Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Commits on Sep 16, 2015
Commits on Jul 6, 2015
  1. html: complete "in head" insertion mode support

    This accepts `noframes`, `noscript`, `script`, `style` and `template`
    elements in the `head` section.
    Aside from correctness, this is needed for various websites that place
    the `Content-Type` meta element after a `script` element.
Commits on Jul 4, 2015
  1. Update to ucd-tools 8.0.0.

Commits on Jun 23, 2015
Commits on Jun 13, 2015
Commits on Jun 12, 2015
  1. engines: properly account for custom offsets

    This makes sure that onevent calls are made for the document items
    prior to the range being spoken.
  2. engines: fix the text offset from a document range

    This ensures that the codepoint ranges passed to ontextrange are
    relative to the start of the document, not the start of the document
  3. document item: track text ranges

    This keeps a running track of the start and end character of the
    document_item events generated from a document_reader.
Commits on Jun 7, 2015
  1. synthesis_callback: add an onevent callback

    This allows users of cainteoir-engine to hook into the document
    events being spoken, allowing them to track additional state
Commits on Jun 6, 2015
Commits on Jun 5, 2015
  1. Rework the vorbis comment API.

Commits on Jun 3, 2015
Commits on May 13, 2015
Commits on Apr 14, 2015
  1. L2P: relax feature matching to support diphthongs

    This allows affricates and diphthongs to match single feature group
  2. L2P: recurse into next_match when ignoring elision

    When in the ignore elision state (i.e. when finding the next phoneme
    to match against), recursively call next_match until a phoneme is
    This ensures that as long as there are characters to match (i.e. the
    elided phoneme is not at the end of the word), there will always be
    a next phoneme.
  3. L2P: use the current phoneme for next matches

    When checking next feature/phoneme matches, use the phoneme from
    the current rule. This is because that phoneme will be the previous
    phoneme on a match of this rule.
  4. L2P: clear the current phoneme on new words

    This fixes previous phoneme matching for the first letter in a new
    word on previous phoneme/feature rules.
Commits on Apr 7, 2015
  1. Rename match_{phoneme => features}_(prev|next).

    This better reflects the intention of the methods as they check
    features expressed in the Kirshenbaum explicit feature syntax. It
    also allows `match_phoneme_(prev|next)` to be used for actual
    phoneme matching using the langdb phonemeset.
  2. L2P: Support multi-feature phoneme checks.

    This extends the feature checks to support multiple features using
    commas to separate the features. This matches the Kirshenbaum
    explicit feature syntax, allowing exact phoneme matches.
Commits on Mar 31, 2015
Commits on Mar 20, 2015
Something went wrong with that request. Please try again.