tari-project / tari / 35840025433
73%

DEFAULT BRANCH: development
Ran
Jobs 1
Files 780
Run time 2min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 63.103% (-0.8%) from 63.925%
35840025433

push

github

web-flow 
test(ledger): fix three assertions and records that overstated themselves (#8029)

Final review follow-up. All three were correct.

1. `annotate_junit` wrote the *default* model list into the JUnit XML,
not the
models that ran. `cmd_test` resolves a positional override into a local
`models` - `./scripts/ledger_speculos.sh test nanosplus` is a documented
form -
   but the annotator read the `MODELS` global, so that run archived
"nanosplus stax" while the console summary correctly said "nanosplus".
That is
exactly the misattribution the annotator was added to prevent, arriving
through the one override the summary already handled, and the XML is the
copy
that outlives the terminal. The resolved list is now passed as an
explicit
argument rather than relying on bash's dynamic scoping. `SEEDS` has no
   positional form and stays global.

2. Three device-drawn strings in `ExpectedReview::check` still reached
failure
   messages - and so the JUnit `<failure>` body - unescaped, while
`Transcript`'s doc asserted it was "the one place left writing a screen
raw".
`review` is the module whose whole job is comparing device screen text,
so it
was the densest source of them. All three now use `{:?}`; the expected
halves
stay plain, which also makes the two sides of a mismatch tell themselves
apart. The completeness claim is replaced by an enumeration of the
sites,
   because a claim like that gets trusted instead of re-checked.

3. The `u_x` comparison added with the network scenario was tautological
as an
argument about the network, and its failure message would have
misdiagnosed.
`u_x` is `r_x + e * x` and `r_x` is drawn afresh every call, so two
calls give
different `u_x` whether or not the network ever reached `e` - the check
cannot
fail because of a network bug, and equal `u_x` means nonce reuse rather
than
   anything about the network.

The check is kept, because nonce reuse is exactly what it does detect
and that
is worth having. What changed is what it claim... (continued)

90036 of 142680 relevant lines covered (63.1%)

467557.09 hits per line

Coverage Regressions

Lines Coverage ∆ File
624
54.44
-11.7% base_layer/core/src/chain_storage/blockchain_database.rs
266
34.14
-46.1% base_layer/core/src/base_node/rpc/query_service.rs
171
51.67
-3.97% base_layer/core/src/chain_storage/lmdb_db/lmdb_db.rs
39
71.12
-4.17% base_layer/core/src/base_node/sync/horizon_state_sync/synchronizer.rs
32
71.6
-3.88% base_layer/core/src/test_helpers/blockchain.rs
31
91.44
-4.74% base_layer/core/src/base_node/sync/header_sync/validator.rs
9
62.39
-2.56% base_layer/core/src/chain_storage/db_transaction.rs
8
96.46
-0.44% base_layer/core/src/validation/header/header_full_validator.rs
6
80.3
-9.09% base_layer/common_types/src/types/fixed_hash.rs
4
88.64
-4.55% base_layer/core/src/test_helpers/block_spec.rs
3
43.94
-3.03% base_layer/core/src/chain_storage/block_add_result.rs
3
98.43
-1.18% base_layer/core/src/consensus/base_node_consensus_manager.rs
3
87.04
-0.3% base_layer/core/src/proof_of_work/cuckaroo_pow.rs
3
22.64
-5.66% base_layer/core/src/validation/error.rs
3
67.44
-6.98% base_layer/node_components/src/blocks/historical_block.rs
3
67.57
-8.11% comms/core/src/peer_manager/error.rs
3
87.98
-0.38% comms/dht/src/actor.rs
2
78.13
-3.13% comms/core/src/pipeline/outbound.rs
2
82.41
-0.93% comms/core/src/protocol/messaging/inbound.rs
1
69.81
-0.24% base_layer/core/src/mempool/sync_protocol/mod.rs
1
82.22
-2.22% base_layer/core/src/proof_of_work/sha3x_pow.rs
1
89.82
-0.18% base_layer/p2p/src/services/liveness/service.rs
1
91.45
-0.29% comms/core/src/multiplexing/yamux.rs
1
84.29
0.16% comms/core/src/protocol/rpc/client/mod.rs
Jobs
ID Job ID Ran Files Coverage
1 35840025433.1 780
63.1
GitHub Action Run
Source Files on build 35840025433