Ecmaspeak (was: `String.prototype.symbolAt()`)
# Michael Dyck (11 years ago)
On 14-02-15 07:44 PM, C. Scott Ananian wrote:
Tooling could help. I've filed a number of minor bugs against the draft spec about undefined identifiers, unused variables, etc. In theory at least those issues could be automatically linted.
Some of the bugs I submit are uncovered by a lint-like process. However, there's a substantial amount of crufty code involved in converting the spec into a lint-able form, and probably most of the bugs I submit arise when that code fails to convert some chunk of spec text.
On Feb 15, 2014 9:47 AM, "Brendan Eich" <brendan at mozilla.com> wrote:
Tooling could help. I've filed a number of minor bugs against the draft spec about undefined identifiers, unused variables, etc. In theory at least those issues could be automatically linted.
It's tempting to say that at some point JS should be defined in terms of "core JS", a strongly typed (!) subset of the language. (Strongly typed because verification is the point. Alternatively you could just make an executable test suite normative...)
FWIW I did a decent amount of work writing a JS runtime in "TurtleScript", a differently-minded JS subset. One could also write specs in/compile to asm.js but that's probably swinging the pendulum too far (and exposing a lot of unnecessary implementation). Ian Piumarta has done a lot of good work on self-hosting languages which look very similar to a proto-JS.
I'm sure you mean ungood, which is plusgood goodthinkful.
The original text under review here was discussing an optimization of the String.at polyfill, so performance was indeed the point.