[vbreuss]

AWT217 only restated CS7003 for a partially open [InjectProperty(typeof(...), name)] target, so it could only ever appear next to a compiler error, and it missed the one case C# accepts: typeof(Outer<>.Inner), which Roslyn marks unbound with arity 0 and which [InjectProperty] silently treated as an open entry matching every closing of Outer.Inner, while [ImportService(typeof(...))] rejected it as AWT215.

Both typeof attributes now follow one rule: the argument must be a closed type or an unbound generic definition, and anything else, including null, is AWT215, which names the attribute. A partially open argument is recognised from syntax for both attributes, which also fixes [ImportService(typeof(Outer.Inner<>))]: it was accepted next to CS7003 and only reported as an unused import (AWT176).

AWT217 never shipped (it was only listed as unshipped); its descriptor, docs section and tests are removed, and the id stays reserved and is never reused.

[github-actions]

Test Results

   27 files  ± 0     27 suites  ±0   1h 13m 46s ⏱️ - 35m 53s
1 673 tests + 5  1 673 ✅ + 5  0 💤 ±0  0 ❌ ±0 
8 130 runs  +15  8 130 ✅ +15  0 💤 ±0  0 ❌ ±0 

Results for commit 2ae5588. ± Comparison against base commit 460c78f.

This pull request removes 14 and adds 19 tests. Note that renamed tests count towards both.

Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinition(typeof(IClock))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinition(typeof(ILog<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinition(typeof(ILog<int>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinition(typeof(Outer<>·Inner<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ ReportsANullType
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ ReportsATypeNestedInAnUnboundGeneric
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215ImportServiceTypeNotClosedOrOpen ‑ ReportsOnAnImportedModule
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt217InjectPropertyPartiallyOpen ‑ DoesNotReportAwt217ForAClosedTypeOrAnUnboundDefinition(typeof(Outer<>·Inner<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt217InjectPropertyPartiallyOpen ‑ DoesNotReportAwt217ForAClosedTypeOrAnUnboundDefinition(typeof(Outer<>·Leaf))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt217InjectPropertyPartiallyOpen ‑ DoesNotReportAwt217ForAClosedTypeOrAnUnboundDefinition(typeof(Outer<int>·Inner<string>))
…

Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnImportService(typeof(IClock))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnImportService(typeof(ILog<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnImportService(typeof(ILog<int>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnImportService(typeof(Outer<>·Inner<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnInjectProperty(propertyName: "Clock", implementation: typeof(Outer<>·Inner<>))
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnInjectProperty(typeof(Outer<>·Inner<>), "Clock")
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnInjectProperty(typeof(Outer<int>·Inner<string>), "Clock")
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ DoesNotReportAClosedTypeOrAnUnboundGenericDefinitionOnInjectProperty(typeof(Outer<int>·Leaf), "Clock")
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ ReportsANullType(ImportService(null!), ImportService)
Awaiten.SourceGenerators.Tests.DiagnosticTests+Awt215TypeofArgumentNotClosedOrOpen ‑ ReportsANullType(InjectProperty(null!, "Clock"), InjectProperty)
…

♻️ This comment has been updated with latest results.

[github-actions]

🚀 Benchmark Results

Details

BenchmarkDotNet v0.15.8, Linux Ubuntu 26.04.1 LTS (Resolute Raccoon)

AMD EPYC 7763 2.96GHz, 1 CPU, 4 logical and 2 physical cores

.NET SDK 10.0.303

[Host] : .NET 10.0.14 (10.0.14, 10.0.1426.52326), X64 RyuJIT x86-64-v3
Job=InProcess  Toolchain=InProcessEmitToolchain  IterationCount=15

LaunchCount=1  WarmupCount=10

Realistic	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
baseline*	305.4 ns	1.00 ns	0.89 ns	1.15	568 B	1.00
Awaiten	266.6 ns	2.90 ns	2.72 ns	1.00	568 B	1.00
MsDI	640.9 ns	6.80 ns	6.03 ns	2.40	1104 B	1.94
Autofac	6,770.6 ns	60.19 ns	56.31 ns	25.40	10784 B	18.99
Jab	188.1 ns	1.61 ns	1.42 ns	0.71	432 B	0.76
DryIoc	419.1 ns	3.62 ns	3.39 ns	1.57	944 B	1.66
SimpleInjector	749.6 ns	1.66 ns	1.47 ns	2.81	1096 B	1.93
PureDI	202.9 ns	5.51 ns	5.16 ns	0.76	632 B	1.11

Details

BenchmarkDotNet v0.15.8, Linux Ubuntu 26.04.1 LTS (Resolute Raccoon)

AMD EPYC 7763 2.81GHz, 1 CPU, 4 logical and 2 physical cores

.NET SDK 10.0.303

[Host] : .NET 10.0.14 (10.0.14, 10.0.1426.52326), X64 RyuJIT x86-64-v3
Job=InProcess  Toolchain=InProcessEmitToolchain  IterationCount=15

LaunchCount=1  WarmupCount=10

Resolve	Size	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
baseline*	8	7.050 ns	0.0091 ns	0.0071 ns	0.91	-	NA
Awaiten	8	7.780 ns	0.0134 ns	0.0119 ns	1.00	-	NA
MsDI	8	7.248 ns	0.0113 ns	0.0101 ns	0.93	-	NA
Autofac	8	137.448 ns	2.4831 ns	2.3227 ns	17.67	656 B	NA
Jab	8	3.141 ns	0.0060 ns	0.0050 ns	0.40	-	NA
PureDI	8	5.432 ns	0.0092 ns	0.0081 ns	0.70	-	NA
DryIoc	8	9.071 ns	0.0166 ns	0.0147 ns	1.17	-	NA
SimpleInjector	8	11.303 ns	0.0325 ns	0.0272 ns	1.45	-	NA
baseline*	256	8.086 ns	0.0577 ns	0.0482 ns	0.96	-	NA
Awaiten	256	8.447 ns	0.0119 ns	0.0093 ns	1.00	-	NA
MsDI	256	7.240 ns	0.0188 ns	0.0157 ns	0.86	-	NA
Autofac	256	144.863 ns	2.3103 ns	2.0480 ns	17.15	656 B	NA
Jab	256	42.768 ns	0.0825 ns	0.0732 ns	5.06	-	NA
PureDI	256	8.281 ns	0.0125 ns	0.0110 ns	0.98	-	NA
DryIoc	256	9.032 ns	0.0127 ns	0.0099 ns	1.07	-	NA
SimpleInjector	256	14.556 ns	0.1164 ns	0.1089 ns	1.72	-	NA

Details

BenchmarkDotNet v0.15.8, Linux Ubuntu 26.04.1 LTS (Resolute Raccoon)

AMD EPYC 9V74 2.87GHz, 1 CPU, 4 logical and 2 physical cores

.NET SDK 10.0.303

[Host] : .NET 10.0.14 (10.0.14, 10.0.1426.52326), X64 RyuJIT x86-64-v3
Job=InProcess  Toolchain=InProcessEmitToolchain  IterationCount=15

LaunchCount=1  WarmupCount=10

Build	Size	Mean	Error	StdDev	Ratio	Allocated	Alloc Ratio
baseline*	8	20.91 ns	1.053 ns	0.985 ns	1.01	208 B	1.00
Awaiten	8	20.678 ns	0.3706 ns	0.3466 ns	1.00	208 B	1.00
MsDI	8	1,519.538 ns	22.1124 ns	18.4648 ns	73.50	5688 B	27.35
Autofac	8	34,344.387 ns	84.6261 ns	79.1593 ns	1,661.34	32482 B	156.16
Jab	8	8.366 ns	0.2037 ns	0.1806 ns	0.40	96 B	0.46
PureDI	8	17.099 ns	1.3799 ns	1.2907 ns	0.83	128 B	0.62
DryIoc	8	762.690 ns	13.3069 ns	12.4473 ns	36.89	1528 B	7.35
SimpleInjector	8	11,674.368 ns	112.7413 ns	105.4583 ns	564.72	24761 B	119.04
baseline*	256	171.22 ns	13.908 ns	13.009 ns	1.25	4176 B	1.00
Awaiten	256	136.735 ns	1.5229 ns	1.4245 ns	1.00	4176 B	1.00
MsDI	256	14,103.919 ns	205.5773 ns	192.2971 ns	103.16	61016 B	14.61
Autofac	256	823,979.551 ns	3,577.4380 ns	2,987.3228 ns	6,026.70	738333 B	176.80
Jab	256	63.416 ns	1.4321 ns	1.3396 ns	0.46	2080 B	0.50
PureDI	256	70.727 ns	0.5152 ns	0.4303 ns	0.52	2112 B	0.51
DryIoc	256	47,820.950 ns	502.0500 ns	445.0542 ns	349.77	80645 B	19.31
SimpleInjector	256	362,993.393 ns	3,845.0289 ns	3,596.6424 ns	2,654.98	573092 B	137.23

baseline* rows show the corresponding Awaiten benchmark from the most recent successful main branch build with results, for regression comparison.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
90.9% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

This is addressed in release v0.8.0.