1
0
Fork 0
mirror of https://github.com/denoland/deno.git synced 2024-12-27 17:49:08 -05:00
denoland-deno/cli/lsp/testing/execution.rs

897 lines
28 KiB
Rust
Raw Normal View History

// Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.
use super::definitions::TestDefinition;
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
use super::definitions::TestModule;
use super::lsp_custom;
use super::server::TestServerTests;
use crate::args::flags_from_vec;
use crate::args::DenoSubcommand;
use crate::factory::CliFactory;
use crate::lsp::client::Client;
use crate::lsp::client::TestingNotification;
use crate::lsp::config;
use crate::lsp::logging::lsp_log;
use crate::lsp::urls::uri_parse_unencoded;
use crate::lsp::urls::uri_to_url;
use crate::lsp::urls::url_to_uri;
use crate::tools::test;
use crate::tools::test::create_test_event_channel;
use crate::tools::test::FailFastTracker;
use crate::tools::test::TestFailureFormatOptions;
use deno_core::anyhow::anyhow;
use deno_core::error::AnyError;
use deno_core::error::JsError;
use deno_core::futures::future;
use deno_core::futures::stream;
use deno_core::futures::StreamExt;
use deno_core::parking_lot::RwLock;
use deno_core::unsync::spawn;
use deno_core::unsync::spawn_blocking;
use deno_core::ModuleSpecifier;
use deno_runtime::deno_permissions::Permissions;
use deno_runtime::deno_permissions::PermissionsContainer;
use deno_runtime::tokio_util::create_and_run_current_thread;
use indexmap::IndexMap;
use std::borrow::Cow;
use std::collections::HashMap;
use std::collections::HashSet;
use std::num::NonZeroUsize;
use std::sync::Arc;
use std::time::Duration;
use std::time::Instant;
use tokio_util::sync::CancellationToken;
use tower_lsp::lsp_types as lsp;
/// Logic to convert a test request into a set of test modules to be tested and
/// any filters to be applied to those tests
fn as_queue_and_filters(
params: &lsp_custom::TestRunRequestParams,
tests: &HashMap<ModuleSpecifier, (TestModule, String)>,
) -> (
HashSet<ModuleSpecifier>,
HashMap<ModuleSpecifier, LspTestFilter>,
) {
let mut queue: HashSet<ModuleSpecifier> = HashSet::new();
let mut filters: HashMap<ModuleSpecifier, LspTestFilter> = HashMap::new();
if let Some(include) = &params.include {
for item in include {
let url = uri_to_url(&item.text_document.uri);
if let Some((test_definitions, _)) = tests.get(&url) {
queue.insert(url.clone());
if let Some(id) = &item.id {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
if let Some(test) = test_definitions.get(id) {
let filter = filters.entry(url).or_default();
if let Some(include) = filter.include.as_mut() {
include.insert(test.id.clone(), test.clone());
} else {
let mut include = HashMap::new();
include.insert(test.id.clone(), test.clone());
filter.include = Some(include);
}
}
}
}
}
} else {
queue.extend(tests.keys().cloned());
}
for item in &params.exclude {
let url = uri_to_url(&item.text_document.uri);
if let Some((test_definitions, _)) = tests.get(&url) {
if let Some(id) = &item.id {
// there is no way to exclude a test step
if item.step_id.is_none() {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
if let Some(test) = test_definitions.get(id) {
let filter = filters.entry(url.clone()).or_default();
filter.exclude.insert(test.id.clone(), test.clone());
}
}
} else {
// the entire test module is excluded
queue.remove(&url);
}
}
}
queue.retain(|s| !tests.get(s).unwrap().0.is_empty());
(queue, filters)
}
fn as_test_messages<S: AsRef<str>>(
message: S,
is_markdown: bool,
) -> Vec<lsp_custom::TestMessage> {
let message = lsp::MarkupContent {
kind: if is_markdown {
lsp::MarkupKind::Markdown
} else {
lsp::MarkupKind::PlainText
},
value: message.as_ref().to_string(),
};
vec![lsp_custom::TestMessage {
message,
expected_output: None,
actual_output: None,
location: None,
}]
}
#[derive(Debug, Clone, Default, PartialEq)]
struct LspTestFilter {
include: Option<HashMap<String, TestDefinition>>,
exclude: HashMap<String, TestDefinition>,
}
impl LspTestFilter {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
fn as_ids(&self, test_module: &TestModule) -> Vec<String> {
let ids: Vec<String> = if let Some(include) = &self.include {
include.keys().cloned().collect()
} else {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test_module
.defs
.iter()
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
.filter(|(_, d)| d.parent_id.is_none())
.map(|(k, _)| k.clone())
.collect()
};
ids
.into_iter()
.filter(|id| !self.exclude.contains_key(id))
.collect()
}
}
#[derive(Debug, Clone)]
pub struct TestRun {
id: u32,
kind: lsp_custom::TestRunKind,
filters: HashMap<ModuleSpecifier, LspTestFilter>,
queue: HashSet<ModuleSpecifier>,
tests: TestServerTests,
token: CancellationToken,
workspace_settings: config::WorkspaceSettings,
}
impl TestRun {
pub async fn init(
params: &lsp_custom::TestRunRequestParams,
tests: TestServerTests,
workspace_settings: config::WorkspaceSettings,
) -> Self {
let (queue, filters) = {
let tests = tests.lock().await;
as_queue_and_filters(params, &tests)
};
Self {
id: params.id,
kind: params.kind.clone(),
filters,
queue,
tests,
token: CancellationToken::new(),
workspace_settings,
}
}
/// Provide the tests of a test run as an enqueued module which can be sent
/// to the client to indicate tests are enqueued for testing.
pub async fn as_enqueued(&self) -> Vec<lsp_custom::EnqueuedTestModule> {
let tests = self.tests.lock().await;
self
.queue
.iter()
.filter_map(|s| {
let ids = if let Some((test_module, _)) = tests.get(s) {
if let Some(filter) = self.filters.get(s) {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
filter.as_ids(test_module)
} else {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
LspTestFilter::default().as_ids(test_module)
}
} else {
Vec::new()
};
Some(lsp_custom::EnqueuedTestModule {
text_document: lsp::TextDocumentIdentifier {
uri: url_to_uri(s).ok()?,
},
ids,
})
})
.collect()
}
/// If being executed, cancel the test.
pub fn cancel(&self) {
self.token.cancel();
}
/// Execute the tests, dispatching progress notifications to the client.
pub async fn exec(
&self,
client: &Client,
maybe_root_uri: Option<&ModuleSpecifier>,
) -> Result<(), AnyError> {
let args = self.get_args();
lsp_log!("Executing test run with arguments: {}", args.join(" "));
let flags = Arc::new(flags_from_vec(
args.into_iter().map(|s| From::from(s.as_ref())).collect(),
)?);
let factory = CliFactory::from_flags(flags);
let cli_options = factory.cli_options()?;
// Various test files should not share the same permissions in terms of
// `PermissionsContainer` - otherwise granting/revoking permissions in one
// file would have impact on other files, which is undesirable.
let permission_desc_parser = factory.permission_desc_parser()?.clone();
let permissions = Permissions::from_options(
permission_desc_parser.as_ref(),
&cli_options.permissions_options(),
)?;
let main_graph_container = factory.main_module_graph_container().await?;
feat(cli): evaluate code snippets in JSDoc and markdown (#25220) This commit lets `deno test --doc` command actually evaluate code snippets in JSDoc and markdown files. ## How it works 1. Extract code snippets from JSDoc or code fences 2. Convert them into pseudo files by wrapping them in `Deno.test(...)` 3. Register the pseudo files as in-memory files 4. Run type-check and evaluation We apply some magic at the step 2 - let's say we have the following file named `mod.ts` as an input: ````ts /** * ```ts * import { assertEquals } from "jsr:@std/assert/equals"; * * assertEquals(add(1, 2), 3); * ``` */ export function add(a: number, b: number) { return a + b; } ```` This is virtually transformed into: ```ts import { assertEquals } from "jsr:@std/assert/equals"; import { add } from "files:///path/to/mod.ts"; Deno.test("mod.ts$2-7.ts", async () => { assertEquals(add(1, 2), 3); }); ``` Note that a new import statement is inserted here to make `add` function available. In a nutshell, all items exported from `mod.ts` become available in the generated pseudo file with this automatic import insertion. The intention behind this design is that, from library user's standpoint, it should be very obvious that this `add` function is what this example code is attached to. Also, if there is an explicit import statement like `import { add } from "./mod.ts"`, this import path `./mod.ts` is not helpful for doc readers because they will need to import it in a different way. The automatic import insertion has some edge cases, in particular where there is a local variable in a snippet with the same name as one of the exported items. This case is addressed by employing swc's scope analysis (see test cases for more details). ## "type-checking only" mode stays around This change will likely impact a lot of existing doc tests in the ecosystem because some doc tests rely on the fact that they are not evaluated - some cause side effects if executed, some throw errors at runtime although they do pass the type check, etc. To help those tests gradually transition to the ones runnable with the new `deno test --doc`, we will keep providing the ability to run type-checking only via `deno check --doc`. Additionally there is a `--doc-only` option added to the `check` subcommand too, which is useful when you want to type-check on code snippets in markdown files, as normal `deno check` command doesn't accept markdown. ## Demo https://github.com/user-attachments/assets/47e9af73-d16e-472d-b09e-1853b9e8f5ce --- Closes #4716
2024-09-18 00:35:48 -04:00
main_graph_container
.check_specifiers(&self.queue.iter().cloned().collect::<Vec<_>>(), None)
feat(cli): evaluate code snippets in JSDoc and markdown (#25220) This commit lets `deno test --doc` command actually evaluate code snippets in JSDoc and markdown files. ## How it works 1. Extract code snippets from JSDoc or code fences 2. Convert them into pseudo files by wrapping them in `Deno.test(...)` 3. Register the pseudo files as in-memory files 4. Run type-check and evaluation We apply some magic at the step 2 - let's say we have the following file named `mod.ts` as an input: ````ts /** * ```ts * import { assertEquals } from "jsr:@std/assert/equals"; * * assertEquals(add(1, 2), 3); * ``` */ export function add(a: number, b: number) { return a + b; } ```` This is virtually transformed into: ```ts import { assertEquals } from "jsr:@std/assert/equals"; import { add } from "files:///path/to/mod.ts"; Deno.test("mod.ts$2-7.ts", async () => { assertEquals(add(1, 2), 3); }); ``` Note that a new import statement is inserted here to make `add` function available. In a nutshell, all items exported from `mod.ts` become available in the generated pseudo file with this automatic import insertion. The intention behind this design is that, from library user's standpoint, it should be very obvious that this `add` function is what this example code is attached to. Also, if there is an explicit import statement like `import { add } from "./mod.ts"`, this import path `./mod.ts` is not helpful for doc readers because they will need to import it in a different way. The automatic import insertion has some edge cases, in particular where there is a local variable in a snippet with the same name as one of the exported items. This case is addressed by employing swc's scope analysis (see test cases for more details). ## "type-checking only" mode stays around This change will likely impact a lot of existing doc tests in the ecosystem because some doc tests rely on the fact that they are not evaluated - some cause side effects if executed, some throw errors at runtime although they do pass the type check, etc. To help those tests gradually transition to the ones runnable with the new `deno test --doc`, we will keep providing the ability to run type-checking only via `deno check --doc`. Additionally there is a `--doc-only` option added to the `check` subcommand too, which is useful when you want to type-check on code snippets in markdown files, as normal `deno check` command doesn't accept markdown. ## Demo https://github.com/user-attachments/assets/47e9af73-d16e-472d-b09e-1853b9e8f5ce --- Closes #4716
2024-09-18 00:35:48 -04:00
.await?;
let (concurrent_jobs, fail_fast) =
if let DenoSubcommand::Test(test_flags) = cli_options.sub_command() {
(
test_flags
.concurrent_jobs
.unwrap_or_else(|| NonZeroUsize::new(1).unwrap())
.into(),
test_flags.fail_fast,
)
} else {
unreachable!("Should always be Test subcommand.");
};
// TODO(mmastrac): Temporarily limit concurrency in windows testing to avoid named pipe issue:
// *** Unexpected server pipe failure '"\\\\.\\pipe\\deno_pipe_e30f45c9df61b1e4.1198.222\\0"': 3
// This is likely because we're hitting some sort of invisible resource limit
// This limit is both in cli/lsp/testing/execution.rs and cli/tools/test/mod.rs
#[cfg(windows)]
let concurrent_jobs = std::cmp::min(concurrent_jobs, 4);
let (test_event_sender_factory, mut receiver) = create_test_event_channel();
let fail_fast_tracker = FailFastTracker::new(fail_fast);
let mut queue = self.queue.iter().collect::<Vec<&ModuleSpecifier>>();
queue.sort();
let tests: Arc<RwLock<IndexMap<usize, test::TestDescription>>> =
Arc::new(RwLock::new(IndexMap::new()));
let mut test_steps = IndexMap::new();
let worker_factory =
Arc::new(factory.create_cli_main_worker_factory().await?);
let join_handles = queue.into_iter().map(move |specifier| {
let specifier = specifier.clone();
let worker_factory = worker_factory.clone();
let permissions_container = PermissionsContainer::new(
permission_desc_parser.clone(),
permissions.clone(),
);
let worker_sender = test_event_sender_factory.worker();
let fail_fast_tracker = fail_fast_tracker.clone();
let lsp_filter = self.filters.get(&specifier);
let filter = test::TestFilter {
substring: None,
regex: None,
include: lsp_filter.and_then(|f| {
f.include
.as_ref()
.map(|i| i.values().map(|t| t.name.clone()).collect())
}),
exclude: lsp_filter
.map(|f| f.exclude.values().map(|t| t.name.clone()).collect())
.unwrap_or_default(),
};
let token = self.token.clone();
spawn_blocking(move || {
if fail_fast_tracker.should_stop() {
return Ok(());
}
if token.is_cancelled() {
Ok(())
} else {
// All JsErrors are handled by test_specifier and piped into the test
// channel.
create_and_run_current_thread(test::test_specifier(
worker_factory,
permissions_container,
specifier,
worker_sender,
fail_fast_tracker,
test::TestSpecifierOptions {
filter,
shuffle: None,
trace_leaks: false,
},
))
}
})
});
let join_stream = stream::iter(join_handles)
.buffer_unordered(concurrent_jobs)
.collect::<Vec<Result<Result<(), AnyError>, tokio::task::JoinError>>>();
let mut reporter = Box::new(LspTestReporter::new(
self,
client.clone(),
maybe_root_uri,
self.tests.clone(),
));
let handler = {
spawn(async move {
let earlier = Instant::now();
let mut summary = test::TestSummary::new();
let mut tests_with_result = HashSet::new();
let mut used_only = false;
while let Some((_, event)) = receiver.recv().await {
match event {
test::TestEvent::Register(description) => {
for (_, description) in description.into_iter() {
reporter.report_register(description).await;
// TODO(mmastrac): we shouldn't need to clone here - we can re-use the descriptions
tests.write().insert(description.id, description.clone());
}
}
test::TestEvent::Plan(plan) => {
summary.total += plan.total;
summary.filtered_out += plan.filtered_out;
if plan.used_only {
used_only = true;
}
reporter.report_plan(&plan);
}
test::TestEvent::Wait(id) => {
reporter.report_wait(tests.read().get(&id).unwrap());
}
2024-06-14 07:40:57 -04:00
test::TestEvent::Output(output) => {
reporter.report_output(&output);
}
test::TestEvent::Slow(id, elapsed) => {
reporter.report_slow(tests.read().get(&id).unwrap(), elapsed);
}
test::TestEvent::Result(id, result, elapsed) => {
if tests_with_result.insert(id) {
let description = tests.read().get(&id).unwrap().clone();
match &result {
test::TestResult::Ok => summary.passed += 1,
test::TestResult::Ignored => summary.ignored += 1,
test::TestResult::Failed(error) => {
summary.failed += 1;
summary
.failures
.push(((&description).into(), error.clone()));
}
test::TestResult::Cancelled => {
summary.failed += 1;
}
}
reporter.report_result(&description, &result, elapsed);
}
}
test::TestEvent::UncaughtError(origin, error) => {
reporter.report_uncaught_error(&origin, &error);
summary.failed += 1;
summary.uncaught_errors.push((origin, error));
}
test::TestEvent::StepRegister(description) => {
reporter.report_step_register(&description).await;
test_steps.insert(description.id, description);
}
test::TestEvent::StepWait(id) => {
reporter.report_step_wait(test_steps.get(&id).unwrap());
}
test::TestEvent::StepResult(id, result, duration) => {
if tests_with_result.insert(id) {
match &result {
test::TestStepResult::Ok => {
summary.passed_steps += 1;
}
test::TestStepResult::Ignored => {
summary.ignored_steps += 1;
}
test::TestStepResult::Failed(_) => {
summary.failed_steps += 1;
}
}
reporter.report_step_result(
test_steps.get(&id).unwrap(),
&result,
duration,
);
}
}
test::TestEvent::Completed => {
reporter.report_completed();
}
test::TestEvent::ForceEndReport => {}
test::TestEvent::Sigint => {}
}
}
let elapsed = Instant::now().duration_since(earlier);
reporter.report_summary(&summary, &elapsed);
if used_only {
return Err(anyhow!(
"Test failed because the \"only\" option was used"
));
}
if summary.failed > 0 {
return Err(anyhow!("Test failed"));
}
Ok(())
})
};
let (join_results, result) = future::join(join_stream, handler).await;
// propagate any errors
for join_result in join_results {
join_result??;
}
result??;
Ok(())
}
fn get_args(&self) -> Vec<Cow<str>> {
let mut args = vec![Cow::Borrowed("deno"), Cow::Borrowed("test")];
args.extend(
self
.workspace_settings
.testing
.args
.iter()
.map(|s| Cow::Borrowed(s.as_str())),
);
args.push(Cow::Borrowed("--trace-leaks"));
for unstable_feature in self.workspace_settings.unstable.as_deref() {
let flag = format!("--unstable-{unstable_feature}");
if !args.contains(&Cow::Borrowed(&flag)) {
args.push(Cow::Owned(flag));
}
}
if let Some(config) = &self.workspace_settings.config {
if !args.contains(&Cow::Borrowed("--config"))
&& !args.contains(&Cow::Borrowed("-c"))
{
args.push(Cow::Borrowed("--config"));
args.push(Cow::Borrowed(config.as_str()));
}
}
if let Some(import_map) = &self.workspace_settings.import_map {
if !args.contains(&Cow::Borrowed("--import-map")) {
args.push(Cow::Borrowed("--import-map"));
args.push(Cow::Borrowed(import_map.as_str()));
}
}
if self.kind == lsp_custom::TestRunKind::Debug
&& !args.contains(&Cow::Borrowed("--inspect"))
&& !args.contains(&Cow::Borrowed("--inspect-brk"))
{
args.push(Cow::Borrowed("--inspect"));
}
args
}
}
#[derive(Debug, PartialEq)]
enum LspTestDescription {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
/// `(desc, static_id)`
TestDescription(test::TestDescription, String),
/// `(desc, static_id)`
TestStepDescription(test::TestStepDescription, String),
}
impl LspTestDescription {
fn origin(&self) -> &str {
match self {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
LspTestDescription::TestDescription(d, _) => d.origin.as_str(),
LspTestDescription::TestStepDescription(d, _) => d.origin.as_str(),
}
}
fn location(&self) -> &test::TestLocation {
match self {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
LspTestDescription::TestDescription(d, _) => &d.location,
LspTestDescription::TestStepDescription(d, _) => &d.location,
}
}
fn parent_id(&self) -> Option<usize> {
match self {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
LspTestDescription::TestDescription(_, _) => None,
LspTestDescription::TestStepDescription(d, _) => Some(d.parent_id),
}
}
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
fn static_id(&self) -> &str {
match self {
LspTestDescription::TestDescription(_, i) => i,
LspTestDescription::TestStepDescription(_, i) => i,
}
}
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
fn as_test_identifier(
&self,
tests: &IndexMap<usize, LspTestDescription>,
) -> lsp_custom::TestIdentifier {
let uri = uri_parse_unencoded(&self.location().file_name).unwrap();
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let static_id = self.static_id();
let mut root_desc = self;
while let Some(parent_id) = root_desc.parent_id() {
root_desc = tests.get(&parent_id).unwrap();
}
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let root_static_id = root_desc.static_id();
lsp_custom::TestIdentifier {
text_document: lsp::TextDocumentIdentifier { uri },
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
id: Some(root_static_id.to_string()),
step_id: if static_id == root_static_id {
None
} else {
Some(static_id.to_string())
},
}
}
}
struct LspTestReporter {
client: Client,
id: u32,
maybe_root_uri: Option<ModuleSpecifier>,
files: TestServerTests,
tests: IndexMap<usize, LspTestDescription>,
current_test: Option<usize>,
}
impl LspTestReporter {
fn new(
run: &TestRun,
client: Client,
maybe_root_uri: Option<&ModuleSpecifier>,
files: TestServerTests,
) -> Self {
Self {
client,
id: run.id,
maybe_root_uri: maybe_root_uri.cloned(),
files,
tests: Default::default(),
current_test: Default::default(),
}
}
fn progress(&self, message: lsp_custom::TestRunProgressMessage) {
self
.client
.send_test_notification(TestingNotification::Progress(
lsp_custom::TestRunProgressParams {
id: self.id,
message,
},
));
}
fn report_plan(&mut self, _plan: &test::TestPlan) {}
async fn report_register(&mut self, desc: &test::TestDescription) {
let mut files = self.files.lock().await;
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let specifier = ModuleSpecifier::parse(&desc.location.file_name).unwrap();
let (test_module, _) = files
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
.entry(specifier.clone())
.or_insert_with(|| (TestModule::new(specifier), "1".to_string()));
let Ok(uri) = url_to_uri(&test_module.specifier) else {
return;
};
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let (static_id, is_new) = test_module.register_dynamic(desc);
self.tests.insert(
desc.id,
LspTestDescription::TestDescription(desc.clone(), static_id.clone()),
);
if is_new {
self
.client
.send_test_notification(TestingNotification::Module(
lsp_custom::TestModuleNotificationParams {
text_document: lsp::TextDocumentIdentifier { uri },
kind: lsp_custom::TestModuleNotificationKind::Insert,
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
label: test_module.label(self.maybe_root_uri.as_ref()),
tests: vec![test_module.get_test_data(&static_id)],
},
));
}
}
fn report_wait(&mut self, desc: &test::TestDescription) {
self.current_test = Some(desc.id);
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
let test = desc.as_test_identifier(&self.tests);
self.progress(lsp_custom::TestRunProgressMessage::Started { test });
}
fn report_slow(&mut self, _desc: &test::TestDescription, _elapsed: u64) {}
fn report_output(&mut self, output: &[u8]) {
let test = self
.current_test
.as_ref()
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
.map(|id| self.tests.get(id).unwrap().as_test_identifier(&self.tests));
let value = String::from_utf8_lossy(output).replace('\n', "\r\n");
self.progress(lsp_custom::TestRunProgressMessage::Output {
value,
test,
// TODO(@kitsonk) test output should include a location
location: None,
})
}
fn report_result(
&mut self,
desc: &test::TestDescription,
result: &test::TestResult,
elapsed: u64,
) {
self.current_test = None;
match result {
test::TestResult::Ok => {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
self.progress(lsp_custom::TestRunProgressMessage::Passed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
duration: Some(elapsed as u32),
})
}
test::TestResult::Ignored => {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
self.progress(lsp_custom::TestRunProgressMessage::Skipped {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
})
}
test::TestResult::Failed(failure) => {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
self.progress(lsp_custom::TestRunProgressMessage::Failed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
messages: as_test_messages(
failure.format(&TestFailureFormatOptions::default()),
false,
),
duration: Some(elapsed as u32),
})
}
test::TestResult::Cancelled => {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
self.progress(lsp_custom::TestRunProgressMessage::Failed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
messages: vec![],
duration: Some(elapsed as u32),
})
}
}
}
fn report_uncaught_error(&mut self, origin: &str, js_error: &JsError) {
self.current_test = None;
let err_string = format!(
"Uncaught error from {}: {}\nThis error was not caught from a test and caused the test runner to fail on the referenced module.\nIt most likely originated from a dangling promise, event/timeout handler or top-level code.",
origin,
test::fmt::format_test_error(js_error, &TestFailureFormatOptions::default())
);
let messages = as_test_messages(err_string, false);
for desc in self.tests.values().filter(|d| d.origin() == origin) {
self.progress(lsp_custom::TestRunProgressMessage::Failed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
messages: messages.clone(),
duration: None,
});
}
}
async fn report_step_register(&mut self, desc: &test::TestStepDescription) {
let mut files = self.files.lock().await;
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let specifier = ModuleSpecifier::parse(&desc.location.file_name).unwrap();
let (test_module, _) = files
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
.entry(specifier.clone())
.or_insert_with(|| (TestModule::new(specifier), "1".to_string()));
let Ok(uri) = url_to_uri(&test_module.specifier) else {
return;
};
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let (static_id, is_new) = test_module.register_step_dynamic(
desc,
self.tests.get(&desc.parent_id).unwrap().static_id(),
);
self.tests.insert(
desc.id,
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
LspTestDescription::TestStepDescription(desc.clone(), static_id.clone()),
);
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
if is_new {
self
.client
.send_test_notification(TestingNotification::Module(
lsp_custom::TestModuleNotificationParams {
text_document: lsp::TextDocumentIdentifier { uri },
kind: lsp_custom::TestModuleNotificationKind::Insert,
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
label: test_module.label(self.maybe_root_uri.as_ref()),
tests: vec![test_module.get_test_data(&static_id)],
},
));
}
}
fn report_step_wait(&mut self, desc: &test::TestStepDescription) {
if self.current_test == Some(desc.parent_id) {
self.current_test = Some(desc.id);
}
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
let test = desc.as_test_identifier(&self.tests);
self.progress(lsp_custom::TestRunProgressMessage::Started { test });
}
fn report_step_result(
&mut self,
desc: &test::TestStepDescription,
result: &test::TestStepResult,
elapsed: u64,
) {
if self.current_test == Some(desc.id) {
self.current_test = Some(desc.parent_id);
}
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let desc = self.tests.get(&desc.id).unwrap();
match result {
test::TestStepResult::Ok => {
self.progress(lsp_custom::TestRunProgressMessage::Passed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
duration: Some(elapsed as u32),
})
}
test::TestStepResult::Ignored => {
self.progress(lsp_custom::TestRunProgressMessage::Skipped {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
})
}
test::TestStepResult::Failed(failure) => {
self.progress(lsp_custom::TestRunProgressMessage::Failed {
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
test: desc.as_test_identifier(&self.tests),
messages: as_test_messages(
failure.format(&TestFailureFormatOptions::default()),
false,
),
duration: Some(elapsed as u32),
})
}
}
}
fn report_completed(&mut self) {
// there is nothing to do on report_completed
}
fn report_summary(
&mut self,
_summary: &test::TestSummary,
_elapsed: &Duration,
) {
// there is nothing to do on report_summary
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::lsp::testing::collectors::tests::new_range;
use deno_core::serde_json::json;
#[test]
fn test_as_queue_and_filters() {
let specifier = ModuleSpecifier::parse("file:///a/file.ts").unwrap();
// Regression test for https://github.com/denoland/vscode_deno/issues/890.
let non_test_specifier =
ModuleSpecifier::parse("file:///a/no_tests.ts").unwrap();
let params = lsp_custom::TestRunRequestParams {
id: 1,
kind: lsp_custom::TestRunKind::Run,
include: Some(vec![
lsp_custom::TestIdentifier {
text_document: lsp::TextDocumentIdentifier {
uri: url_to_uri(&specifier).unwrap(),
},
id: None,
step_id: None,
},
lsp_custom::TestIdentifier {
text_document: lsp::TextDocumentIdentifier {
uri: url_to_uri(&non_test_specifier).unwrap(),
},
id: None,
step_id: None,
},
]),
exclude: vec![lsp_custom::TestIdentifier {
text_document: lsp::TextDocumentIdentifier {
uri: url_to_uri(&specifier).unwrap(),
},
id: Some(
"69d9fe87f64f5b66cb8b631d4fd2064e8224b8715a049be54276c42189ff8f9f"
.to_string(),
),
step_id: None,
}],
};
let mut tests = HashMap::new();
let test_def_a = TestDefinition {
id: "0b7c6bf3cd617018d33a1bf982a08fe088c5bb54fcd5eb9e802e7c137ec1af94"
.to_string(),
name: "test a".to_string(),
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
range: Some(new_range(1, 5, 1, 9)),
is_dynamic: false,
parent_id: None,
step_ids: Default::default(),
};
let test_def_b = TestDefinition {
id: "69d9fe87f64f5b66cb8b631d4fd2064e8224b8715a049be54276c42189ff8f9f"
.to_string(),
name: "test b".to_string(),
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
range: Some(new_range(2, 5, 2, 9)),
is_dynamic: false,
parent_id: None,
step_ids: Default::default(),
};
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
let test_module = TestModule {
specifier: specifier.clone(),
defs: vec![
(test_def_a.id.clone(), test_def_a.clone()),
(test_def_b.id.clone(), test_def_b.clone()),
]
.into_iter()
.collect(),
};
tests.insert(specifier.clone(), (test_module.clone(), "1".to_string()));
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
tests.insert(
non_test_specifier.clone(),
(TestModule::new(non_test_specifier), "1".to_string()),
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
);
let (queue, filters) = as_queue_and_filters(&params, &tests);
assert_eq!(json!(queue), json!([specifier]));
let mut exclude = HashMap::new();
exclude.insert(
"69d9fe87f64f5b66cb8b631d4fd2064e8224b8715a049be54276c42189ff8f9f"
.to_string(),
test_def_b,
);
let maybe_filter = filters.get(&specifier);
assert!(maybe_filter.is_some());
let filter = maybe_filter.unwrap();
assert_eq!(
filter,
&LspTestFilter {
include: None,
exclude,
}
);
assert_eq!(
refactor(lsp): store test definitions in adjacency list (#20330) Previously: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: SourceRange, pub steps: Vec<TestDefinition>, } pub struct TestDefinitions { pub discovered: Vec<TestDefinition>, pub injected: Vec<lsp_custom::TestData>, pub script_version: String, } ``` Now: ```rust pub struct TestDefinition { pub id: String, pub name: String, pub range: Option<Range>, pub is_dynamic: bool, // True for 'injected' module, not statically detected but added at runtime. pub parent_id: Option<String>, pub step_ids: HashSet<String>, } pub struct TestModule { pub specifier: ModuleSpecifier, pub script_version: String, pub defs: HashMap<String, TestDefinition>, } ``` Storing the test tree as a literal tree diminishes the value of IDs, even though vscode stores them that way. This makes all data easily accessible from `TestModule`. It unifies the interface between 'discovered' and 'injected' tests. This unblocks some enhancements wrt syncing tests between the LSP and extension, such as this TODO: https://github.com/denoland/vscode_deno/blob/61f08d5a71536a0a5f7dce965955b09e6bd957e1/client/src/testing.ts#L251-L259 and https://github.com/denoland/vscode_deno/issues/900. We should also get more flexibility overall. `TestCollector` is cleaned up, now stores a `&mut TestModule` directly and registers tests as it comes across them with `TestModule::register()`. This method ensures sanity in the redundant data from having both of `TestDefinition::{parent_id,step_ids}`. All of the messy conversions between `TestDescription`, `LspTestDescription`, `TestDefinition`, `TestData` and `TestIdentifier` are cleaned up. They shouldn't have been using `impl From` and now the full list of tests is available to their implementations.
2023-08-30 11:31:31 -04:00
filter.as_ids(&test_module),
vec![
"0b7c6bf3cd617018d33a1bf982a08fe088c5bb54fcd5eb9e802e7c137ec1af94"
.to_string()
]
);
}
}