames, not attrvals.
  Although XML is case-sensitive, HTML is case-insensitive regarding
  tagnames and attribute names.  But of course the attribute *values* are
  still case sensitive, so a blanket (map toLower) on pattern queries is
  wrong.  Here is an example of a document and query which previously
  failed:
  
     <a class="Foo">
  
     Xtract "//a[@class='Foo']"
  
  The solution is to fix the lexer to do downcasing (or more generally, any
  transformation) on the appropriate tokens.

    M ./src/Text/XML/HaXml/Xtract/Lex.hs -35 +39
    M ./src/Text/XML/HaXml/Xtract/Parse.hs -6 +6
    M ./src/tools/Xtract.hs -2 +2
    M ./src/tools/XtractLazy.hs -2 +2


More information about the Cvs-libraries mailing list