Skip to content

Latest commit

 

History

History
43 lines (25 loc) · 3.38 KB

changes.md

File metadata and controls

43 lines (25 loc) · 3.38 KB

Change log

Version 2.0.6

  • The "Topic: URL-to-an-issue" feature, that extracted the title from the issue reference and added a directive for the issue, worked for w3c hosted repositories only, even if the configuration file referred to different repo owners. This has been generalized.

Version 2.0.5

  • Filter out the zakim 'q' command (i.e., the standalone 'q' that displays the queue content with possible comments, one line per entry)

Version 2.0.4

  • Getting the titles of issues (or pull requests) is happening now through the Github/octocat interface layer rather than via direct GET URL-s to the Github API. i.e., the data is queried via the OAUTH approach, ie, as an authenticated request. This is necessary to avoid hitting the limit of 60 queries per minutes which is imposed at unauthenticated API accesses. (If a WG call handles lots of issues this limits may be easily reached…)

Version 2.0.3

  • The zakim "agendum" line output is converted into a bona fide Topic: line

Version 2.0.2

  • Added the possibility to consume the RDF/XML version of the RRSAgent log. This is a fallback that was made necessary by a bug somewhere on the server which, in some cases dumped only a portion of the textual log.
  • If the (Sub)topic is a single issue/pr directive, the title of the (sub)topic is retrieved from the relevant github issue/pr

Version 2.0.1

  • Bug in the string replacement function: the replaceAll function had a nasty bug when the 'to' included the 'from' (got into an infinite cycle). Replaced it by a neat trick found on the Web. (Note: node with version < 15 did not implement the replaceAll function for strings, hence the necessity of having this...).

Version 2.0.0

  • The package has been fully converted to TypeScript. The generated javascript code replaces the old code at the same place, i.e., deployed scripts should not be affected.
    • The TypeScript conversion also resulted in a major review of the code, including the usage of TypeScript-adapted lint. Too many minor changes to record them all here
    • Some older libraries have been removed because unnecessary (e.g., safe-regex and moment) or its use greatly reduces (the underscore package is now restricted to the client-side script, should be removed at later).
    • Better adaptation to modern js environments, primarily the usage of async and await: some older modules in the code still painstakingly generated Promises "by hand", which is unnecessary by now.
  • A proper documentation has been added, generated by typedoc.
  • Minor feature changes have also been incorporated (tolerance of scribe errors when typing extra spaces after the nickname, better display of generated link lines). By and large the functionality remained unchanged, however.
  • The JSON-LD header incorporates now the resolutions and the actions, if applicable. This means that post-processing steps can get to the data without forced to interpret the markdown (or HTML) content. That is what metadata is for…
    • Note that while the schema.org vocabulary contains the notion of actions which (more or less…) aligns with what is required here, there is no term for the minute resolutions. As a consequence, the JSON-LD vocabulary extends the schema.org one by specific terms. The corresponding vocabulary document is in the ./vocab directory.

(This list is incomplete, was introduced starting at version 2.0.0 only… My sloppiness)