profile
viewpoint
Fumiya Karasawa fmy MeetsMore Inc. Tokyo, JAPAN Co-Founder & CTO @ MeetsMore Inc.

fmy/github2qiita 3

Github :point_right: Qiita

fmy/dotfiles 2

dotfiles

fmy/I-Feel-Fine 1

Examples of Objective-C

fmy/atc 0

build scheduler and web interface

fmy/bingo_flat 0

bingo machine with flat design

fmy/bootstrap 0

Sleek, intuitive, and powerful front-end framework for faster and easier web development.

fmy/browser-extension 0

OctoLinker – Available on Chrome and Firefox.

fmy/crystal 0

The Crystal Programming Language

issue commentnodejs/node

Http connections aborted after 5s / keepAliveTimeout

In case this helps - I think there's an additional use case which isn't related to the transmission size, but rather it's a race condition between the start / end of requests on the same connection.

  1. First request, new connection opened, keep-alive timer starts.
  2. First request ends
  3. Keep-alive timer kicks in, "timeout" event is about to be emitted on the socket 3.1 First bytes of second request come in on the soon-to-be-destroyed socket
  4. "timeout" event is emitted on the socket, resulting in the destruction of the socket 4.1 Second request gets dropped
  5. Second requests parsing (parseOnIncoming..) starts, trying to reset the timeout on the already-destroyed socket

Important: Note that in this test (see "Reproduce" below) we're not transmitting large amounts of data on the connection - which is (I think) why f6a725e doesn't fix this issue. Put differently, I think it's a timing issue from the time one requests ends (and keepAlive timer is about to end) - and just at the (almost) same time - a new request comes in.

*I'm not 100% sure this is the correct flow - it's just my best educated-guess, based on some debugging

Reproduce:

  • I simply used the test code in this repo: https://github.com/yoavain/node8keepAliveTimeout
  • Pretty consistent result - socket gets destroyed after the first couple of requests (easy to reproduce)

Tested Versions:

  • v8.16.1
  • v12.18.0
  • v12.19.1
  • v12.20.0
  • v14.7.0

Naive Approach to a Fix Assuming my debug analysis is correct, I would assume that there are 2 main ways to go about a fix for this:

  • either prevent the timeout event from firing on the socket if the socket is still in use
  • or, when ever the timeout event has fired, we then check if the socket is still in use, in which case we do nothing

Here's a naive code listing of the second approach:

// file: _http_server.js
function socketOnTimeout() {
  // "this" is the socket
  if (this.isInUse) {
    return;
  }  

  const req = this.parser && this.parser.incoming;
  const reqTimeout = req && !req.complete && req.emit('timeout', this);
  const res = this._httpMessage;
  const resTimeout = res && res.emit('timeout', this);
  const serverTimeout = this.server.emit('timeout', this);

  if (!reqTimeout && !resTimeout && !serverTimeout)
    this.destroy();
}
pbininda

comment created time in 3 minutes

issue commentmui-org/material-ui

[RFC] v5 styling solution 💅

Can you give us example of how customization of complex components will work with styled API? If I want to change list class on Menu component for example, how will that work with styled api. I know I can generate class with emotion, and set classes.list, but since lot of people prefer styled-components and styled api, I am just curious what would be solution for that.

mnajdova

comment created time in 6 minutes

pull request commentmui-org/material-ui

Upgrade to argos-cli@0.3.1

<!-- 0 failure: 0 warning:

1 markdown notices DangerID: danger-id-Danger; -->

Bundle size will be reported once Azure build #19142 finishes. <p align="right"> Generated by :no_entry_sign: <a href="https://danger.systems/js">dangerJS</a> against 5f7dd5fc8977de4fc62b818f0fb38fec34fd795d </p>

oliviertassinari

comment created time in 11 minutes

PR opened mui-org/material-ui

Upgrade to argos-cli@0.3.1 core

#23623 breaks the integration with Material-UI X. I have released the change directly in https://github.com/argos-ci/argos-cli/releases/tag/v0.3.1.

+4 -80

0 comment

3 changed files

pr created time in 14 minutes

pull request commentmui-org/material-ui

[docs] Remove unused abstraction

<!-- 0 failure: 0 warning:

1 markdown notices DangerID: danger-id-Danger; -->

Bundle size will be reported once Azure build #19140 finishes. <p align="right"> Generated by :no_entry_sign: <a href="https://danger.systems/js">dangerJS</a> against 1d5fbeb5e62fe77158d51211aa00747f80c4c24b </p>

oliviertassinari

comment created time in 21 minutes

push eventmui-org/material-ui

Matt

commit sha 16b9f53d9785ff62260939af21f6290356befcd9

[docs] API pages i18n (#23214)

view details

push time in 23 minutes

PR opened mui-org/material-ui

[docs] Remove unused abstraction docs

One chunk of #23721

+15 -15

0 comment

2 changed files

pr created time in 26 minutes

pull request commentmui-org/material-ui

[docs] Improve classes descriptions

<!-- 0 failure: 0 warning:

1 markdown notices DangerID: danger-id-Danger; -->

Bundle size will be reported once Azure build #19138 finishes. <p align="right"> Generated by :no_entry_sign: <a href="https://danger.systems/js">dangerJS</a> against 58166f662e8b50a7320fb104997a3256016bb8b7 </p>

oliviertassinari

comment created time in 29 minutes

PR opened mui-org/material-ui

[docs] Improve prop descriptions

One chunk of #23721

+3 -3

0 comment

3 changed files

pr created time in 32 minutes

pull request commentnodejs/node

errors: eliminate all overhead for hidden calls

CI: https://ci.nodejs.org/job/node-test-pull-request/34564/

mmomtchev

comment created time in an hour

pull request commentnodejs/node

lib: add span API to diagnostics_channel

I believe @RafaelGSS was working on making a fastify pugin

Yes, here https://github.com/fastify/fastify/issues/2697 is the tracking issue. I haven't read the code though but seems reasonable to have a unique way to deal with Spans.

Qard

comment created time in an hour

issue commentnodejs/node

buffer.write performance issue

Seeing this on another Buffer code-path

image

ledbit

comment created time in an hour

issue commentnodejs/node

The long standing bugs that I kept ignoring, yet here I report....

@colicrod seems the parse itself priorize # as end of hostname.

if (!hostlessProtocol.has(lowerProto) &&
    (slashes || (proto && !slashedProtocol.has(proto)))) {

  // there's a hostname.
  // the first instance of /, ?, ;, or # ends the host.
  //
  // If there is an @ in the hostname, then non-host chars *are* allowed
  // to the left of the last @ sign, unless some host-ending character
  // comes *before* the @-sign.
  // URLs are obnoxious.
  //
  // ex:
  // http://a@b@c/ => user:a@b host:c
  // http://a@b?@c => user:a host:b path:/?@c
colicrod

comment created time in an hour

push eventmui-org/material-ui

Mhayk Whandson

commit sha 02b273d239302af3c21cb7b2cafbd32035170398

[TextareaAutosize] Deprecate rowsMax->maxRows & rowsMin->minRows (#23530)

view details

push time in an hour

PR merged mui-org/material-ui

[TextareaAutosize] Deprecate rowsMax->maxRows & rowsMin->minRows component: TextareaAutosize deprecation

<!-- Thanks so much for your PR, your contribution is appreciated! ❤️ -->

Add deprecations for #21873

+204 -77

2 comments

34 changed files

mhayk

pr closed time in an hour

pull request commentmui-org/material-ui

[TextareaAutosize] Deprecate rowsMax->maxRows & rowsMin->minRows

Thanks @mhayk!

mhayk

comment created time in an hour

pull request commentnodejs/node

deps: V8: cherry-pick 86991d0587a1

CI: https://ci.nodejs.org/job/node-test-pull-request/34563/

bcoe

comment created time in an hour

push eventconcourse/concourse

Aidan Oldershaw

commit sha 805fb859db192d6ef4394e68a80289bd1b1dc114

containerd: behaviour: return exe not found error concourse/concourse#6098 doesn't work for containerd because our runtime doesn't try to detect exe not found errors. This was copied from gdn: https://github.com/cloudfoundry/guardian/blob/3da697a62154c6d1f797aede3a0fb23f2297e6cd/rundmc/runcontainerd/runcontainerd.go#L201-L221 Also, use fake iptables in the network test so I can run it on mac Signed-off-by: Aidan Oldershaw <aoldershaw@pivotal.io>

view details

Aidan Oldershaw

commit sha 7908bc73aaf6822d2cb8582a82b2a574a0c1d36e

testflight: add coverage for default shell fallback It's a feature that requires cooperation between worker, ATC, and fly, and is an easy thing to miss without higher-level coverage. Signed-off-by: Aidan Oldershaw <aoldershaw@pivotal.io>

view details

Aidan Oldershaw

commit sha 9104ed6a18089ebb7349a2e9b509e661f0bb4ff1

containerd: behaviour: fix exe not found check Should be bound to proc.Start rather than task.Exec, which doesn't execute the process Signed-off-by: Aidan Oldershaw <aoldershaw@pivotal.io>

view details

Aidan Oldershaw

commit sha 0b4a952e1f6f3a4871d56ea5f54a45ae84d65a53

Merge pull request #6304 from concourse/sh-fallback-containerd `fly intercept` falls back to `sh` when `bash` is missing (containerd runtime)

view details

push time in an hour

delete branch concourse/concourse

delete branch : sh-fallback-containerd

delete time in an hour

PR merged concourse/concourse

`fly intercept` falls back to `sh` when `bash` is missing (containerd runtime) bug misc

<!-- Hi there! Thanks for submitting a pull request to Concourse!

The title of your pull request will be used to generate the release notes. Please provide a brief sentence that describes the PR, using the imperative mood. Please refrain from adding prefixes like 'feature:', and don't include a period at the end.

Examples: "Add feature to doohickey", "Fix panic during spline reticulation"

We will edit the title if needed so don't worry about getting it perfect!

To help us review your PR, please fill in the following information. -->

What does this PR accomplish?

<!-- Choose all that apply. Also, mention the linked issue here. This will magically close the issue once the PR is merged. --> Bug Fix | Feature | Documentation

#6098 changed the behaviour of fly intercept to fallback to sh as the default command if the container doesn't have bash. However, this requires cooperation with the runtime to return a specific garden error, which containerd did not follow.

Changes proposed by this PR:

<!-- Tell the reviewer What changed, Why, and How were you able to accomplish that? -->

  • Return a garden.ExecutableNotFoundError when the executable is not found (logic copied from Guardian)
  • Add testflight coverage

Notes to reviewer:

<!-- Leave a message to whoever is going to review this PR. Mainly, pointers to review the PR, and how they can test it. -->

Contributor Checklist

<!-- Most of the PRs should have the following added to them, this doesn't apply to all PRs, so it is helpful to tell us what you did. -->

Reviewer Checklist

<!-- This section is intended for the reviewers only, to track review progress. -->

  • [ ] Code reviewed
  • [ ] Tests reviewed
  • [ ] Documentation reviewed
  • [ ] Release notes reviewed
  • [ ] PR acceptance performed
  • [ ] New config flags added? Ensure that they are added to the BOSH and Helm packaging; otherwise, ignored for the integration tests (for example, if they are Garden configs that are not displayed in the --help text).
+128 -11

0 comment

5 changed files

aoldershaw

pr closed time in an hour

pull request commentmui-org/material-ui

[docs] Add settings panel to allow system mode

<!-- 0 failure: 0 warning:

1 markdown notices DangerID: danger-id-Danger; -->

Bundle size will be reported once Azure build #19136 finishes. <p align="right"> Generated by :no_entry_sign: <a href="https://danger.systems/js">dangerJS</a> against 0bf7c05f3662ae1d69cdd5710175d70bb53e9a35 </p>

mbrookes

comment created time in an hour

PR opened mui-org/material-ui

[docs] Add settings panel to allow system mode docs

This has been on my list of things to do for longer than I care to remember. 😄

Closes #23683

+247 -94

0 comment

7 changed files

pr created time in an hour

pull request commentmui-org/material-ui

[docs] API pages i18n (#23214)

<!-- 0 failure: 0 warning:

1 markdown notices DangerID: danger-id-Danger; -->

Bundle size will be reported once Azure build #19135 finishes. <p align="right"> Generated by :no_entry_sign: <a href="https://danger.systems/js">dangerJS</a> against 1403e9d3ccc664846e86aac177030d9626fe1f8b </p>

oliviertassinari

comment created time in 2 hours

Pull request review commentnodejs/node

dns: Update to c-ares 1.17.1

 #include "ares_nowarn.h" #include "ares_private.h" +#ifndef T_OPT+#  define T_OPT  41 /* EDNS0 option (meta-RR) */+#endif

Here it is: https://github.com/c-ares/c-ares/pull/385

lxdicted

comment created time in 2 hours

PR opened mui-org/material-ui

[docs] API pages i18n (#23214) docs

Second try :) https://github.com/mui-org/material-ui/pull/23214#issuecomment-728954686

+14182 -11139

0 comment

435 changed files

pr created time in 2 hours

create barnchmui-org/material-ui

branch : docs-api-i18n

created branch time in 2 hours

issue openednodejs/node

The long standing bugs that I kept ignoring, yet here I report....

# node -p 'require("url").parse("http://username:pass#123cba@myproxy.com:8080")'; node --version
Url {
  protocol: 'http:',
  slashes: true,
  auth: null,
  host: 'username',
  port: null,
  hostname: 'username',
  hash: '#123cba@myproxy.com:8080',
  search: null,
  query: null,
  pathname: '/:pass',
  path: '/:pass',
  href: 'http://username/:pass#123cba@myproxy.com:8080'
}
v14.15.1

Is just me who using # in the proxy password? as i believe this is an old bug from old nodejs up till now in 14.x

created time in 2 hours

Pull request review commentnodejs/node

dns: Update to c-ares 1.17.1

 #include "ares_nowarn.h" #include "ares_private.h" +#ifndef T_OPT+#  define T_OPT  41 /* EDNS0 option (meta-RR) */+#endif

This piece of code is used in other places of c-ares, too. I'm going to provide a pull request for it in c-ares. Don't know how this should be handled in nodejs.

lxdicted

comment created time in 2 hours

pull request commentnodejs/node

dns: Update to c-ares 1.17.1

@richardlau I can't see, how these failing tests relate to this pull request:

12:24:29 not ok 2497 parallel/test-webcrypto-derivebits-hkdf
12:24:29   ---
12:24:29   duration_ms: 122.197
12:24:29   severity: fail
12:24:29   exitcode: -15
12:24:29   stack: |-
12:24:29     timeout
12:24:29   ...
12:24:29 not ok 2498 parallel/test-webcrypto-derivebits-node-dh
12:24:29   ---
12:24:29   duration_ms: 121.988
12:24:29   severity: fail
12:24:29   exitcode: -15
12:24:29   stack: |-
12:24:29     timeout
12:24:29   ...
12:25:44 not ok 2499 parallel/test-webcrypto-derivebits-pbkdf2
12:25:50   ---
12:26:05   duration_ms: 180.679
12:26:05   severity: fail
12:26:05   exitcode: -15
12:26:05   stack: |-
12:26:05     timeout
12:26:05   ...

Help would be very welcome to track this down, if this is related to this pull request.

lxdicted

comment created time in 2 hours

issue commentnodejs/node

Proposal: mark AsyncResource and AsyncLocalStorage as stable

Sounds reasonable. :)

Qard

comment created time in 2 hours

more