Markup.ml 0.7.4

• Decrease parsing time of HTML parser by 20-95%, depending on properties of the
  input (#15, #16, af90cdf, e1d78e7). The XML parser seems to have benefitted as
  a side effect, with parsing times decreased by about 18%.
• Fix stack growth in Markup_lwt and deprecate Markup_lwt.ensure_tail_calls
  (9e0219d, 1e0a078).
• Prefix internal module names to avoid conflicts.
• Install .cmx files for internal modules, for Flambda switches.
• Add Markup_lwt.load.
• Move opam to the root directory for easier pinning.

Markup.ml 0.7.3

• Adapt to Uutf 1.0.0 (#14).
• As a consequence, Markup.ml now requires OCaml at least 4.01, due to a
  constraint of Uutf.

Markup.ml 0.7.2

• Exposed input streams (CR-LF translation and locations assignment) in the public API, though this is deliberately omitted from the documentation for the time being.
• Added compare_locations.
• Fixed parse_html not honoring the ~context argument.
• Made parse_html ~context element name case-insensitive, in line with the HTML treatment of element name (prompted Gabriel Radanne).
• Added row variables to some polymorphic variant arguments (requested Gabriel Radanne).
• Now installing .cmx files for OCaml 4.03 (reported Gabriel Radanne).
• Fixed an instance of warning 52 (matching for Failure with a specific string).

Markup.ml 0.7.1

• New trees function for converting a signal stream to a stream of
  tree data structures.
• Fixed bug in existing tree function that caused it to consume all
  signals, instead of only those needed to construct one tree.
• Comments, processing instructions, XML declarations, and doctype
  declarations can now be optionally included in tree assembly.
• Added tree deconstruction helper from_tree. This converts a tree to a
  signal stream (prompted by Angelos Oikonomopoulos).
• Added predefined namespace URIs, e.g. Markup.Ns.html is the string
  "http://www.w3.org/1999/xhtml".
• Added transform function, which maintains an accumulator like fold
  does, but produces a derived stream instead of the final value of the
  accumulator (prompted by Angelos Oikonomopoulos).
• Renamed the internal module Xml_parser, to avoid conflict with
  Xml-Light's internal module of the same name (reported by Angelos
  Oikonomopoulos).

Markup.ml 0.7

• Implemented the adoption agency algorithm in the HTML parser (a major internal complication).
• Fixed many bugs. Thanks to Leo Wzukw and Jerome Vouillon for reports.
• Added more HTML parser tests for more confidence.

Markup.ml 0.6

Breaking changes:

• Renamed content to text.
• Changed Text of string` to Text of string list` (suggested by Daniel Buenzli).
• Removed drop_locations and replaced with an explicit type family 's parser and function signals.

Other changes:

• Added a new function content, which filters out non-element and non-text signals (prompted by Daniel Buenzli).
• Added conformance status to documentation.

Markup.ml 0.5

Initial release.