Unexpected token ...

[juvenrui]

// blabla.js
const a = {a: 1, b: 2};
const b = {...a, c: 3};

node r.js -o blabla.js

output error: Error: Line 2: Unexpected token ...

[jrburke]

This looks to be an issue with Esprima: last release for it was 4.0.1 (2 years ago), and it seems like it is not keeping up with the new standards: jquery/esprima#1949

In that Esprima issue, another project switched away from Esprima to espree. I just tried dropping in espree into r.js, but some of the r.js tests fail (4 failing tests), so it is not a drop in replacement. I don't have time to look more into it now. So for now the workaround is to avoid object spread in the source code.

[restarian]

As a follow up, is there an easy way to pass in espima options with the config or cli? Maybe a {tolerant: true} option.

[jrburke]

Looking back in the code, there currently is not a way to provide options to pass to Esprima.

[ejdaly]

Looking at Esprima, it seems that the master branch is being updated fairly regularly (but no releases, as you mentioned).

I've tried pulling in a build from the master branch (along with a couple pending PRs related to ES2019 syntaxes), and adding to build/jslib/esprima.js

This seems to work fine (one test fails, but it also fails with the current r.js... It seems to be an outdated test file..?).

This would add support for:

• Optional chaining
• Optional catch
• Nullish coalescing
• for await...of (async iterators)

If you like, I can open a PR for this change?

[chrswk]

I would love to see an update to the latest Esprima master. Is there something I can help with directly, or something I can "encourage" through a bounty?

[chrswk]

Looking at Esprima, it seems that the master branch is being updated fairly regularly (but no releases, as you mentioned).

I've tried pulling in a build from the master branch (along with a couple pending PRs related to ES2019 syntaxes), and adding to build/jslib/esprima.js

This seems to work fine (one test fails, but it also fails with the current r.js... It seems to be an outdated test file..?).

This would add support for:

• Optional chaining
• Optional catch
• Nullish coalescing
• for await...of (async iterators)

If you like, I can open a PR for this change?

Do you mind creating a PR with all your changes? I only got started with fixing the test case, but you have already done the work integrating Esprima master.

[ejdaly]

@chrswk - apologies, I hadn't seen the reply. I'll take a look at opening a PR here when I get a chance.

I did the upgrade of Esprima-to-master in my fork if you want to check that out: https://github.com/ejdaly/r.js

(I seem to have made 2 small commits after that, so you may / may not want those commits...)

[ejdaly]

Aside: there may be some momentum starting to get a new Esprima release, which would be fantastic...

jquery/esprima#2078

[chrswk]

@ejdaly No worries. I've been using your version/fork for a couple of weeks now, so thanks again for your efforts!

[abhijithsreenivas]

Still optional chaining is failing r.js, web-packing. When this will get resolved?

[theloveofcode]

This causes problems with a lot of new JS functionality like optional chaining, spread, etc. I recommend requirejs upgrade to espira-next

[prantlf]

The parser bundled to r.js is too old. You'll need to upgrade it. See #998.