wiki:XmlNotes

Version 1 (modified by davea, 15 years ago) (diff)

--

XML notes

BOINC originally used a line-oriented parser with primitives like

parse_str(tag, buf, len);

This is not a general XML parser; it's able to parse only the XML generated by BOINC itself. For example, it can parse

<file>
   <url>http://foo.bar</url>
</fil>

but not

<file><url>http://foo.bar</url></file>

or

<file>
   <url>
      http://foo.bar
   </url>
</file>

More recently, we added a better parser (class XML_PARSER). However, much of the code still uses the old parser.

XML encoding

Both of the parsers decode XML entities when parsing strings.

The functions that generate XML don't generally encode XML entities in strings. Most strings are generated by BOINC itself and will never contain special characters. The following fields may contain special characters, and are XML-encoded by the functions that generate XML:

client:
PROJECT::user_name, team_name
FILE_INFO::url
APP_INIT_DATA::user_name, team_name
PROXY_INFO::http_user_name, http_user_passwd, socks5_user_name, socks5_passwd
HOST_INFO::p_vendor, p_model, os_name, os_version, serialnum

server (db_dump and db_purge):
USER::name, url
TEAM::name, url, name_hteml, description
RESULT::stderr_out

In addition, the RESULT::stderr_out (client) is enclosed in <![CDATA[ ...]]>