Age | Commit message (Collapse) | Author | Files | Lines |
|
The `lv2_validate` tool is meant for checking data files against the LV2 spec.
The plugins should not be included here since they aren't part of the spec, and
might not even be available if LV2 is built/installed without them.
|
|
Using someValuesFrom here is redundant and overly complex, since the property
range specifies that it is for literals anyway. So, simply replace it with a
minCardinality of 1.
|
|
These cause warnings with some versions of Doxygen, but all used default values
anyway, so they can simply be removed without any effect.
|
|
|
|
It is more or less meaningless to have an explicit type of rdfs:Resource, and
some tools don't like it, since everything (including classes and properties)
is a Resource. Since this is an individual, owl:Thing is a better type.
|
|
This was obviously wrong. I'm not sure how this happened, probably a
copy/paste mistake.
|
|
I am not sure if anonymous classes like this are "officially" invalid for OWL,
but they cause problems with some tools. So, just replace them with the class
we actually use. With this, tools have no problem loading dcs.ttl as an OWL
Full ontology.
|
|
This adds more specific property types to support validation, and removed
doap:anon-root since it is unused by LV2 (and frankly I don't even know what
it's supposed to mean).
|
|
These are required by OWL tools to use vocabulary from another ontology.
Though we don't use OWL tools in LV2, they are at least useful for sanity
checking the semantic side of things. These properties are also nice to have
anyway, since they clearly describe the dependencies between specifications.
|
|
This avoids the need to do any datatype conversion for handling bounds. In
particular, it avoids troublesome range issues, like the maximum
xsd:unsignedLong being too large to fit in an xsd:long, which is likely what an
implementation without bigint support uses to represent xsd:integer.
So, avoid all of that by simply using the exact datatype in bounds for that
datatype, which makes more sense anyway.
|
|
This data costs memory, and I think it is better to not present things that
aren't useful to LV2 implementations at all here.
The whitespace annotations could be useful for cleaning up literals, but this
isn't currently used (and tends to be baked in to implementations anyway), so
remove them as well. If necessary they can be added later easily enough.
|
|
|
|
This fixes the patterns to properly match the spec (several were too strict or
otherwise impresive), and also reduces the regex syntax used to a basic and
nearly universal subset (without the use of counted replication) that should be
supported and handled consistently by even a very basic implementation.
|
|
|
|
It's unfortunate to diverge from the canonical document, but most of these
third-party schemas already do anyway. This may be technically fine according
to rdfs (I'm not sure), but it's better for the LV2 universe as a whole to
consistently eliminate such things.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
For consistency, and because clang-tidy can't figure out if a define is an
include guard unless it ends the file.
|
|
|
|
|
|
|
|
The initial gain coefficient in the code is 1.0 (full volume). The parameter
itself is in dB, though, so the matching external default value is zero.
|
|
This avoids warnings about potential null pointer dereferences, which is true,
but in this case the buffer are static so it's really a programming error if
there is ever an overrun.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Not all of these actually show up in the output, but it's unnecessary to use
these in Markdown anyway.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|