Sort of. If your runtime is reused while bindings are updated, then some code may break if it change

HHello, I’m Allie!Sort of. If your runtime is reused while bindings are updated, then some code ma...

A

Andrezzz•7/7/23, 2:37 PM

btw do u know how to debug correctly, since it's a webpack and then it goes thousands of lines

H

hanpolo•7/7/23, 4:04 PM

Folks, I tried upgrading node from 16.19.1 to 18.16.1. With 16.19.1 & wrangler 3.1.2 everything works. But when I try node 18.16.1 I get:

workerd/server/server.c++:2362: error: Uncaught exception: kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f00e350 10f011530
workerd/server/server.c++:2274: error: exception = kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f0011f0 10f0622d0 10e41ad30
[mf:inf] GET / 502 Bad Gateway (7ms)
✘ [ERROR] Error: Could not proxy request: Error: connect ECONNREFUSED ::1:3341

workerd/server/server.c++:2362: error: Uncaught exception: kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f00e350 10f011530
workerd/server/server.c++:2274: error: exception = kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f0011f0 10f0622d0 10e41ad30
[mf:inf] GET / 502 Bad Gateway (7ms)
✘ [ERROR] Error: Could not proxy request: Error: connect ECONNREFUSED ::1:3341

workerd/server/server.c++:2362: error: Uncaught exception: kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f00e350 10f011530
workerd/server/server.c++:2274: error: exception = kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f0011f0 10f0622d0 10e41ad30
[mf:inf] GET / 502 Bad Gateway (7ms)
✘ [ERROR] Error: Could not proxy request: Error: connect ECONNREFUSED ::1:3341

workerd/server/server.c++:2362: error: Uncaught exception: kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f00e350 10f011530
workerd/server/server.c++:2274: error: exception = kj/async-io-unix.c++:186: disconnected: ::write(fd, buffer, size): Broken pipe
stack: 110129a83 110129e81 10f032246 10f0326b7 10f0011f0 10f0622d0 10e41ad30
[mf:inf] GET / 502 Bad Gateway (7ms)
✘ [ERROR] Error: Could not proxy request: Error: connect ECONNREFUSED ::1:3341

H

hanpolo•7/7/23, 4:05 PM

Any ideas what is happening?

H

hanpolo•7/7/23, 4:07 PM

Some extra info: on the browser window I see:

[wrangler] Could not proxy request: TypeError: fetch failed

[wrangler] Could not proxy request: TypeError: fetch failed

[wrangler] Could not proxy request: TypeError: fetch failed

[wrangler] Could not proxy request: TypeError: fetch failed

The command to launch my app is:

 npx wrangler pages dev -- npm start

 npx wrangler pages dev -- npm start

 npx wrangler pages dev -- npm start

 npx wrangler pages dev -- npm start

J

James•7/7/23, 4:22 PM

This sounds highly likely related to the

localhost

localhost

localhost

localhost IPv6 vs IPv4 changes in Node 18 from Node 16.

J

James•7/7/23, 4:25 PM

Can you share a repro if possible with your code and wrangler.toml?

H

hanpolo•7/7/23, 6:00 PM

Repo @ https://github.com/safaalai/stencil-latest

GitHub

GitHub - safaalai/stencil-latest

Contribute to safaalai/stencil-latest development by creating an account on GitHub.

H

hanpolo•7/7/23, 6:00 PM

I've installed Node globally

H

hanpolo•7/7/23, 6:01 PM

Should I open a bug?

H

hanpolo•7/7/23, 6:03 PM

Note there is no wrangler.toml .. I've basically created the simplest repo that produces the bug

H

hanpolo•7/7/23, 9:47 PM

Ok .. so verified what you said, @cherryjimbo. If I run

node --dns-result-order=ipv4first node_modules/wrangler/bin/wrangler.js pages dev -- npm start

node --dns-result-order=ipv4first node_modules/wrangler/bin/wrangler.js pages dev -- npm start

node --dns-result-order=ipv4first node_modules/wrangler/bin/wrangler.js pages dev -- npm start

node --dns-result-order=ipv4first node_modules/wrangler/bin/wrangler.js pages dev -- npm start

Then everything works

H

hanpolo•7/7/23, 9:48 PM

not quite sure how else to pass node command line arguments (i.e. --dns-result-order) to npx .. so I'm stuck with this rather long-winded node command. Any suggestions?

G

gutterball•7/10/23, 9:08 AM

upgraded to wrangler 3.1.2 to use local https option, but get the following error:

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
*** Fatal uncaught kj::Exception: kj/async-io-unix.c++:945: failed: ::bind(sockfd, &addr.generic, addrlen): Address already in use; toString() = 127.0.0.1:9999
stack: 1021761bf 102175f57 1021737ef 100911083 100911244 10091194c 1008dc91f 1008e032b 1008e015f 1008e0147 1021ac0db 1021ac337 1021aaa8f 1021aa85f 1008d41ab 186ce7f27

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
*** Fatal uncaught kj::Exception: kj/async-io-unix.c++:945: failed: ::bind(sockfd, &addr.generic, addrlen): Address already in use; toString() = 127.0.0.1:9999
stack: 1021761bf 102175f57 1021737ef 100911083 100911244 10091194c 1008dc91f 1008e032b 1008e015f 1008e0147 1021ac0db 1021ac337 1021aaa8f 1021aa85f 1008d41ab 186ce7f27

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
*** Fatal uncaught kj::Exception: kj/async-io-unix.c++:945: failed: ::bind(sockfd, &addr.generic, addrlen): Address already in use; toString() = 127.0.0.1:9999
stack: 1021761bf 102175f57 1021737ef 100911083 100911244 10091194c 1008dc91f 1008e032b 1008e015f 1008e0147 1021ac0db 1021ac337 1021aaa8f 1021aa85f 1008d41ab 186ce7f27

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
*** Fatal uncaught kj::Exception: kj/async-io-unix.c++:945: failed: ::bind(sockfd, &addr.generic, addrlen): Address already in use; toString() = 127.0.0.1:9999
stack: 1021761bf 102175f57 1021737ef 100911083 100911244 10091194c 1008dc91f 1008e032b 1008e015f 1008e0147 1021ac0db 1021ac337 1021aaa8f 1021aa85f 1008d41ab 186ce7f27

using node 16.20.0 (tried switching to 18 but the error still occurs).
wrangler.toml setting:

[dev]
ip = "127.0.0.1"
port = 9999
local_protocol = "https"

[dev]
ip = "127.0.0.1"
port = 9999
local_protocol = "https"

[dev]
ip = "127.0.0.1"
port = 9999
local_protocol = "https"

[dev]
ip = "127.0.0.1"
port = 9999
local_protocol = "https"

EDIT: strangely, it works when passing the flags inline:

wrangler dev src/main.ts --live-reload --local-protocol https --port 9999

wrangler dev src/main.ts --live-reload --local-protocol https --port 9999

wrangler dev src/main.ts --live-reload --local-protocol https --port 9999

wrangler dev src/main.ts --live-reload --local-protocol https --port 9999

anyone knows why?

Ggutterball upgraded to wrangler 3.1.2 to use local https option, but get the following erro...

M

MrBBot•7/10/23, 11:09 AM

Hey!

Could you double check that you don't have any other

wrangler dev

wrangler dev

sessions running? Ideally restart your computer, but otherwise try

pkill workerd

pkill workerd

pkill workerd

pkill workerd.

MMrBBot Hey! 👋 Could you double check that you don't have any other `wrangler dev` sess...

G

gutterball•7/10/23, 3:32 PM

hey @mrbbot , thanks for replying! I restarted my machine and indeed it works! I'm still getting the errors in the console, but the app works which is what's important. thanks!

Ggutterball hey @mrbbot , thanks for replying! I restarted my machine and indeed it works! I...

M

MrBBot•7/10/23, 5:22 PM

Excellent! If you're seeing this regularly it's possible we've got an issue. Please open a GitHub issue if this becomes a problem for you.

K

kian•7/11/23, 11:21 AM

So the

put

put

promise never resolves?

V

vassbence•7/11/23, 4:38 PM

i'm getting

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
workerd/jsg/util.c++:277: info: exception = kj/async-io-unix.c++:1526: disconnected: connect(): Connection refused

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
workerd/jsg/util.c++:277: info: exception = kj/async-io-unix.c++:1526: disconnected: connect(): Connection refused

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
workerd/jsg/util.c++:277: info: exception = kj/async-io-unix.c++:1526: disconnected: connect(): Connection refused

workerd/util/symbolizer.c++:95: warning: Not symbolizing stack traces because $LLVM_SYMBOLIZER is not set. To symbolize stack traces, set $LLVM_SYMBOLIZER to the location of the llvm-symbolizer binary. When running tests under bazel, use `--test_env=LLVM_SYMBOLIZER=<path>`.
workerd/jsg/util.c++:277: info: exception = kj/async-io-unix.c++:1526: disconnected: connect(): Connection refused

V

vassbence•7/11/23, 4:38 PM

in

unstable_dev()

unstable_dev()

unstable_dev()

unstable_dev()

V

vassbence•7/11/23, 4:44 PM

it will has to do something with running in vitest setup files, as it does not happen when i start the

unstable_dev

unstable_dev

unstable_dev

unstable_dev in a regular test file

M

micromashor•7/11/23, 7:46 PM

is there a flag to force wrangler to run in non-interactive mode?

M

micromashor•7/11/23, 7:47 PM

I guess the hacky way is to pass

/dev/null

/dev/null

/dev/null

/dev/null as its stdin

Original message was deleted

M

micromashor•7/11/23, 8:26 PM

thanks

M

micromashor•7/11/23, 8:28 PM

actually that's probably the easiest way bc I'm running it from

make

make

make

make, so that would avoid having to add

</dev/null

</dev/null

</dev/null

</dev/null every time

M

micromashor•7/11/23, 8:29 PM

yup

H

hanpolo•7/12/23, 9:11 PM

So does anyone know of a guide on how to test Durable Objects with the new Wrangler 3.1.2? I cannot get any of my old tests to work. For instance, should I still be specifying environment: minflare? Because when I do it simply fails to find my durable object. Or should I set
"vitest-environment-miniflare/globals" in tsconfig.json?

Hhanpolo So does anyone know of a guide on how to test Durable Objects with the new Wrang...

M

MrBBot•7/13/23, 9:45 AM

Miniflare's Vitest environment hasn't been upgraded to v3 yet. We're working on it, but it's going to take a while...

If you've got existing Vitest tests written, I'd stick with v2 for the time being.

M

Muhymin•7/13/23, 11:26 AM

Hi

! I am having an issue with

wrangler tail

wrangler tail

wrangler tail

wrangler tail that it always exits automatically after some time. Is it the expected behavior?? I didn't find any solution for this online.

MMrBBot Miniflare's Vitest environment hasn't been upgraded to v3 yet. We're working on ...

H

hanpolo•7/13/23, 11:43 AM

Thanks for the response. I came to that conclusion yesterday, removed miniflare from my tests and wrote my own mocks so I could run tests for 3.1.2. looking forward to getting the miniflare update sometime in the future.

M

Muhymin•7/13/23, 12:03 PM

Off-Topic: @mrbbot I really thought you were a bot who creates threads. And I wondered why the bot didn't create a thread from my message

!

Ssimpsojo Seems it should just work (TM), but didn't for me (On windows, I got the error: ...

A

Adam•7/13/23, 8:40 PM

If there's an issue with expected functionality, please help us by filing an issue here https://github.com/cloudflare/workers-sdk/issues so we can take a look!

GitHub

Issues · cloudflare/workers-sdk

Home to Wrangler, the CLI for Cloudflare Workers® - Issues · cloudflare/workers-sdk

B

beelzabub•7/14/23, 9:01 PM

Does Miniflare 3 have any mechanism to overwrite what domain names map to (within the API, not wrangler)? For example, if my worker is doing a fetch to

https://my.custom.domain.com

https://my.custom.domain.com

https://my.custom.domain.com

https://my.custom.domain.com, I would like that to get routed to a localhost server for CI testing purposes without changing the URL within the code.

K

kian•7/14/23, 9:30 PM

Not yet https://github.com/cloudflare/miniflare/discussions/588#discussioncomment-5995047

Bbeelzabub Does Miniflare 3 have any mechanism to overwrite what domain names map to (withi...

M

MrBBot•7/14/23, 9:54 PM

Put a PR up, https://github.com/cloudflare/miniflare/pull/629, still planning to add support for fetch mocks like in Miniflare 2

A

ajgeiss0702•7/15/23, 1:16 AM

secrets are seperate for each worker, right?

J

James•7/15/23, 1:18 AM

yep!

J

James•7/15/23, 1:18 AM

I believe account-wide secrets is a feature that will be coming, but no ETA

A

ajgeiss0702•7/15/23, 1:21 AM

ok, cool. If i put a secret before publishing the worker, will it pick it up when published? Or will I need to wait to put secrets until after publishing?

A

ajgeiss0702•7/15/23, 1:22 AM

I guess ill find out in a few minutes anyway

J

James•7/15/23, 1:23 AM

I think you will have to create and publish the worker once before you add secrets? But honestly I've not tried that workflow in quite a while - let me know what happens

A

ajgeiss0702•7/15/23, 1:25 AM

will do

A

ajgeiss0702•7/15/23, 1:26 AM

looks like the pre-put secrets worked

J

James•7/15/23, 1:26 AM

awesome to hear!

V

vassbence•7/15/23, 6:38 PM

My tests using

unstable_dev()

unstable_dev()

unstable_dev()

unstable_dev() and R2 fail in CI. Has anyone seen something similar? Sadly giving a longer timeout (5+sec) for the test doesn't seem to fix the problem. The test finishes in local in 24ms.

J

James•7/15/23, 6:41 PM

That's not a lot of info to work with unfortunately. Are these tests that used to work and no longer? Is it just the R2 tests - do other tests work? What kind of debugging have you done to see where they seemingly hang? What versions of wrangler, etc. are you using? Can you create a minimal reproducible example of a single test that works locally but then fails in CI so we can reproduce?

V

vassbence•7/15/23, 6:42 PM

Sadly

loglevel: debug

loglevel: debug

loglevel: debug

loglevel: debug doesnt show any errors, test just hang with a simple

env.BUCKET.put

env.BUCKET.put

env.BUCKET.put

env.BUCKET.put

Sort of. If your runtime is reused while bindings are updated, then some code may break if it change

Similar Threads

Similar Threads

Similar Threads