18 Mar 2014 18 Mar '14
Herm Fischer, 17.03.2014 19:58:
For XBRL usage we routinely have humongous files well over millions of lines, and I think we need the XML_PARSE_BIG_LINES libxml parser option so we don't get line numbers crunched to short ints.
Other than recompiling from source, is there a way to expose this option (or just use it for everybody)? (End users don't like the thought of custom-hacked lxml.)
I implemented it in latest master. It's automatically switched on when using libxml2 2.9.0 or later.
Note that it only works for the XML parser, not the HTML parser (AFAICT).
Please give it a try.