Upload page content

You can upload content for the page named below. If you change the page name, you can also upload content for another page. If the page name is empty, we derive the page name from the file name.

File to load page content from
Page name
Comment

    PaulHowarth/Blog/2011-06-18

Saturday 18th June 2011

Fedora Project

  • Updated libxml in F-13, F-14, F-15 and Rawhide to fix the broken XPath implementation, which was failing the regression tests and, more significantly, segfaulting on 32-bit builds; this was a result of a pointer being cast from type xmlDocPtr to xmlNodePtr (which have only a few structure members in common) and trying to access non-existent elements of the resulting xmlNode (upstream said that the cast was generally safe to do but that was because there were lots of special-case tests in the code to see if the xmlNodePtr was really a xmlDocPtr and act accordingly, but one had been missed, resulting in the segfault and failed XPath tests)

  • This simple addition to the code fixed it:
  • --- libxml-1.8.17/xpath.c.orig  2011-06-18 16:26:57.430229932 +0100
    +++ libxml-1.8.17/xpath.c       2011-06-18 19:27:03.295989406 +0100
    @@ -1537,6 +1539,8 @@
             return(ctxt->context->node->childs);
         }
     
    +    if (cur == (xmlNodePtr) ctxt->context->doc)
    +        return(ctxt->context->doc->root);
         if (cur->childs != NULL) return(cur->childs);
         if (cur->next != NULL) return(cur->next);

Local Packages

  • Updated libxml as per the Fedora version

  • Updated perl-XML-LibXML to 1.73:

    • Calculating $err->column() properly, so it won't be maxed out at 80 (CPAN RT#66642)

    • Update the repository in the documentation to point to the bitbucket.org one
    • Revamped Makefile.PL:

      • Got rid of "\t" characters

      • Add "use strict" and "use warnings"

      • Add resources and keywords to the META_MERGE

      • Other changes
    • Fix CPAN RT#53632: when calling normalize on a node, processing of children nodes will stop when an empty element node is encountered

    • Apply the patch from Daniel Fetter's InputCallbackFix branch:

      • A partial fix to CPAN RT#4263

      • Call two $parser->parse_string() in succession

    • Apply the NestedParsing patch (more of CPAN RT#4263):

      • Updated how legacy parser local callbacks are utilized by init_callbacks so that the XML::LibXML::InputCallback object doesn't have to be temporarily modified during the parsing process; this change could break code for users that have subclassed XML::LibXML::InputCallback and overridden the init_callbacks method

    • Documentation fixes patch from Daniel Fetter


Recent