The percentage rounding was wrong: 76 was rounded to 70 instead of 75.
Moreover, we now round to the greated multiple of 5 (except for 0).
Therefore, 76 is now rounded to 80.
o Covered modules are now explicitly listed in each testsuite. This
permits to cover multiple modules by one testsuite.
o An index.html file is generated. It lists all testsuites and
their covered modules.
Those modules add standard tags/nodes, Erlang-specific nodes and a few
nodes specific to Yakaz. The latter will probably move to another place
or be standardized in any way.
The yaml_app module is an application behaviour used to check the
configuration. The yaml_sup module is a supervisor behaviour only
included because an application must start a supervision tree.
This option must be specified like this:
{default_tags, [
{Prefix, Value}
]}
Prefix may be:
o {non_specific, "!"} or {non_specific, "?"}
o a string
Value must be a string.
The parser would crash if a coma (',') was found outside a flow context.
This is now fixed.
While here, improve errors handling API. Now, when an exception is
thrown, the parser is insde a {yaml_parser, Parser} tuple. Thus, the
caller doesn't need to rely on the #yaml_parser record to identify the
exception and therefore, the definition of this record remains private.
The document version can be forced using the following option:
{doc_version, {Major, Minor}}
where Major and Minor are positive integers.
When this option is specified, {Major, Minor} will be used, even if a
%YAML directive indicates another version.
YAML 1.1 supports non-ASCII line breaks:
o Next Line (0x85)
o Line Separator (0x2028)
o Paragraph Separator (0x2029)
If the document is YAML 1.1, handle these line breaks. If it's YAML 1.2
or above, add a warning and treat them as non-break characters.
This parser is written against the YAML 1.2 specification. It may
therefore be used as a JSON parser as well.
The project is an Erlang application. The module "yaml_parser" parses a
YAML source (a string, a file, whatever) and runs a callback function
for each token scanned. It's more or less like a SAX XML parser.
The testsuite provided isn't complete yet, though it contains all the
example found in the specification.
