Renaming this makes it more clear what is being signed.
Depends on D4240
Details
Details
- Reviewers
deadalnix - Group Reviewers
Restricted Project
Should still succeed:
./gitian-build.py -c -n -b -o m gitianuser v0.20.3
This runs and then errors part way through since we currently don't maintain this codepath.
At least the script doesn't break in unexpected ways.
./gitian-build.py -c -n -s -o m gitianuser v0.20.3
Diff Detail
Diff Detail
- Repository
- rABC Bitcoin ABC
- Branch
- gitian-split-sign
- Lint
Lint Passed - Unit
No Test Coverage - Build Status
Buildable 7744 Build 13527: Bitcoin ABC Buildbot (legacy) Build 13526: arc lint + arc unit
Event Timeline
Comment Actions
First, I'm not convinced this is ambiguous. What would you sign in there that makes any sense? Is that worth breaking things? Second, the test plan doesn't cover the feature whatsoever.
Comment Actions
Read through build() and it should give a better picture why it's ambiguous. Running the script with -b set but -s not still requires the signer argument and uses it as part of gsign which is called during build(), not sign().
This forces you to generate keys on every machine you run the build, even though you aren't interested in signing. It's unnecessary, so that was the rationale for D4242.