Bundle zx package about zx HOT 5 CLOSED

It might at least help #24 / #25 (I was already working on deno port via rollup).

Anyway, I fully understand and respect your priorities as maintainer. Will continue to progress on a new repo. Hopefully, we might rejoin works or partially share useful utils while evolving in parallel :)

(feel free to close issue if it is not helping for roadmap)

from zx.

I decided not to bundle zx.

from zx.

But on other hands it adds completely. Now we need build process, errors will originate from a different source.

I’m not sure what benefits you described is really worth the hassle.

Install cost? How much it is?
Standalone zx script? What benefits?
Memory footprint seems small.
What a dist analysis? It’s easier to no on source without any transformation.
ts is great, but project really simple.
hack with cjs is ugly, but will go away with json modules in future.

from zx.

The main reason I thought to bundle was to distribute zx as a standalone js shell that can be embedded or distributed easily for pipelines. (like into a docker image or a curl script for automation). While is possible to depend on npm/npx, this would increase complexities, edge-cases (dep regressions, npm failure), and considerations for usage environment (npm needs to be installed and non-node runtimes cannot be supported since depending on node_modules resolution).

By #51 build happens automatically during npm publish and only affects stack-trace of zx internal errors if exist so it is not much more complexity...

Anyway, if goal of zx project is to keep internal simplicity rather than portability and user performance, it is understandable to not bundle node_modules :)

I would be happy to continue these changes as a new project keeping credits to your awesome work on zx idea <3

from zx.

to distribute zx as a standalone js shell that can be embedded or distributed easily for pipelines

This is is more reasonable thought.

For now, I like it's minimalist design as only a wrapper around a few other libs.

Let's not introduce premature optimization, and see what real issues users will encounter.

from zx.