fix: builds on various archs/platforms re: PR 4331

[wileyj]

ref: #4331

these are the equivalent changes for master and develop.

fix building & cross compilation, including stacks-signer, tested building for the following:

• linux-glibc-x86-64
• linux-glibc-arm64
• linux-glibc-armv7
• linux-musl-x86_64
• linux-musl-arm64
• linux-musl-armv7
• windows-x86_64
• macos-x86_64
• macos-arm64

note: target branch here is for develop since there is a workaround in place already in master. this change won't affect that workaround the next time develop is merged to master

cc: @zone117x @hugocaillard

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (228e9fb) 82.35% compared to head (28f2cc3) 48.34%.

Additional details and impacted files

@@             Coverage Diff              @@
##           develop    #4335       +/-   ##
============================================
- Coverage    82.35%   48.34%   -34.02%     
============================================
  Files          404      404               
  Lines       292642   292642               
============================================
- Hits        241018   141482    -99536     
- Misses       51624   151160    +99536     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jcnelson]

Is there a way to simply build without TLS? If not, then this is fine.

[wileyj]

following discussion in #4331 , converting this to draft.
there may be another option vs modifying the cargo.toml's here - eg using a native arm runner to build the binaries.

[wileyj]

@xoloki can you confirm this is OK to do, or should be continue looking for build alternatives where we're not cross-compiling?

[xoloki]

The general approach of using default-features = false is correct.

I'm a little concerned that we are using different versions of wsts in the root and stacks-signer Cargo.toml files, but that was present before this PR so not a blocker.

[xoloki]

This is the correct way to include wsts to avoid including bindgen and friends.

[xoloki]

Is there a reason why we have such an old version of wsts here in the stacks-signer Cargo.toml? And why we have different versions here and in the root Cargo.toml?

[wileyj]

Can't say for certain, but best to leave as-is since this branch has been genesis-sync tested as working.

[wileyj]

The general approach of using default-features = false is correct.

I'm a little concerned that we are using different versions of wsts in the root and stacks-signer Cargo.toml files, but that was present before this PR so not a blocker.

any concerns with the similar PR #4331 ?

[xoloki]

The general approach of using default-features = false is correct.
I'm a little concerned that we are using different versions of wsts in the root and stacks-signer Cargo.toml files, but that was present before this PR so not a blocker.

any concerns with the similar PR #4331 ?

#4331 is for next which uses a workspace dependency for wsts so no concerns there.