Git version looks older than NPM

[martianboy]

Hi. I was wondering if you could push the latest version of this project to its Git repo as well, cause unless I'm missing something, it looks like the NPM version is ahead. I can't even successfully build the project without having to hack my way through. This is a great work, and I'd love to hack on it a bit. Thank you!

[GerHobbelt]

Bleeding edge is b0rked at the moment as work towards better action code block checking via transpilation through babel was aborted in 2019 due to RL priorities taking over and the realization that the chosen direction was not really doable.

I'll try to revive the bleeding edge in the next few months -- expect slow going.

[martianboy]

@GerHobbelt Thank you! Looking forward to it.

[GerHobbelt]

Please see comment #51 (comment) and onwards from that for info on the bleeding edge.

As stated there, if you feel very daring, you can try the stuff in the t3 branch and if you just want to run the bleeding edge jison-gho you could go and copy its /dist/ files into your /node_modules/jison-gho/dist/ as described as "patching" in https://github.com/GerHobbelt/jison/blob/t3/CONTRIBUTING.md#big-fat-warning-for-when-that-make-fails--the-need-to-patch-node_modulesjison-ghodist -- be sure to spin up your prayer wheels to full speed while you do this. 😉 😉 😇

What's new then?

What you will get is a jison-gho which /should/ be ES6 capable, as may be observed from its own grammar spec files, such as https://github.com/GerHobbelt/jison/blob/t3/packages/lex-parser/lex.y :

• look for lines containing bits like let lst = or let srcCode = in their "action code blocks".
  • all jison / jison-gho releases will go 💥bang💥 as they will, at best, generate code which has scope collisions as multiple let variable declarations with the same variable name end up in the same scope inside the large parser kernel switch/case statement. Bleeding edge jison-gho add curly braces where needed to scope these fellas.
• %{ import ... %} at the top /should/ work as expected.

I HAVE NOT tested jison-gho output in a rollup merge process like someone once asked a long time ago, but we might be getting much closer to a doable situation there.

Anyway, that's where we are today. There's movement, but haphazardly as time and other duties allow. 👼 🙏

[GerHobbelt]

If you try anything dev-like, be reminded: master is b0rked today. t3 branch is the only branch currently doing anything sort of sane when you try to build it using make.

Note: master will become viable again later, but there's no time frame for that yet, as I had bit of a fight with the code and will need to mix the remaining good bits into t3 before I overrule the bad parts of master by merging t3 into it.