Skip to content

Meeting minutes 2021 01 07

Roger Sheen edited this page Jan 7, 2021 · 5 revisions

Attendance: Radu Coravu, John Pashley, Frank Wegmann, Robert Anderson, Jarno Elovirta, Jason Fox, Lief Erickson, George Bina, Roger Sheen

Item 1: Any updates about prior releases?

Project board for 3.6.1: https://github.com/orgs/dita-ot/projects/28

Couple of items in there now but not much.

Item 2: DITA-OT 3.Next Development status and updates

Project board for next release: https://github.com/orgs/dita-ot/projects/27

Discussion of Java upgrade, from 8 to 11. May cause some churn but need to move to a more recent version. Would this open up new options for code? Jarno mentions can do some new things but a lot of the upcoming cool features are in Java 16 or 17. Nothing major, just that 8 is really old by now. Some discussion around Java, input/output streams, and how files are managed on disk ... internal caching in newer versions may cause slow downs. Several on call will try it out. Issue with Java reported here: https://bugs.java.com/bugdatabase/view_bug.do?bug_id=JDK-8258196 For those who try it out, please report results on our Java 11 upgrade ticket: https://github.com/dita-ot/dita-ot/pull/3653

DITA 2.0 status? Robert brings up Jarno's idea about how to redo chunking. Especially with 2.0, we could just do a two pass model, with one pass to split files and one pass to combine them. Seems like this would significantly simplify processing, in so many ways -- better code, easier to maintain. Worry that if you have a lot of link rewriting, and have to do it twice, that could slow things down; but seems rare to use chunking in both directions (split and combine in one root map). Also need to address what happens if you have a mix -- DITA 2.0 chunk tokens in a root map and DITA 1.x tokens in a submap.

Jason Fox and Docker proposal: been using plugins to test. With Travis issues have switched over to Github Actions, want to use DITA-OT in container. Can't directly use the Docker packages because of where they are stored, have to recreate them and then use them. Would be nice to have a DITA-OT Action that can run it. Now have a docker package built to ghcr.io which is where needed. One outstanding issue that Jason encountered, if doing sort on content, uses a default that results in wrong sort order in index. Seems like this is an issue with the Java code that it should ensure the sort is using the right language, not something that is set in the Docker container? Yes, but the Docker default seems to be one that nobody would have on their system. But ... should be fixed in Java code so that regardless of system default, you get whatever is set in your content. If you build the same code on the same toolkit on one Swedish and one Japanese system, we should get the same result. It's sounding like there is a bug somewhere in the index collation that it's not picking up the right default. Seems that there are a couple of outcomes: we can set UTF8 as default in the docker image, and then fix the index collation issue that is causing the problem. Jason would also like to transfer the new DITA-OT action over to ownership of dita-ot organization; seems reasonable, even if Jason is the one to continue updating. Roger is willing to update docs, which is likely more work than remaining changes for the action.

Item 3: Doc updates and plans

  • Build issues with recent Gradle versions
    • Can be fixed by moving settings from properties files to Gradle build
    • Any reason to keep properties files? (no responses)
      (originally shared with Ant builds, no longer required for that)
  • Considering refactoring docs CSS in Sass for extensibility

Item 4: DITA-OT Day

Convex is this spring, which probably removes a fall conference

Item 5: Other topics

...

Item 6: Backlog discussion

Project boards: Project boards: https://github.com/orgs/dita-ot/projects/

Time to go through pull requests, issues tracker, determine how to handle open issues.

Clone this wiki locally