Preserialization dylib, Connor's work

[maackle]

Summary

Running @Connoropolous's #2203 on CI

TODO:

• CHANGELOG(s) updated with appropriate info
• Just before pressing the merge button, ensure new entries to CHANGELOG(s) are still under the UNRELEASED heading

[maackle]

@Connoropolous also could you just write (here is fine) a line for the CHANGELOG about this change?

[Connoropolous]

@Connoropolous as for actually running the tests, I'm not sure what you're running into that wouldn't allow it. I did update the wasmer dep, was that what was stopping you?

Holochain being a complex project, i dont know what commands ro run

[Connoropolous]

@Connoropolous also could you just write (here is fine) a line for the CHANGELOG about this change?

Yes sure. When by? Having a busy day.

[Connoropolous]

I made a commit to make this test pass but we should revert it if you think we should update the test to do wasm serialization f777f60

That makes sense. Leave it in i suggest

[maackle]

Yes sure. When by? Having a busy day.

Passing tests and your changelog are the only things left before merging. So whenever. @Connoropolous

[maackle]

oh yeah, and another review would help!

[thedavidmeister]

why is this in zome types? this crate is for types we need compiled into both the wasm host and guest, but this seems like something we only need on the host?

[maackle]

Well it's been in this file for years, it just got shifted up a few lines in this PR. But yes, sounds like it should probably be moved to holochain_types

[maackle]

@thedavidmeister I agree but there's a bunch of stuff that would have to be rearranged so I'll add some TODOs but want to leave this out of this PR

[thedavidmeister]

@maackle np

[Connoropolous]

@maackle
should there be a changelog item for each changed crate?

[maackle]

@Connoropolous only for the public-facing stuff. I know that's a bit fuzzy since it depends who's using what crate but at least something in hc_bundle with the high level details. We're not too meticulous about making each crate's Changelog be complete on its own.

[Connoropolous]

@maackle

hc dna pack command now takes --dylib-ios option, which produces iOS optimized Zomes. These can be utilized by passing dylib: PathBuf for Zome configurations in dna.yaml files and other data structures based on ZomeManifest where Zomes are constructed.

ZomeManifest now takes a dylib argument, with the type Option<PathBuf>. It can be safely ignored in cases other than trying to execute on native iOS. It is used with artifacts produced by hc dna pack when it has been called with the --dylib-ios option.

[Connoropolous]

@thedavidmeister can we get a clear on your change request?