1
0
Fork 0
mirror of https://github.com/denoland/deno.git synced 2024-12-04 17:18:23 -05:00
Commit graph

106 commits

Author SHA1 Message Date
David Sherret
2dfc0aca7c
fix(publish): make include and exclude work (#22720)
1. Stops `deno publish` using some custom include/exclude behaviour from
other sub commands
2. Takes ancestor directories into account when resolving gitignore
3. Backards compatible change that adds ability to unexclude an exclude
by using a negated glob at a more specific level for all sub commands
(see https://github.com/denoland/deno_config/pull/44).
2024-03-07 20:16:32 -05:00
Matt Mastracci
0fdb33c3aa
fix(cli): limit test parallelism on Windows to avoid pipe error (#22776)
One last attempt to fix the parallelism issue on Windows.
2024-03-07 10:04:39 -07:00
David Sherret
594d8397ad
fix(publish): properly display graph validation errors (#22775)
The graph validation errors were displaying cryptically during publish.
This fixes that.
2024-03-07 17:30:30 +01:00
Matt Mastracci
72d34a79ac
fix(cli): remove possible deadlock in test channel (#22662)
The stderr stream could possibly starve the other bits of the
output-redirecting event loop.

Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
2024-03-04 20:14:21 -07:00
cui fliter
d4b3b39cc0
chore: fix typos (#22677) 2024-03-04 14:28:57 +00:00
Matt Mastracci
2e4a1fc3e8
fix(cli): force flush output after test unloads (#22660)
Fixes flake in load_unload test
2024-03-01 21:26:57 +00:00
Matt Mastracci
736b91edd0
perf(cli): use new deno_core timers (#22569)
Improves #19100 

Fixes #20356

Replaces #20428



Changes made in deno_core to support this:

 - [x] Errors must be handled in setTimeout callbacks
 - [x] Microtask ordering is not-quite-right
 - [x] Timer cancellation must be checked right before dispatch
 - [x] Timer sanitizer
 - [x] Move high-res timer to deno_core
 - [x] Timers need opcall tracing
2024-03-01 11:15:18 -07:00
Bartek Iwańczuk
ea3651a1ac
fix(publish): disable provenance if not in GHA (#22638)
The condition was wrong
2024-02-29 21:19:30 +00:00
Matt Mastracci
b6e44f91ad
fix(cli): ensure that pre- and post-test output is flushed at the appropriate times (#22611)
Some `deno_std` tests were failing to print output that was resolved
after the last test finished. In addition, output printed before tests
began would sometimes appear above the "running X tests ..." line, and
sometimes below it depending on timing.

We now guarantee that all output is flushed before and after tests run,
making the output consistent.

Pre-test and post-test output are captured in `------ pre-test output
------` and `------ post-test output ------` blocks to differentiate
them from the regular output blocks.

Here's an example of a test (that is much noisier than normal, but an
example of what the output will look like):

```
Check ./load_unload.ts
------- pre-test output -------
load
----- output end -----
running 1 test from ./load_unload.ts
test ...
------- output -------
test
----- output end -----
test ... ok ([WILDCARD])
------- post-test output -------
unload
----- output end -----
```
2024-02-28 22:12:21 +00:00
Matt Mastracci
a3f982c1d5
chore(cli): rename --trace-ops to --trace-leaks (#22598)
As we add tracing to more types of runtime activity, `--trace-ops` is
less useful of a name. `--trace-leaks` better reflects that this feature
traces both ops and timers, and will eventually trace resource opening
as well.

This keeps `--trace-ops` as an alias for `--trace-leaks`, but prints a
warning to the console suggesting migration to `--trace-leaks`.

One test continues to use `--trace-ops` to test the deprecation warning.

---------

Signed-off-by: Matt Mastracci <matthew@mastracci.com>
2024-02-28 09:12:43 -07:00
Matt Mastracci
96cfe82664
perf(cli): reduce overhead in test registration (#22552)
- Removes the origin call, since all origins are the same for an isolate
(ie: the main module)
- Collects the `TestDescription`s and sends them all at the same time
inside of an Arc, allowing us to (later on) re-use these instead of
cloning.

Needs a follow-up pass to remove all the cloning, but that's a thread
that is pretty long to pull

---------

Signed-off-by: Matt Mastracci <matthew@mastracci.com>
2024-02-27 20:30:17 -07:00
Matt Mastracci
47c2a63d87
chore: bump deno_core (#22596)
Migrations:

 - snapshot code updated
 - runtime stats API tweaks
2024-02-27 08:05:57 -07:00
Matt Mastracci
5193834cf2
refactor(cli): clean up test runner channels (#22422)
Gets us closer to solving #20707.

Rewrites the `TestEventSender`:

- Allow for explicit creation of multiple streams. This will allow for
one-std{out,err}-per-worker
- All test events are received along with a worker ID, allowing for
eventual, proper parallel threading of test events.
 
In theory this should open up proper interleaving of test output,
however that is left for a future PR.

I had some plans for a better performing synchronization primitive, but
the inter-thread communication is tricky. This does, however, speed up
the processing of large numbers of tests 15-25% (possibly even more on
100,000+).

Before

```
ok | 1000 passed | 0 failed (32ms)
ok | 10000 passed | 0 failed (276ms)
```

After

```
ok | 1000 passed | 0 failed (25ms)
ok | 10000 passed | 0 failed (230ms)
```
2024-02-23 11:11:15 -07:00
David Sherret
f90889e5ee
perf(jsr): fast check cache and lazy fast check graph (#22485) 2024-02-20 21:29:57 +00:00
Asher Gomez
eb542bc185
chore(fs): rename op_fs_file_stat_{async/sync} ops (#22476)
Renames `op_fs_fstat_{sync/async}` to `op_fs_file_stat_{async/sync}` in
preparation of the upcoming removal of `Deno.fstat()` in v2.
2024-02-19 06:33:42 -07:00
Matt Mastracci
67a4231bb6
refactor(cli): move op sanitizer to Rust (#22413)
The format of the sanitizers will change a little bit:

- If multiple async ops leak and traces are on, we repeat the async op
header once per stack trace.
- All leaks are aggregated under a "Leaks detected:" banner as the new
timers are eventually going to be added, and these are neither ops nor
resources.
 - `1 async op` is now `An async op`
- If ops and resources leak, we show both (rather than op leaks masking
resources)

Follow-on to https://github.com/denoland/deno/pull/22226
2024-02-16 21:22:12 +00:00
Matt Mastracci
f705906256
chore(ext/web): refactor timer ops before landing op sanitizer (#22435)
Splitting the sleep and interval ops allows us to detect an interval
timer. We also remove the use of the `op_async_void_deferred` call.

A future PR will be able to split the op sanitizer messages for timers
and intervals.
2024-02-16 15:35:51 +00:00
Matt Mastracci
e23fc6d88c
chore: deno_core bump (#22407)
- Adding `None` flag for warmup script.
 - Modify opcall trace interface to match new Rust implementation
2024-02-13 19:44:37 -07:00
David Sherret
d2477f7806
fix: cache bust jsr deps on constraint failure (#22372)
Removes the `FileFetcher`'s internal cache because I don't believe it's
necessary (we already cache this kind of stuff in places like deno_graph
or config files in different places). Removing it fixes this bug because
this functionality was already implemented in deno_graph and lowers
memory usage of the CLI a little bit.
2024-02-10 10:02:31 -05:00
David Sherret
e5e2c45998
fix: upgrade to deno_ast 0.33 (#22341)
* Uses diagnostics from deno_ast
* Real fix for https://github.com/denoland/deno/pull/22310
* Moves `deno lint --json` code here
* Upgrades swc
   
 Closes #22117
 Closes #22109
 Closes #21927
 Closes #20993
2024-02-09 01:40:26 +00:00
Bartek Iwańczuk
043fee48fd
chore: upgrade deno_core to 0.259.0 (#22311)
This update brings number of ops available to user code down to 45.
2024-02-07 02:16:08 +01:00
Divy Srivastava
e568cb8bf8
chore(ws): remove unused op_ws_send_pong (#22283) 2024-02-06 13:37:20 +01:00
Matt Mastracci
2c621f5894
refactor(cli): Move op descriptions into Rust and prepare for op import (#22271)
This moves the op sanitizer descriptions into Rust code and prepares for
eventual op import from `ext:core/ops`. We cannot import these ops from
`ext:core/ops` as the testing infrastructure ops are not always present.

Changes:
- Op descriptions live in `cli` code and are currently accessible via an
op for the older sanitizer code
 - `phf` dep moved to workspace root so we can use it here
- `ops.op_XXX` changed to to `op_XXX` to prepare for op imports later
on.
2024-02-05 20:26:59 +00:00
Matt Mastracci
0a3d329dd8
Reland refactor(cli): use new sanitizer for resources (#22226)
Originally in #22125
Reverted in #22153 because of #22148

Fixed in deno_core https://github.com/denoland/deno_core/pull/538

Test plan: 

1. Check out: https://github.com/poolifier/poolifier-deno.git

2. `PATH=.../deno/target/release/:$PATH deno task test`

3. `ok | 13 passed (188 steps) | 0 failed (18s)`
2024-02-05 12:21:29 -07:00
Matt Mastracci
aecad7f353
refactor(cli): Add TestFailureDescription (#22267)
Extract zero-risk changes from #22226
2024-02-05 17:27:17 +00:00
David Sherret
4b7c6049ef
refactor: load bytes in deno_graph (#22212)
Upgrades deno_graph to 0.64 where deno_graph is now responsible for
turning bytes into a string. This is in preparation for Wasm modules.
2024-02-01 03:15:22 +00:00
Bartek Iwańczuk
7281775381
Revert "refactor(cli): use new sanitizer for resources (#22125)" (#22153) 2024-01-27 20:37:36 +01:00
Matt Mastracci
84fb2ad71b
refactor(cli): use new sanitizer for resources (#22125)
Step 1 of the Rustification of sanitizers, which unblocks the faster
timers.

This replaces the resource sanitizer with a Rust one, using the new APIs
in deno_core.
2024-01-26 17:24:16 -05:00
David Sherret
4e72ca313a
refactor: use globbing from deno_config (#21925) 2024-01-15 19:15:39 -05:00
David Sherret
e212e1fc35
perf: skip expanding exclude globs (#21817)
We were calling `expand_glob` on our excludes, which is very expensive
and unnecessary because we can pattern match while traversing instead.

1. Doesn't expand "exclude" globs. Instead pattern matches while walking
the directory.
2. Splits up the "include" into base paths and applicable file patterns.
This causes less pattern matching to occur because we're only pattern
matching on patterns that might match and not ones in completely
unrelated directories.
2024-01-08 17:18:42 +00:00
David Sherret
7e72f3af61
chore: update copyright to 2024 (#21753) 2024-01-01 19:58:21 +00:00
Matt Mastracci
76a6ea5775
refactor(cli): update to new deno_core promise/call methods (#21519) 2023-12-13 08:07:26 -07:00
David Sherret
ddfbe71ced
feat(lsp): provide quick fixes for specifiers that could be resolved sloppily (#21506) 2023-12-08 09:57:06 -05:00
David Sherret
78566753c8
feat: add suggestions to module not found error messages for file urls (#21498) 2023-12-07 15:59:13 -05:00
Matt Mastracci
4a9f429501
refactor(cli): refactor bench/test for future module changes (#21460)
Extracting some refactorings for the module work that will land in
https://github.com/denoland/deno_core/pull/359/
2023-12-05 09:26:06 -07:00
Bartek Iwańczuk
a8c24d2a8b
fix: 'Promise was collected' error in REPL/jupyter (#21272)
Fixes #20528
2023-11-22 03:45:34 +01:00
David Sherret
02822d309f
fix(test): --junit-path should handle when the dir doesn't exist (#21044)
Closes https://github.com/denoland/deno/issues/21022
2023-11-01 19:59:51 +00:00
Bartek Iwańczuk
1713df1352
feat: deno run --unstable-hmr (#20876)
This commit adds `--unstable-hmr` flag, that enabled Hot Module Replacement.

This flag works like `--watch` and accepts the same arguments. If
HMR is not possible the process will be restarted instead.

Currently HMR is only supported in `deno run` subcommand.

Upon HMR a `CustomEvent("hmr")` will be dispatched that contains
information which file was changed in its `details` property.

---------

Co-authored-by: Valentin Anger <syrupthinker@gryphno.de>
Co-authored-by: David Sherret <dsherret@gmail.com>
2023-10-31 01:25:58 +01:00
Bartek Iwańczuk
8d9fef3b89
refactor: add WatcherCommunicator helper struct (#20927)
This commit introduces "WatcherCommunicator" struct that
is used facilitate bi-directional communication between CLI
file watcher and the watched function.

Prerequisite for https://github.com/denoland/deno/pull/20876
2023-10-19 07:05:00 +02:00
Nayeem Rahman
ac464ead49
feat(jupyter): support Deno.test() (#20778) 2023-10-05 21:45:11 +01:00
Nayeem Rahman
551a081450
refactor(test): support custom writer in PrettyTestReporter (#20783) 2023-10-05 11:25:15 +01:00
Luca Casonato
aef56f3d70
fix(cli): panic with __runtime_js_sources (#20704)
Also a drive-by cleanup elsewhere (removing unused enum).

Fixes #20702
2023-09-27 10:55:54 +00:00
Nayeem Rahman
fa18878f54
fix(lsp): include JSON modules in local import completions (#20536) 2023-09-17 07:50:30 +01:00
Nayeem Rahman
11f0ccf805
fix(test): share fail fast tracker between threads (#20515) 2023-09-15 15:46:48 +00:00
Luca Casonato
851f795001
fix: output traces for op sanitizer in more cases (#20494)
This adds traces for the "started outside test, closed inside test"
case.
2023-09-14 16:38:15 +02:00
Nayeem Rahman
17276a1df9
fix: empty include in config file excludes all (#20404) 2023-09-08 15:04:45 +01:00
Nayeem Rahman
2cbd1b40cb
fix(test): apply filter before checking for "only" (#20389) 2023-09-06 17:07:37 +01:00
Marvin Hagemeister
e0a269c23a
fix: don't show filtered test suites as running (#20385) 2023-09-06 14:54:21 +02:00
Nayeem Rahman
d28384c3de
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:
61f08d5a71/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 16:31:31 +01:00
Nayeem Rahman
e1fe31508c
fix(lsp/testing): use full ancestry to compute static id of step (#20297)
Fixes https://github.com/denoland/vscode_deno/issues/656.

Test steps were ID'd by a checksum of `[origin, level, step_name]` which
is incorrect. Now it's `[origin, ...ancestor_names, step_name]`.
2023-08-27 11:16:09 +02:00
Valentin Anger
a526cff0a9
feat(cli/tools): add TAP test reporter (#14390) (#20073)
This PR adds a test reporter for the [Test Anything
Protocol](https://testanything.org).

It makes the following implementation decisions:
- No TODO pragma, as there is no such marker in `Deno.test`
- SKIP pragma for `ignore`d tests
- Test steps are treated as TAP14 subtests
  - Support for this in consumers seems spotty
- Some consumers will incorrectly interpret these markers, resulting in
unexpected output
- Considering the lack of support, and to avoid implementation
complexity,
subtests are at most one level deep (all test steps are in the same
subtest)
- To accommodate consumers that use comments to indicate test-suites
(unspecced)
  - The test module path is output as a comment
  - This is disabled for `--parallel` testing
- Failure diagnostics are output as JSON, which is also valid YAML
- The structure is not specified, so the format roughly follows the spec
example:
  ```
  ---
  message: "Failed with error 'hostname peebles.example.com not found'"
  severity: fail
  found:
    hostname: 'peebles.example.com'
    address: ~
  wanted:
    hostname: 'peebles.example.com'
    address: '85.193.201.85'
  at:
    file: test/dns-resolve.c
    line: 142
  ...
  ```
2023-08-26 01:19:23 +02:00
Matt Mastracci
b1ce2e4167
fix(ext/web): add stream tests to detect v8slice split bug (#20253)
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
2023-08-23 17:03:05 -06:00
Alexander Michaud
b5839eefcf
fix(test): JUnit reporter includes file, line and column attributes (#20174)
Closes #20156
2023-08-17 23:41:29 +02:00
Bartek Iwańczuk
db287e216d
refactor: use '--reporter' and '--junit-path' flags for 'deno test' (#20031)
This commit adds "--reporter" and "--junit-path" flags to "deno test"
subcommand instead of using "--dot" and "--junit" flags.
2023-08-02 22:05:34 -04:00
Bartek Iwańczuk
029bdf0cd5
feat(cli): Add dot test reporter (#19804)
This commit adds a "dot" reporter to "deno test" subcommand,
that can be activated using "--dot" flag.

It provides a concise output using:
- "." for passing test
- "," for ignored test
- "!" for failing test

User output is silenced and not printed to the console.

In non-TTY environments each result is printed on a separate line.
2023-08-02 18:38:10 +02:00
Bartek Iwańczuk
806137bb96
refactor(cli/test): move reporters to a separate directory (#19957)
Just a small cleanup that will make #19804 easier.
2023-07-27 02:16:12 +02:00