Aaron O'Mullan
5f41f822e7
perf(webidl): optimize createRecordConverter() ( #12286 )
...
Cuts self-time by ~6x, 172ns/iter => 22ns/iter benched on 1M Response builds / HeadersInit calls
2021-10-04 15:39:32 +02:00
Aaron O'Mullan
ea7a63cd5a
refactor(core): split opcall into sync/async ( #12312 )
2021-10-04 12:34:53 +02:00
Aaron O'Mullan
4a1300edde
fix(core/runtime): sync_ops_cache if nuked Deno ns ( #12302 )
...
Decouple JsRuntime::sync_ops_cache() from the availability of the Deno.* namespace in the global scope
This avoids crashes when calling sync_ops_cache() on a bootstrapped WebWorker who has dropped its Deno.* namespace
It's also just cleaner and more robust ...
2021-10-04 11:45:41 +02:00
Aaron O'Mullan
11acdf1ea8
perf(core): use opcall() directly ( #12310 )
...
Instead of the wrapper dispatch() func, also now forbids passing opIds to opSync()/opAsync() callers must always pass names
2021-10-03 21:02:50 +02:00
Andreu Botella
2170a41d97
feat(web): Implement DOMException's stack property. ( #12294 )
...
As per WebIDL (https://heycam.github.io/webidl/#es-DOMException-specialness ),
if `Error` objects have a `stack` property, so should `DOMException`
instances.
2021-10-03 17:21:49 +02:00
Divy Srivastava
8884141c3f
fix(ext/crypto): missing Aes key typings ( #12307 )
2021-10-03 15:24:46 +02:00
Ahab
198f5b9514
fix(ext/net): should not panic when listening to unix abstract address ( #12300 )
2021-10-02 19:09:59 +02:00
Divy Srivastava
21520f5ff1
feat(ext/crypto): decode RSAES-OAEP-params with default values ( #12292 )
2021-10-02 15:20:53 +02:00
Aaron O'Mullan
26de516558
feat(core): implement Deno.core.isProxy() ( #12288 )
2021-10-01 20:25:33 +02:00
Divy Srivastava
1dfa35b2ba
fix(ext/crypto): use NotSupportedError for importKey() ( #12289 )
2021-10-01 13:16:11 +02:00
Aaron O'Mullan
f68825eda0
perf(fetch): optimize fillHeaders() key iteration ( #12287 )
...
Reduces self-time by ~70x (~70ms => ~1ms on 1M iters)
for...in filtered by hasOwnProperty yields the same set of keys as Object.keys()
2021-10-01 12:17:16 +02:00
Divy Srivastava
5065c7bcd9
feat(ext/crypto): implement wrapKey ( #12125 )
2021-10-01 11:39:49 +02:00
Nayeem Rahman
b354eaa247
fix(runtime/js/workers): throw errors instead of using an op ( #12249 )
2021-10-01 11:30:55 +02:00
Divy Srivastava
c0b6c0eea5
fix(ext/crypto): decode id-RSASSA-PSS with default params ( #12147 )
2021-10-01 11:14:16 +02:00
Aaron O'Mullan
7f390612a3
perf(web): optimize byteLowerCase() ( #12282 )
2021-10-01 01:03:51 +02:00
Aaron O'Mullan
b3ceafaa5d
tools(bench): rebootstrap ( #12281 )
...
Enable deno devs to bench/profile/test JS code changes without doing a full --release rebuild.
Incremental release builds take ~4mn on M1s, often more on other machines ...
2021-10-01 00:42:24 +02:00
Casper Beyer
6bf5c850e6
fix(runtime/testing): format aggregate errors ( #12183 )
2021-09-30 21:54:56 +02:00
David Sherret
c3e441c5b5
fix: worker environment permissions should accept an array ( #12250 )
2021-09-30 15:50:59 -04:00
David Sherret
cdac190583
chore: ensure first process in shared flock test does not enter and exit lock before second enters ( #12255 )
2021-09-30 15:45:13 -04:00
Aaron O'Mullan
72836478f2
perf(webidl): optimize createDictionaryConverter() ( #12279 )
...
On a benchmark constructing Responses with headers this shaves off 25%
2021-09-30 21:33:12 +02:00
Andreu Botella
a2632c86b1
fix: Don't panic when a worker is closed in the reactions to a wasm operation. ( #12270 )
2021-09-30 19:52:58 +02:00
Aaron O'Mullan
68e5cdaff0
perf(web): ~400x faster http header trimming ( #12277 )
...
Use a regex substring match with a first/last char fastpath instead of 2 regex replaces. Roughly ~400x faster (423ms vs 0.7ms in profiled runs)
2021-09-30 18:39:55 +02:00
Sylvain Cau
ee2e25fba7
feat(cli/uninstall): add uninstall command ( #12209 )
2021-09-30 11:38:07 -04:00
Yoshiya Hinosawa
f602d63f48
fix(repl): avoid panic when assigned to globalThis ( #12273 )
2021-10-01 00:25:58 +09:00
Luca Casonato
0d7a417f33
feat(tls): custom in memory CA certificates ( #12219 )
...
This adds support for using in memory CA certificates for
`Deno.startTLS`, `Deno.connectTLS` and `Deno.createHttpClient`.
`certFile` is deprecated in `startTls` and `connectTls`, and removed
from `Deno.createHttpClient`.
2021-09-30 09:26:15 +02:00
mynane
62920e4ef5
fix(ext/http): merge identical if/else branches ( #12269 )
2021-09-29 11:55:10 -07:00
Aaron O'Mullan
4cf744032d
perf(web/Event): move last class field to constructor ( #12265 )
2021-09-29 20:50:25 +02:00
Ahab
923d9c7786
fix(ext/fetch): avoid panic when header is invalid ( #12244 )
2021-09-29 18:42:06 +02:00
Aaron O'Mullan
c896ba2e19
perf(webidl/DOMString): don't wrap string primitives ( #12266 )
2021-09-29 18:06:29 +02:00
Aaron O'Mullan
cd913e51ef
bench(cli): add deno_http_native_headers.js ( #12267 )
2021-09-29 18:06:20 +02:00
Andreu Botella
cdb252af0a
feat: support serializing WebAssembly.Module objects ( #12140 )
2021-09-29 10:47:24 +02:00
Luca Casonato
5b526e5d17
feat: stabilize URLPattern API ( #12256 )
2021-09-29 10:42:32 +02:00
Aaron O'Mullan
3ab682918a
1.14.2 ( #12254 )
2021-09-28 14:00:26 +02:00
Aaron O'Mullan
e8901281da
chore: bump crate versions for 1.14.2 ( #12253 )
2021-09-28 12:19:03 +02:00
Aaron O'Mullan
9167f0c6bd
perf(fetch): optimize newInnerRequest blob url check ( #12245 )
...
Avoid "blob:" prefix check on requests built in the http module since those can never be blob objects
Reduces cost of `newInnerRequest()` from 20ms to 0.1ms in my profiled run on ~2.5M reqs
2021-09-27 13:19:24 +02:00
Luca Casonato
ff3a17b72d
fix: subprocess kill support on windows ( #12134 )
2021-09-27 12:18:02 +02:00
Aaron O'Mullan
0964685486
refactor(fetch/Request): inline defaultInnerRequest ( #12241 )
...
Similar to #12235
2021-09-27 11:13:27 +02:00
Aaron O'Mullan
8aba521e18
perf(fetch/Response): avoid class fields ( #12237 )
2021-09-27 10:51:32 +02:00
Aaron O'Mullan
34a15545c9
refactor(fetch/response): inline defaultInnerResponse ( #12235 )
...
Not useful to have the defaults externally defined when they're only used in `newInnerResponse()`. Also match order in `newInnerResponse()` and `cloneInnerResponse`
2021-09-26 23:46:56 +02:00
Aaron O'Mullan
5788f2e082
perf(web): optimize Event constructor ( #12231 )
...
Assign in constructor instead of using class initializers which are currently ~10x slower
2021-09-26 20:41:05 +02:00
Aaron O'Mullan
1749e79f97
perf(webidl/ByteString): 3x faster ASCII check ( #12230 )
2021-09-26 20:40:37 +02:00
Aaron O'Mullan
7f2976b3e6
perf(fetch): optimize InnerBody constructor ( #12232 )
...
Avoid initializers due to overhead
2021-09-26 20:40:16 +02:00
Ben Noordhuis
2b6f8d0187
fix(ext/http): include port number in h2 urls ( #12181 )
2021-09-26 20:26:16 +02:00
Aaron O'Mullan
6c007aa5ab
perf(fetch/headers): optimize appendHeader ( #12234 )
...
Use a single regex to check for `\0`, `\n`, `\r` instead of 3 `String.includes(...)` calls
2021-09-26 20:19:02 +02:00
Luis Malheiro
b095157c1d
perf(ext/fetch): Use the WebIDL conversion to DOMString rather than USVString for Response constructor ( #12201 )
2021-09-25 15:30:31 +02:00
Dan Rose
09f2cdbc72
fix(ext/web): FileReader error messages ( #12218 )
2021-09-25 15:28:17 +02:00
Andreu Botella
1380defb80
chore: update wpt ( #12206 )
2021-09-25 15:27:34 +02:00
Aaron O'Mullan
3c88dffd32
fix(http): panic when responding to a closed conn ( #12216 )
...
Our oneshot receiver in `HyperService::call` would unwrap and panic, the `.await` on the oneshot receiver happens when the sender is dropped.
The sender is dropped in `op_http_response` because:
1. We take `ResponseSenderResource`
2. Then get `ConnResource` and early exit on failure (conn already closed)
3. The taken sender then gets dropped in this early exit before any response is sent over the channel
Fallbacking to returning a dummy response to hyper seems to be a fine quickfix
2021-09-25 13:22:19 +02:00
Ben Noordhuis
16ea39ee48
fix(ext/http): fortify "is websocket?" check ( #12179 )
...
Check for expected headers more rigorously and check that it's a
HTTP/1.1 GET request. The logic mirrors what Deno Deploy and the
tungstenite crate do.
The presence of "Sec-Websocket-Version: 13" is now also enforced.
I don't expect that to break anything: conforming clients already
send it and tungstenite can't talk to older clients anyway.
The new code is more efficient due to heap-allocating less and aligns
more closely with the checks in ext/http/01_http.js now.
2021-09-25 10:02:26 +02:00
Andreu Botella
1a6249c971
chore(workers): Test that closing a worker closes any child workers ( #12215 )
...
Before #12156 , closing a worker which had children would cause a panic
(https://github.com/denoland/deno/issues/11342#issuecomment-918327693 ).
After that PR, closing a worker will also close any child workers.
2021-09-24 19:26:57 -07:00
