refactor: nip01 immutable

[1-leo]

No description provided.

[1-leo]

• signers now return a event because its immutable
• => problem in NdkBroadcastResponse because publishEvent need to be a future TBD
• check AuthEvent extend
• check ZapRequest extend

[codecov]

Codecov Report

❌ Patch coverage is 88.30189% with 31 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.22%. Comparing base (f368849) to head (22f7e13).

Files with missing lines	Patch %	Lines
...es/ndk/lib/domain_layer/entities/nip_01_utils.dart	73.68%	10 Missing ⚠️
...kages/ndk/lib/shared/isolates/isolate_manager.dart	65.38%	9 Missing ⚠️
...in_layer/usecases/proof_of_work/proof_of_work.dart	60.00%	4 Missing ⚠️
...layer/repositories/signers/nip46_event_signer.dart	0.00%	2 Missing ⚠️
...ndk/lib/domain_layer/usecases/bunkers/bunkers.dart	0.00%	2 Missing ⚠️
.../ndk/lib/data_layer/models/nip_01_event_model.dart	98.38%	1 Missing ⚠️
...ayer/repositories/signers/bip340_event_signer.dart	50.00%	1 Missing ⚠️
...s/ndk/lib/domain_layer/usecases/relay_manager.dart	93.75%	1 Missing ⚠️
packages/ndk/lib/shared/nips/nip13/nip13.dart	93.33%	1 Missing ⚠️

Additional details and impacted files

@@                          Coverage Diff                           @@
##           fix-websocket-performance-issues-2     #332      +/-   ##
======================================================================
+ Coverage                               70.87%   71.22%   +0.35%     
======================================================================
  Files                                     139      142       +3     
  Lines                                    5157     5262     +105     
======================================================================
+ Hits                                     3655     3748      +93     
- Misses                                   1502     1514      +12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[1-leo]

@frnandu
We have this in both our engines with the refactor, the not signed event is returned.
I think the best way forward is to return a future (with a completer), but this would require more refactoring on all the clients...
What do you think?

[1-leo]

• refactor nip13 to non static (object model)
• isolate without Singleton? As usecase

[frnandu]

My proposed changes (to be discussed):

• re-naming Nip01EventService -> Nip01Utils and not in the usecases package
• dart:isolate is not supported on dart4web -> we need to either try/catch or use stubs depending on dart.library.html???

DartError: Unsupported operation: dart:isolate is not supported on dart4web
dart-sdk/lib/_internal/js_dev_runtime/private/ddc_runtime/errors.dart 274:3       throw_
dart-sdk/lib/_internal/js_dev_runtime/patch/isolate_patch.dart 149:3              _unsupported
dart-sdk/lib/_internal/js_dev_runtime/patch/isolate_patch.dart 117:19             get sendPort
package:ndk/shared/isolates/isolate_manager.dart 51:67                            <fn>

• in Nip01Event we can use late final String id; and then this.id = id ?? Nip01Utils.calculateEventIdSync in the constructor and still keep it immutable. No need for such big refactoring and break existing client app's code.

• why we need an isolate manager for json decoding, can't we also use Isolate.run for only such simple things as json decode? Or maybe try to use a ready lib like isolate_manager or work_manager instead of trying to do it ourselves?

see #348

[frnandu]

see previous comment

[1-leo]

#332 (comment)
=>
Todo

• fix web
• relay pool, IsolateManager

[1-leo]

noticed some weird behavior, if encodingIsolatePoolSize > 1 the tests fail. Looks like our code depends on the order of processed events, suggesting thats its not safe for concurrency. (see tests above)
@frnandu did you have the same problem when using Isolate.spawn() directly?

[1-leo]

I think the problem is that processing the eose event is always faster compared to a kind 1 etc.
Rn I dont see another option then to use debounce or to wait until all events are processed

[1-leo]

@frnandu please take a close look at 22f7e13 I am not sure if its the best approach to use a chain here. We may only care about the eose to be in order