Textx-lang-irirefs

Fixed

- Added the missing `IRIRefResolveError` exception.

Added

- The method `recompose` in class `IRIReference` now admits an optional parameter `absolute = False`. When this parameter is called with value `True`, the fragment identifier is omitted from the recomposition.

Changed

- the fields `query` and `fragment` now correctly contain `"query"` or `"fragment"` instead of `"?query"` or `"?fragment"`

Fixed

- IRegNames now correctly accept an empty input

Changed

- changed the method `resolve(self, base)` in class `IRIReference` to `resolve(self, relative)` in class `IRI`.
- in line with the previous change, the resolve command-line script now accepts `irirefresolve [-h] [-c] base relative` instead of `irirefresolve [-h] [-c] relative base`

Added

- in line with the previous change, added a method `resolve_seglist(self, seglist)` in class `ÌRI`, allowing to write:
python
base = parser.model_from_str('http://www.lirmm.fr')
resolved = base.resolve_seglist(['foo', 'bar'])
print(resolved.recompose)

to obtain `http://www.lirmm.fr/foo/bar`

Changed

- The root of the grammar is no more an IRIRef, but a IRIRefContainer, in order to allow for an IRIRef to have a parent object.

Added

- scheme based normalization

Changed

- The method `resolve` of class `textxirirefs.iri.IRIRef` now returns a model instead of a string. Tests and irrefresolve modified according to that change.
- Changed grammars and classes to properly classify AbsoluteURI and AbsoluteIRI programmatically, for efficiency reasons.

Fixed

- Resolving `bar` against `//foo` now correctly returns `//foo/bar` instead of `//foobar`. [Issue 2](https://gitlab.inria.fr/jfbaget/textx-lang-irirefs/-/issues/2) now closed.
- IRI is now correctly defined and its fragment is optional, an important feature when importing the grammar.

Added

- Added the `irirefresolve` command.