Raptor 1.4.9 with Tutorial
Yesterday I released Raptor 1.4.9 and the major visible addition this time is the first version of the new Raptor Tutorial covering all the parsing and serializing functions with working full examples. Used together with the updated Raptor Reference Manual which now covers 100% of the public API of functions, structures and defines, these provide a complete set of docs. (There is also the libraptor manual page but that’s primarily for command-line use in unix/linux).
There are of course, more fixes and improvements:
- The
rapperutility can now pretty-print RDF using namespaces from parsing as hints in serializing. - The Turtle parser has gained boolean literals which were accidently left out last time, oops!
- Requests for content to parse now send appropriate HTTP
Accept:headers depending on the parser used. - It is no longer required to use libxml2 for the rss-tag-soup (Atom, RSS*) parser
- Various Win32 fixes and VC build files updates from John Barstow
I previously described how Raptor was refactored in detail and the results of the latest changes are mostly internal with respect to the SAX2 API. The public result is that the Atom support inside the RSS Tag Soup parser is ready but not fully enabled to handle the enveloping of X-in-Atom that people are now trying. I guess this is better than using SOAP or (ugh) XML-RSS.
23 April 2006 at 4:28 pm
Novedades de Redland
Para todos aquellos interesados, ahora existe un tutorial de Raptor, el parser de le excelente librería Redland. Esto, sumado a la versión para Windows es una grata, grata noticia.
23 April 2006 at 7:11 pm
Did you really have to name those boolean literals TRUE and FALSE - and give them values other than 0 and 1? Can’t we call them TURTLE_TRUE and TURTLE_FALSE?
Other than that very happy to get rapper.exe working with the new pretty-print; unfortunately I think I need to wait for the up-to-date librdf before I can distribute on Win32 (using 1.0.3 with updated Raptor keeps throwing errors…)
24 April 2006 at 10:08 am
John: those are tokens in the grammar to make the lexer/parser error messages read well, since it will give the token names when they are defined. The values are not specified by me, but flex generates them. I guess you are saying that something in windows defines TRUE and FALSE already - which seems rather rude to do for you (I think C99 bool doesn’t do that). Maybe you could #undef them in win32_raptor_config.h which everything will include early on?