Add `noCheck` API option #57934

weswigham · 2024-03-25T18:41:12Z

This PR adds noCheck, a compiler option which instructs the compiler to skip the semantic diagnostics phase. This means the compiler skips right form program construction, parsing, and binding, to emit. Implementation-wise, the basically just considers every file the same way we do declaration files under skipLibCheck, or project reference redirects.

As-is, there are some limitations on this - declaration emit works wonderfully already - every test with declaration emit on also tests declaration emit with noCheck, while JS emit needs some functionality disentangled from the checker's full-tree-error-walk to consistently work with this. As such, noCheck requires emitDeclarationOnly for now. (This limitation could be lifted with some additional work to make a lazy version of the .referenced node links calculations in checker, and by making lazy versions of all the calculations implied by the getNodeCheckFlags calls in the JS transforms - which would be required to do #50699 and for this to fulfill #29651) Additionally, passing noCheck alongside noEmit is an error - there's nothing wrong with this combination of options, it just seems like it doesn't do anything.

noCheck is not currently exposed on the commandline or to tsconfig files, and is an internal API, as there was some concern at the design meeting that it might need to change a bit when we add support for JS emit to it.

…/declarations without full checking

…sentagling NodeCheckFlags and .referenced from the checkers full traversal)

…king, accept baselines with extra output due to less errors, or union order changes

…ly, refine noCheck test conditions to match

typescript-bot · 2024-03-25T18:41:27Z

Looks like you're introducing a change to the public API surface area. If this includes breaking changes, please document them on our wiki's API Breaking Changes page.

Also, please make sure @DanielRosenwasser and @RyanCavanaugh are aware of the changes, just as a heads up.

weswigham · 2024-03-25T18:46:22Z

src/compiler/checker.ts

@@ -9038,6 +9038,7 @@ export function createTypeChecker(host: TypeCheckerHost): TypeChecker {
            }

            function serializeSymbol(symbol: Symbol, isPrivate: boolean, propertyAsAlias: boolean): void {
+                void getPropertiesOfType(getTypeOfSymbol(symbol)); // resolve symbol's type and properties, which should trigger any required merges


The exhaustive testing in the harness revealed that some JS declaration emit cases were a bit broken with out-of-order declaration emit (and were emitting declarations twice) - these two changes fix them by ensuring the checker work to unify weird JS merges is always done before we look at the symbol's identity (so the merge symbol should always exist if it will exist).

RyanCavanaugh · 2024-03-25T18:55:19Z

skipDefaultLibCheck
skipLibCheck
skipCheck

weswigham · 2024-03-25T20:02:09Z

But //@ts-nocheck and --noEmit 🤔

sheetalkamat · 2024-03-25T20:51:02Z

This is not really --no-check its just dont report diagnostics right? Because getEmitResolver still checks sourceFile and ensures that file is typechecked

weswigham · 2024-03-25T21:59:41Z

This is not really --no-check its just dont report diagnostics right? Because getEmitResolver still checks sourceFile and ensures that file is typechecked

It doesn't request any diagnostics from the checker, so it skips the whole exhaustive check traversal - on our own codebase this skips basically all checker time - the bits of the checker declaration emit actually pulls on to produce output are minimal. In theory, yes, it could cascade into a complete program type check. In practice, it never does, for the same reason completions basically never results in a whole program typecheck - the checker's lazy.

weswigham · 2024-03-25T22:07:07Z

Specifically: The skipTypeChecking function is used inside checkSourceFile, so even though createEmitResolver tries to call getDiagnostics, with noCheck on, it does nothing~

sheetalkamat · 2024-03-27T20:54:37Z

src/compiler/program.ts

+            if (!options.emitDeclarationOnly) {
+                createDiagnosticForOptionName(Diagnostics.Option_0_cannot_be_specified_without_specifying_option_1, "noCheck", "emitDeclarationOnly");
+            }
+        }


for now we should disable this for incremental as the build info generated will not handle this correctly.

or we need to make sure noCheck recalculates semantic diagnostics and stores the that flag into buildinfo

As in I should add an error for using noCheck alongside incremental?

or we need to make sure noCheck recalculates semantic diagnostics and stores the that flag into buildinfo

That's just setting affectsSemanticDiagnostics in the option descriptor when we add it back to being enabled for cli/tsconfig in the future, right?

"in option" yes but now it needs to be handled in builder or reported as error to have incremental and nocheck as buildinfo written will not be correct. It will show no errors and next time you run tsc it will not report errors . (without storing noCheck value in buildInfo and checking for that in builder)

Is 051a1d1 along the lines of what you were thinking? Since it's not in the commonOptionsWithBuild, commandOptionsWithoutBuild, or optionDeclarations lists, it's never included in help or valid on the CLI/in a config, but since it's built into the OptionsNameMap, all the normal option serializing/deserializing machinery works with it - build info included.

We need test cases:

Run through API with --no-check , followed by tsc .(with and without incremental).

Run through API with --no-check , followed by tsc --b (with and without incremental) - have a feeling that tsc - in non incremental is going to find that everything is uptodate (there is no buildinfo to check, only output file stamps) and will not report errors?

You can add these testcases by using verifyTsc where the initial fs is generated through running API with --no-check

Run through API with --no-check

We.... don't have the harness for this right now. At least not a harness that then feeds into the existing verifyTsc stuff (obviously, all the compiler-style tests build against the API without using the executeCommandLine entrypoint). This is quite a lot to put together for something that's only supposed to be API-only temporarily... I started putting it together, and such a harness is looking like it's going to be more than the entire rest of the PR. 😵

So I'm not going to. This is gonna be a normal command line option, accepted on both the CLI and via tsconfig, except with a

extraValidation(value) { return [Diagnostics.Unknown_compiler_option_0, "noCheck"]; },

so anything that runs validation (API methods that take CompilerOptions obviously do not) reports that it doesn't exist, but otherwise, it uses all the usual pipelines for things. And then in the test where I'd like to consider it a real option, I'll override that validator with a noop temporarily, which I can do for the tsc -b style tests, to validate it will work correctly with -b once it's actually public and usable via.... anything other than the API.

…`noCheck`, it is serialized correctly

sheetalkamat · 2024-04-12T17:47:16Z

src/compiler/program.ts

@@ -4357,6 +4357,15 @@ export function createProgram(rootNamesOrOptions: readonly string[] | CreateProg
            }
        }

+        if (options.noCheck) {
+            if (options.noEmit) {


i think noEmitOnError and noCheck should also be combo that does not make sense.

noEmitOnError blocks emit on the presence of options, syntactic, global, semantic, and declaration diagnostics. noCheck only blocks semantic from being generated, so noEmitOnError does still have meaning alongside noCheck, though it certainly will block emit less.

sheetalkamat · 2024-04-12T22:34:42Z

src/testRunner/unittests/tsbuild/noCheck.ts

+                loadProjectFromFiles({
+                    "/src/a.ts": getATsContent(declAText),
+                    "/src/tsconfig.json": jsonToReadableText({
+                        compilerOptions: { noCheck: true, emitDeclarationOnly: true, declaration: true },


Having this in json and not reporting error is a bug because then it doesnt become API only option per design meeting.,

The scenario i am talking is missed here because:
modifying tsconfig.json will check that config file is newer than output and hence program will be rebuilt. The problem with doing this with API is that output will be updated but not inputs which means the build will report that everything is uptodate

Here is how to write this test case
1 load project files (without noCheck option) in your test case as is . No errors
2. As part of first edit: Do these steps to emulate API usage: Change a.ts to introduce error, run API (by creating program and emitting with noCheck on, use sys= new vfs.sys(fs)) to create compiler host
3. This will fail because after edit, test framework will run tsc --build again and it will find that everything is UpToDate and not report error. (test will not create discrepancy baseline because unfortunately we don't check discrepancies in error reporting just file emit validation)

Having this in json and not reporting error is a bug because then it doesnt become API only option per design meeting.,

And it does normally report an error, which is what the other test block that doesn't override the validator function verifies (and naturally, these collapse into one test once it's actually exposed via CLI)~

The problem with doing this with API is that output will be updated but not inputs which means the build will report that everything is uptodate

If you build via API and pass options not in your on-disk config file (as you would need to, since noCheck is API only right now), your up-to-dateness checks for configs are going to all be invalidated anyway, since your API-fueled build will have different options applied than a commandline one. Precisely

The problem with doing this with API is that output will be updated but not inputs which means the build will report that everything is uptodate

is in no way unique to this new flag - any time you do a API build with different flags than what your on-disk configs specify this'll happen, and I'm pretty sure how -b handles that is just entirely out of scope, unless we're talking specifically incremental output, which just comes down to the options serialized in the .buildinfo which is why:

This will fail because after edit, test framework will run tsc --build again and it will find that everything is UpToDate and not report error. (test will not create discrepancy baseline because unfortunately we don't check discrepancies in error reporting just file emit validation)

It both will and won't. If --incremental isn't on, yeah, it won't build again because all the timestamps say it's newer, and build doesn't know that your last build was with different settings. Again, that's pretty out-of-scope and isn't going to get fixed, since there's no incremental artifact. Nothing new there, that issue has nothing to do with noCheck - you can replicate that with any semantic diagnostic affecting option. For --incremental, that won't be the case, because the .buildinfo emitted by the API build will have different compiler options than the new build (semantic diagnostic affecting ones), which'll force a recheck (though not reemit), and is precisely what this test verifies without needing all the API-layering hoopla by temporarily enabling it in configs/cli (since an incremental build via config is going to produce the same buildinfo as one via api). AFAIK, the only critical thing to test is that noCheck is serialized into the buildinfo just like every other semantic diagnostic-affecting option, when it's allowed, which this test verifies.

As an aside "via API" is ambiguous, since we expose executeCommandLine now, but I've taken that to mean "only allowed where we take CompilerOptions-type objects" for now, which just means any API that's post-option-parsing/validating. Absolutely none of those APIs (eg, createProgram) are even going to emit a .buildinfo file on their own without extra instrumentation on behalf of the caller of the API, so it is quite literally up to the API user to keep things up-to-date, provided our emitBuildInfo function does serialize the option for them when they call it to do that, which, again, this test does verify.

The problem with writing this test case this way is that it is modifying timestamp of tsconfig files and hence build will take place. To check this correctly, you want to use API emit as edit as mentioned above.

Or I could allow --noCheck as a build option (why not? seems like a reasonable thing to want to turn on/off for the whole project graph) and pass it down from the top-level CLI invocation, rather than by editing the tsconfig?

sheetalkamat · 2024-04-16T18:06:18Z

src/compiler/builder.ts

+        else if (!canCopySemanticDiagnostics) {
+            // No emit being done, but semantic diagnostics must be recalculated - since these (and the options that change these)
+            // are in the build info file, we need to emit it to keep the info on-disk up-to-date
+            state.buildInfoEmitPending = true;


This should be set only if (!outputFile)

As mentioned somewhere below, you want this condition to be useOldState && !!oldState!.compilerOptions.noCheck !== !!compilerOptions.noCheck

I have no clue why this would be noCheck-specific, though? noCheck seems to me to be no different from any other compiler option that varies the errors we produce.

sheetalkamat · 2024-04-16T18:08:59Z

src/compiler/emitter.ts

@@ -848,8 +848,8 @@ export function emitFiles(resolver: EmitResolver, host: EmitHost, targetSourceFi
        const filesForEmit = forceDtsEmit ? sourceFiles : filter(sourceFiles, isSourceFileNotJson);
        // Setup and perform the transformation to retrieve declarations from the input files
        const inputListOrBundle = compilerOptions.outFile ? [factory.createBundle(filesForEmit)] : filesForEmit;
-        if (emitOnly && !getEmitDeclarations(compilerOptions)) {
-            // Checker wont collect the linked aliases since thats only done when declaration is enabled.
+        if ((emitOnly && !getEmitDeclarations(compilerOptions)) || compilerOptions.noCheck) {


Also need to add forceDtsEmit check here? shouldnt signature generation behave like noCheck?

AFAIK it already does - the || compilerOptions.noCheck being tacked on here is so that noCheck builds do the same data collection signature generation runs already do.

sheetalkamat · 2024-04-16T18:11:41Z

src/compiler/utilities.ts

@@ -9938,6 +9938,7 @@ export function skipTypeChecking(sourceFile: SourceFile, options: CompilerOption
    // '/// <reference no-default-lib="true"/>' directive.
    return (options.skipLibCheck && sourceFile.isDeclarationFile ||
        options.skipDefaultLibCheck && sourceFile.hasNoDefaultLib) ||
+        options.noCheck ||


Adding this here also skips grammar checks .. May be i am remembering it incorrectly but i thought we wanted to do grammar checks ?

Adding this here also skips grammar checks

Not... quite. More in a bit.

May be i am remembering it incorrectly but i thought we wanted to do grammar checks ?

I wanna say "no". We ruminated on it briefly, but "grammar checks" aren't a defined subset of the diagnostics we publicly recognize (and some things you'd think are in that category are emitted in the parser, binder, or checker passes over the AST - roll a d6 to find out where), so while are somewhat well-defined in ECMA-262, are incredibly poorly defined for us. Additionally, what diagnostics we skip vs emit isn't in any way what's important for this flag, what is is that we skip the checker's full tree walk (and all the work it does) for diagnostics and jump right to emit (which is what inclusion in this function does). Doing a walk specifically for grammar errors would run pretty counter to that goal, though I admit if you're in a checker context and already traversing an AST most grammar-style errors are usually trivial to emit. Again, not terribly important for this, though - we just want to skip the big walk that does all the expensive work. (Usually so it can be done asynchronously.)

sheetalkamat · 2024-04-16T18:31:39Z

...e/emitDeclarationOnly-false-on-commandline-with-declaration-and-incremental-discrepancies.js

@@ -115,127 +115,4 @@ IncrementalBuild:
    ]
  },
  "version": "FakeTSVersion"
-}
-TsBuild info text without affectedFilesPendingEmit:: /src/project2/src/tsconfig.tsbuildinfo.readable.baseline.txt::


This is incorrectly emitting buildInfo because "emitDeclarationsOnly" does not affect semantic diagnostics. The condition in builder to check for this is:

const canCopySemanticDiagnostics = useOldState && oldState!.semanticDiagnosticsPerFile && !!state.semanticDiagnosticsPerFile && !compilerOptionsAffectSemanticDiagnostics(compilerOptions, oldCompilerOptions!);

where as you want it to be compilerOptionsAffectSemanticDiagnostics(compilerOptions, oldCompilerOptions!); when emitting buildInfo again.

I think you just want to check if "noCheck" is changed and then mark the buildInfo emit pending.

sheetalkamat · 2024-04-16T18:36:48Z

src/testRunner/unittests/tsbuild/noCheck.ts

+                loadProjectFromFiles({
+                    "/src/a.ts": getATsContent(declAText),
+                    "/src/tsconfig.json": jsonToReadableText({
+                        compilerOptions: { noCheck: true, emitDeclarationOnly: true, declaration: true },


The problem with writing this test case this way is that it is modifying timestamp of tsconfig files and hence build will take place. To check this correctly, you want to use API emit as edit as mentioned above.

weswigham · 2024-04-24T18:14:44Z

Per our last design meeting discussion on this, this PR enables noCheck for transpileDeclaration by default. On this workstation, without noCheck, a simple transpileDeclaration on each file in our src directory (single thread, I/O time not included) takes 27144ms, with noCheck, that is cut to 8404ms. A substantial improvement, considering semantic diagnostics aren't even available via the transpile APIs!

weswigham · 2024-04-24T18:20:15Z

@sheetalkamat Since the option isn't available on the CLI yet, nor does it have any special behavior W.R.T. build mode compared to any other semantic diagnostics, I'm not sure there's much to do in this PR. As we said at the last design meeting we likely want to clean up how build mode handles changing flags that aren't from tsconfig files before we make this available via CLI (since a primary usecase is toggling it for a whole project graph), but that's not something that's going to happen in this PR, nor likely for 5.5. I could revert the builder.ts change here just so we can keep all that cleanup together, if you'd like.

robpalme · 2024-04-24T19:38:52Z

The new baselines show that enabling noCheck can re-order union items compared to checked declaration emit. Do we know of any other declaration emit differences?

jakebailey · 2024-04-24T20:06:25Z

tests/baselines/reference/outModuleConcatES6.js

+
+
+
+!!!! File all.d.ts missing from original emit, but present in noCheck emit


In general, how are we ending up with these messages? Are they all cases where we don't emit due to errors?

Yea, a checker error blocks the .d.ts emit without noCheck on.

jakebailey · 2024-04-24T20:07:10Z

...nes/reference/tsbuild/configFileErrors/reports-syntax-errors-in-config-file-discrepancies.js

This seems like it needs fixing?

Ah, there are other files with this (and other threads), forgot.

I've removed the small builder change in this PR - we can look into it separately. The issue there isn't related to the new noCheck flag, but rather applies to any semantic-diagnostic-affecting option you build-via-api with a different setting than you build-via-cli with, which, as we discussed, needs some more extensive builder changes to properly support (and to support passing builder-wide semantic settings via CLI generally).

weswigham · 2024-04-24T20:35:47Z

The new baselines show that enabling noCheck can re-order union items compared to checked declaration emit. Do we know of any other declaration emit differences?

Yeah, this can happen in incremental mode, too. It's a pretty well-known issue that we haven't found a great way to fix, but usually doesn't have semantic implications. Other stuff that depends on check order is regarded as a bug, since it could make diagnostics in the editor inconsistent.

weswigham · 2024-04-24T20:49:14Z

src/compiler/commandLineParser.ts

+        affectsSemanticDiagnostics: true,
+        affectsBuildInfo: true,
+        extraValidation() {
+            return [Diagnostics.Unknown_compiler_option_0, "noCheck"];


For anyone reading, this validation is what blocks this option from being passed on the CLI. This is a nicer way to do it than our usual method (just adding an entry to the CompilerOptions interface), since it still provides metadata for .buildinfo serialization and showConfig, plus makes it easier to test as though it were CLI-allowed, like is done in the noCheck unittest.

sheetalkamat · 2024-04-24T21:29:43Z

@weswigham Will review soon probably tonight or tomrorow morning, but i think we need to make sure buildInfo is not written when "noCheck" is on accidently. This can be done by changing "getTsBuildInfoEmitOutputFilePath" to return undefined if noCheck is true.

weswigham · 2024-04-24T21:40:23Z

but i think we need to make sure buildInfo is not written when "noCheck" is on accidently. This can be done by changing "getTsBuildInfoEmitOutputFilePath" to return undefined if noCheck is true.

noCheck isn't any different from any other option that affects checker diagnostics; I don't see why we'd do a one-off fix for it that doesn't apply to every other semantic-diagnostic-affecting option. noCheck isn't in any way special, other than it always adding an option parsing diagnostic when you pass it through the command line helper (which is also what any non-extant flag would do).

weswigham · 2024-04-24T21:47:32Z

For additional context, allowNonTsExtensions and suppressOutputPathCheck are other internal-only compiler option that definitely affect diagnostics, which we don't do any special handling for at all. I assume our approach for internal only options is "use at your own risk", since it's not like they can be passed through normal build/cli/config pipelines.

And noCheck shouldn't even be internal only for long. 🤷‍♂️

weswigham · 2024-04-26T20:18:12Z

Per out-of-band discussions, I've opened #58336 to formally track future work with --build and --noCheck to help clarify the long-term goals for this. As-is, I'm going to merge this, and any other changes we want to make can be taken as follow-ups.

zanminkian · 2024-05-07T13:43:49Z

Will this land in 5.5?

weswigham and others added 4 commits March 23, 2024 22:42

Add --noCheck flag, mirroring --noEmit, allowing emitting transpiling…

79961d4

…/declarations without full checking

Test .d.ts emit on noCheck (js is pretty hopelessly broken without di…

540ce28

…sentagling NodeCheckFlags and .referenced from the checkers full traversal)

Fix harness issues, js declaration emit issues with out of order chec…

3f89f5c

…king, accept baselines with extra output due to less errors, or union order changes

Add errors on using noCheck with noEmit and without emitDeclarationOn…

bf82544

…ly, refine noCheck test conditions to match

weswigham requested review from DanielRosenwasser, jakebailey and RyanCavanaugh March 25, 2024 18:41

typescript-bot assigned weswigham Mar 25, 2024

typescript-bot added Author: Team For Uncommitted Bug PR for untriaged, rejected, closed or missing bug labels Mar 25, 2024

weswigham commented Mar 25, 2024

View reviewed changes

Format

a3d4f73

weswigham added 4 commits March 27, 2024 13:22

Merge branch 'main' into nocheck

ce65c79

Update baselines with main

736eac2

Remove command-line accesible noCheck option, API only until JS is done

5c5e8e0

Merge branch 'main' into nocheck

b33afe7

weswigham changed the title ~~Add --noCheck~~ Add noCheck API option Mar 27, 2024

sheetalkamat reviewed Mar 27, 2024

View reviewed changes

Remove unused baseline

1a05d96

DanielRosenwasser mentioned this pull request Mar 28, 2024

Design Meeting Notes, 3/26/2024 #57980

Open

weswigham added 2 commits March 28, 2024 14:20

Add private option description so if build is invoked via API with …

051a1d1

…`noCheck`, it is serialized correctly

Increase comment verbosity

1231be1

PR Backlog automation moved this from Not started to Needs merge Apr 11, 2024

Merge branch 'main' into nocheck

8c667f4

sheetalkamat requested changes Apr 12, 2024

View reviewed changes

PR Backlog automation moved this from Needs merge to Waiting on author Apr 12, 2024

weswigham requested a review from sheetalkamat April 16, 2024 17:35

sheetalkamat requested changes Apr 16, 2024

View reviewed changes

weswigham mentioned this pull request Apr 19, 2024

Add transpileDeclaration API method #58261

Merged

zanminkian mentioned this pull request Apr 22, 2024

feat: support noCheck rnmjs/tscx#7

Open

weswigham added 2 commits April 24, 2024 10:37

Merge branch 'main' into nocheck

e0dfe56

Enable noCheck for transpileDeclaration

2f65540

weswigham requested a review from sheetalkamat April 24, 2024 18:15

jakebailey reviewed Apr 24, 2024

View reviewed changes

Remove builder change

0f1b4c6

weswigham commented Apr 24, 2024

View reviewed changes

This was referenced Apr 24, 2024

Lint to ensure CompilerOptions and options in commandLineParser are synchronized #58312

Open

Allow --noCheck on the CLI with top-level --build #58336

Open

weswigham enabled auto-merge (squash) April 26, 2024 20:18

weswigham disabled auto-merge April 26, 2024 20:21

weswigham merged commit f76727d into microsoft:main Apr 26, 2024
25 checks passed

PR Backlog automation moved this from Waiting on author to Done Apr 26, 2024




		!!!! File all.d.ts missing from original emit, but present in noCheck emit

Add noCheck API option #57934

Add noCheck API option #57934

Conversation

weswigham commented Mar 25, 2024 • edited

typescript-bot commented Mar 25, 2024

weswigham Mar 25, 2024 • edited

Choose a reason for hiding this comment

RyanCavanaugh commented Mar 25, 2024 • edited

weswigham commented Mar 25, 2024

sheetalkamat commented Mar 25, 2024 • edited

weswigham commented Mar 25, 2024

weswigham commented Mar 25, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

weswigham Apr 12, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

weswigham commented Apr 24, 2024

weswigham commented Apr 24, 2024

robpalme commented Apr 24, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

weswigham commented Apr 24, 2024

Choose a reason for hiding this comment

sheetalkamat commented Apr 24, 2024

weswigham commented Apr 24, 2024 • edited

weswigham commented Apr 24, 2024

weswigham commented Apr 26, 2024

zanminkian commented May 7, 2024

Add `noCheck` API option #57934

Add `noCheck` API option #57934

weswigham commented Mar 25, 2024 •

edited

weswigham Mar 25, 2024 •

edited

RyanCavanaugh commented Mar 25, 2024 •

edited

sheetalkamat commented Mar 25, 2024 •

edited

weswigham Apr 12, 2024 •

edited

weswigham commented Apr 24, 2024 •

edited