Skip to content

Python 2 and 3 compatibility#4

Merged
dstanchfield merged 1 commit into
keepkey:masterfrom
kyuupichan:master
Oct 22, 2016
Merged

Python 2 and 3 compatibility#4
dstanchfield merged 1 commit into
keepkey:masterfrom
kyuupichan:master

Conversation

@kyuupichan

Copy link
Copy Markdown
Contributor

Updated build_pb.sh to hack the right thing with sed.
Included my compiled files.
Used protoc version 2.6.1 from protobuf3.

Updated build_pb.sh to hack the right thing with sed.
Included my compiled files.
Used protoc version 2.6.1 from protobuf3.
@romanz

romanz commented Oct 21, 2016

Copy link
Copy Markdown
Contributor

Thanks @kyuupichan! This PR works well for me (I was using protoc-3.0.0).
@dstanchfield Could you please merge this PR (and update the *_pb2.py files as well)?
I'd be happy to use this package with Python 3 :)

@dstanchfield
dstanchfield merged commit d3a0b76 into keepkey:master Oct 22, 2016
pkhx02 added a commit that referenced this pull request Dec 9, 2016
keepkeyjon pushed a commit that referenced this pull request Sep 5, 2018
* add tagged release downloader

* add hmac calculation for release binaries
BitHighlander added a commit that referenced this pull request Jul 7, 2026
…gaps

Audit of fork develop CI run 28679110349 (PR #284 clearsign v2 static-
schema, 526 junit tests) against the who/what/why standard for every
major daily-driver EVM tx format. Ten fixes:

1. Lower 8 stale requires_firmware("7.15.1") gates to 7.15.0 (signing
   guards x5, blind_sign_blocked/allowed, bip39-word-reject) — the
   behaviors ship in the 7.15.0 line under test; the gate zeroed out
   the EIP-1559 wrong-signer regression suite and the blind-sign
   policy negative path on every 7.15.0 CI run.
2. Document the ETH-section amount-unit rule (Wei below 1 gwei, scaled
   ETH/ticker above) so toy conformance-vector screenshots aren't
   misread as a "never wei" violation.
3. Disclose EIP-712 typed-data hash-signing as a known gap (E16): the
   device signs two host-computed hashes with zero readable domain/
   message display — a daily-driver format with no who/what/why today.
4. Fix the ERC-4337 handleOps flow's overclaiming "innerCall: decoded
   separately, not raw" placeholder — the representative UserOp's
   inner callData is actually empty; the claim is now honest about
   what this flow does and does not prove (regenerated its frozen
   reference-blob snapshot to match).
5. Add 2 new on-device tests for the v2 static-schema headline security
   property, previously unit-tested but never proven on real device
   behavior: calldata/schema length-mismatch falls back to the
   blind-sign gate (VS6), and an unsupported dynamic arg format is
   independently rejected as MALFORMED by the device parser (VS7).
6. Surface the 3 silently-skipped Uniswap V2 liquidity tests (E17-E19)
   with their documented emulator-limitation reason instead of leaving
   them absent from SECTIONS entirely.
7. Surface the 3 passing-but-invisible THORChain-router EVM deposit
   tests (E20-E22), including the router-pin blind-sign-gate proof
   that closed a CRITICAL bypass.
8. Surface 6 passing-but-invisible malformed-blob rejection tests
   (V7a-V7f: empty/truncated/trailing-bytes/wrong-version/zero-sig/
   empty-slot) — the WHY = fail-closed story had no rejection-path
   evidence in the release-gate document.
9. Fix the stale "7.15.1 NEW FEATURES" banner comment and the V8
   context text that claimed "covered in 7.15.0+" while pointing at a
   test gated to 7.15.1 (a self-contradiction in the PDF's own prose,
   resolved by fix #1).
10. Fix print_clearsign_flows() (--flows external-signer reference
    dump): KeyError on flow['shows'] (no catalog flow has that key —
    regression from the move to the shared catalog module) and a
    missing required key_id arg to flow_blob().

Verified: all 6 edited files compile; SECTIONS loaded in the real
firmware-checkout env (245 unique entries, 224->245, zero ID
collisions); every new/changed (module, method) SECTIONS reference
resolves to a real test function; all 51 REFERENCE_BLOB_SNAPSHOTS
still match (only the ERC-4337 entry's hash/length changed, matching
the ARG_FORMAT_STRING value edit in fix #4).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants