Closed chewselene closed 3 years ago
Merging #984 (b5a1dbe) into main (4fd8dda) will not change coverage. The diff coverage is
100.00%
.
@@ Coverage Diff @@
## main #984 +/- ##
=======================================
Coverage 94.33% 94.33%
=======================================
Files 113 113
Lines 8987 8987
=======================================
Hits 8478 8478
Misses 509 509
Impacted Files | Coverage Δ | |
---|---|---|
graphql_compiler/compiler/compiler_frontend.py | 98.57% <ø> (ø) |
|
graphql_compiler/compiler/cypher_query.py | 100.00% <ø> (ø) |
|
graphql_compiler/cost_estimation/statistics.py | 93.05% <ø> (ø) |
|
...aphql_compiler/schema_generation/graphql_schema.py | 95.89% <ø> (ø) |
|
...schema_generation/orientdb/schema_graph_builder.py | 95.87% <ø> (ø) |
|
...hema_generation/sqlalchemy/sqlalchemy_reflector.py | 97.95% <ø> (ø) |
|
...l_compiler/compiler/ir_lowering_cypher/__init__.py | 100.00% <100.00%> (ø) |
|
..._compiler/compiler/ir_lowering_gremlin/__init__.py | 100.00% <100.00%> (ø) |
|
...ql_compiler/compiler/ir_lowering_match/__init__.py | 100.00% <100.00%> (ø) |
|
...ql_compiler/compiler/ir_self_consistency_checks.py | 97.05% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 4fd8dda...b5a1dbe. Read the comment docs.
Instead of
self_consistency_check
andself_consistency_checks
, what aboutassert
andassertions
? They describe what's actually happening, are unambiguous to the reader, and are much easier to read, type, and say.
I originally had this as validate/validation! gonna loop in @obi1kenobi who said:
would it be okay to use self-consistency check instead of validate when possible for this? there’s a subtle distinction, but compiler IR can be locally valid (all real blocks with valid arguments etc.) but self-inconsistent (invariants that span multiple blocks are not upheld) so for example, if there’s a Traverse without a matching Backtrack, or a Traverse to a location that is never marked with MarkLocation
I think assert/assertion likely has a similar issue to validate/validation, but Branen also pointed out that all these functions do raise assertion errors so maybe it is appropriate?
Remove ableist language.