Mandatory

Theon Documentation Completion

  • Complete updates to TheonBook so that it is consistent with the current state of play. This includes adding some extended generic examples (such as updating "self" and using "bookworld", "pgcatalog" as a demonstration). Incorporate recipes as generic workflow examples.
  • Complete the XSLT that automatically produces the HTML model schema documentation.
  • Complete the XSLT that automatically produces the HTML model desktop documentation
  • Implement the DocBook XSLT transforms that will automatically produce documentation on a model for subsequent Book inclusion. Refactor the existing XSLT for the HTML outputs above so it can be included (re-used) to build both the HTML and the DocBook outputs.
  • Add relevant "info" entries for all elements in XSDDB so that the automatically produced HTML documentation for the "self" model is complete.

Theon Completion

  • Complete schema management desktops.
  • Fix: When using the XSDDB desktops to add an entity the pgcat_name is set but NOT pgcat_entity_name, however the latter is used predominantly in XSD generation.
  • Fix: The pgcat_* fields that should be set from foreign key values (and update when they change) are not being given values on insert/update (e.g. when adding attribute the pgcat_entity_name and pgcat_database_name need to be set to the values from the owning entity). This causes a subsequent import to duplicate the record rather than attaching to it.
  • Fix: After a "reload" the _th_COLUMN_changedby value is left as the owner running the reload, a subsequent "gather" does not then ever update the corresponding COLUMN value until the _th_COLUMN_changedby value is NULL.
  • Fix: The local "xsddb" is still always being used by "gather" due to it not loading any live profile for forge connection parameters.
  • Fix: A blank model_tag for imported/reloaded model should be made illegal as causes subsequent export problems (duplicates due to profile fallback), so need to explicitly set to name/pgcat_name etc.
  • Fix: on "core" reload - the derived output is written to one place but looked for at another, requires manual copy for now.
  • Fix: Need to remove all pre-existing non-model folder content in "derive".
  • Fix: Complete the correct generation of ui.conf (includes profile based db etc) and put into the right place so the UI finds it as the default config file.
  • Fix: A change in schema (to "self") requires ddldata.xsl to be re-generated but as this is done as a sub-layer (hidden) a pre-existing version will mask/stop regeneration.
  • Fix: All issues documented in individual recipes so that these can be presented as clean examples.
  • Fix: The "server" action needs to default to using from_path as a base for the config file (and we need to generate the ui default.conf correctly, see other item).

Theon and Hypatia Service Completion

  • Decommission and remove the old infdb/theon LCFG headers.
  • Complete a full infdb -> xsddb -> schemat drift correction process. In principle this is working but some tidying is required to remove noise change from schemat including suppressing all th named elements and removing th from role names (both of these probably done at the import layer).
  • Merge back SVN branch into trunk (on completion of mandatory changes above) and then migrate to the individual Git repositories. Publish Git repositories to the world so packages are visible via the yum repository. Then Blog.

Optional

Hypatia Documentation Completion

  • Review and update the "info" entries so that the automatically generated documentation is comprehensive and correct.

Theon and Hypatia Service Completion

  • Re-structure the Theon/Hypatia LCFG headers, including the ways packages are specified and folding back of last minute additions. This needs done at the lcfg, dice and live layers as well as individual machine profiles.
  • Package for portal cgi-bin needs fixed as it is owned by root on installation.

Theon Completion

  • Write the XSLT to generate the SVG for relation/process structure in documentation.
  • Write the XSLT to generate the SVG for desktop structure in documentation.
  • Add/check container support for unsupported sql.
  • Add aggregate function support (for pagila etc).
  • Add unique index for "uuid" in all "self" tables.
  • Code scan/review - go through it all and tidy up.
  • There is a "skip" flag in the stream table that is passed down through XSD and XSLT into the coupler configuration files, but the toolkit does not use it as an option for the stream method data load - hence still need to override skip header via the actual Incoming call parameters. Arguably if we support skip lines at this level we ought to also support the encoding option to Incoming as well.
  • Need "without_cascade" at per-bond level to sort "institute_role" couples with non-unique pkey lookup think this affects a feed.
  • Toolkit functions (and command line) don't return sensible (or any) error codes on failure - they should do to avoid things like the post-log grep we now have to do in the incoming script to detect failure.
  • Model "stream" command is missing the zero length check on the input file (over and above the same check post load in the coupler function itself).
  • Retest full cycle with "other" db's.
  • Complete "diff ddl" in toolkit and implement at least one alteration rule (e.g. add table column) and fail fallback as default for others.
  • Complete "release" action - it needs a "dryrun" option which can repeatedly build releases during testing but does not alter indexes and uses "unstaged" content.

-- TimColles - 07 Dec 2016

Topic revision: r10 - 14 Aug 2017 - 12:19:16 - TimColles
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies