Have make dist in the web_client package build the client

[natefoo]

Not ideal because of package Makefile divergence. Also you have to use the Makefile, python -m build without it will get you an empty package. Maybe we would prefer to do this entirely in galaxy-release-util (although that does not address the second issue).

How to test the changes?

(Select all options that apply)

• I've included appropriate automated tests.
• This is a refactoring of components with existing test coverage.
• Instructions for manual testing are as follows:
  1. cd packages/web_client
  2. make dist

License

• I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[dannon]

Not a -1 or anything if we want this for this release, but I think you are right that it'd be best to link this to galaxy-release-util with the npm/depot publishing of the client build there.

[natefoo]

I just don't like that as-is the package will be empty if you build it without an external util. The same is true of the meta package, although in that case I believe it will update and commit the changes so that future direct builds will work.

[mvdbeek]

I do have a version of this that uses hatchet to build the client (but fails to actually put it in the wheel 😆 ), maybe that would be ideal ?

[natefoo]

I'm gonna need some context there - I found 3 projects called hatchet, none of which seem relevant to this.

[mvdbeek]

that's cause it's called hatch 😆. 2668ad7 is the non-functional WIP

[mvdbeek]

I spent a good bit of time getting the hatch build to work, but no real progress. I don't love the makefile procedure but this is better than nothing.

[mvdbeek]

Somehow I can't comment on the diff, but python -m build should probably go to line 67 so that the client build happens before be build the package ?

[natefoo]

Make deps (galaxy/web_client/client_build_hash.txt) ensure that does it not?

[mvdbeek]

Can you tell i've never done more than copy-paste a makefile ? 😆