SpiderNode for Firefox chrome code

Myk Melez myk at mykzilla.org
Wed Dec 14 17:38:24 UTC 2016


> Richard Newman <mailto:rnewman at mozilla.com>
> 2016 December 13 at 17:21
> I'd put measuring this right at the top of the list. Gecko is already 
> huge, and if we're talking adding 15MB, this wouldn't be feasible to 
> ship on Android at all, which would mean no cross-platform features 
> could use it, which kinda defeats the point. We want Gecko to be 
> getting smaller, not larger.
Understood.

> I think that Node 7 has async/await, no?
It does, but it isn't clear if/when Node's core APIs will support it. Of 
course we can still promisify and then await them.

> That raises a broader concern: which Node version to support. I'd 
> perhaps be inclined to start with 7/8, given async/await, but for 
> comparison, FxA recently wrestled bugs just to get Node v4 into 
> production, so there's quite a spread within Mozilla already.
SpiderNode is tracking Node 7, so that (or 8, when it ships in April) is 
the version that we'd support.

> I also have a mild concern about Node native modules — in my limited 
> experience, just about anything interesting one would do with Node 
> ends up requiring some native module down in the guts, which might 
> have an impact on how much value we'd get from npm. Do you have a 
> native module compatibility story for SpiderNode yet?
I don't yet have a complete story, but I'm tracking the ABI Stable Node 
API project (NAPI) <https://github.com/nodejs/abi-stable-node>, which is 
designing an API for native modules that is intended to be ABI-stable 
across both Node (and thus V8) versions and JavaScript engines.

-myk

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/firefox-dev/attachments/20161214/f87e484b/attachment.html>


More information about the firefox-dev mailing list