Noted, and also noted that AdvancedPool is optional.

Fixed, thanks

Here's what I did on Windows using Go's Qt wrapper: https://github.com/cretz/ffembedpoc. That takes Firefox's window and makes it a control. Granted it has to turn off its tabs and communicates over a control point, but idea is there.

(I have plenty of familiarity w/ Qt C++, but trying to find a binding that encourages more contributions)

That file comes from https://github.com/mattn/go-sqlite3. Are you able to build that project?

The dialer is just a SOCKS proxy at https://github.com/cretz/bine/blob/2d69661a2441f449d2d9ec15ab033252b82af08f/tor/dialer.go#L81 like any other socks proxy using Tor. You can probably set StartConf.NoHush to true which will not set --hush on the CLI when starting Tor. You can also create a torrc file with Log options (see manual) and set that as StartConf.TorrcFile.

You can probably also get lots of information over the control port. I haven't investigated to see which ones might help you, but Tor.Control has several things you can run to inspect Tor's state (like seeing which circuits are open, or trying to build them, or even getting descriptors for onion services). The details of calls that can be made on the control port are at https://github.com/torproject/torspec/blob/main/control-spec.txt.

Nice! Afraid I don't have any intentions or insight on doing this, but I do agree it would be cool.

Looks like just too many constants. I'd have to dig into seeing what constants asmble is generating, but short of changing how asmble works, I don't think there's an easy fix.

Feature Description

Today at https://github.com/k6io/k6/blob/cc0361ca4a5443e5b90d22b4c0e62a9f15c9757d/js/common/bridge.go#L151 you must have more than one return value to propagate Go errors. This surprised my developing a side-effect-only function that just returned a single error. I now return (struct{}, error) to have my Go error seen as a real error and not just a single value return.

Suggested Solution

Assuming this is by intention, should document this somewhere. Currently k6 Extensions in the docs links to https://k6.io/blog/extending-k6-with-xk6/ which states:

If err is not nil, it will be thrown with the appropriate message in the JS script.

Though it's understandable that a high-level blog post won't have this detail. Unsure where to document it, just thought I'd point out the surprise. I suppose a single-error-only-return func test has value at https://github.com/k6io/k6/blob/master/js/common/bridge_test.go to affirm behavior.

Given a proto w/ a message named Foo in somedir1/foo.proto and a service named Foo in somedir2/foo.proto that references the Foo message from somedir1/foo.proto, you'll get an error on the import { Foo } from "../somedir1/foo"; with:

Import declaration conflicts with local declaration of 'Foo'

Can the answer be placed here and/or can a public link to the mailing list archive be provided?

Onionbalance for v3 relies on a recent PR at https://github.com/torproject/tor/pull/1647 (discussion at https://gitlab.torproject.org/tpo/core/tor/-/issues/32709) which only applies to torrc-based onion services, not ephemeral ones via ADD_ONION. Mailing list discussion started at https://lists.torproject.org/pipermail/tor-dev/2021-June/014570.html.

I may also look into impl'ing a poc of this PDF which discusses multiple instances using the same hostname and priv key.

At least need an example showing it, if not helpers to make it easy.

Adding help wanted tag. Until the Go issues mentioned in the PR at #13 are done, this would be a manual effort w/ syscall.

Is it possible to change the length of onion ID that we get, i.e. can ask for greater sized onion ID that we get from the following code?

onion, err := t.Listen(listenCtx, &tor.ListenConf{Version3: true, RemotePorts: []int{80}})

Closing old issue. Let mw know if this issue persists.

linux 4.19, tor 0.4.0.5. I was trying to run the example client code to create communicating hidden services, but i couldn't get the example code to run on my manjaro linux.

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x63cf53]

goroutine 24 [running]:
golang.org/x/net/proxy.(*socks5).Dial(0xc00009ec30, 0x6f2455, 0x3, 0xc0000966a0, 0xf, 0xf, 0x0, 0x28a, 0x0)
	/home/h/go/src/golang.org/x/net/proxy/socks5.go:71 +0x63
github.com/cretz/bine/tor.(*Dialer).DialContext.func1(0xc000085180, 0x6f2455, 0x3, 0xc0000966a0, 0xf, 0xc00009aa80, 0x754200, 0xc000096010, 0xc00009aae0)
	/home/h/go/src/github.com/cretz/bine/tor/dialer.go:95 +0x7e
created by github.com/cretz/bine/tor.(*Dialer).DialContext
	/home/h/go/src/github.com/cretz/bine/tor/dialer.go:94 +0xfa
exit status 2

the code was:

package main

import (
	"context"
	"fmt"
	"io/ioutil"
	"log"
	"net/http"
	"time"

	"github.com/cretz/bine/tor"
)

func main() {
	if err := run(); err != nil {
		log.Fatal(err)
	}
}

func run() error {
	// Start tor with default config (can set start conf's DebugWriter to os.Stdout for debug logs)
	fmt.Println("Starting tor and fetching title of https://eludemaillhqfkh5.onion, please wait a few seconds...")
	t, err := tor.Start(nil, nil)
	if err != nil {
		return err
	}
	defer t.Close()
	// Wait at most a minute to start network and get
	dialCtx, dialCancel := context.WithTimeout(context.Background(), time.Minute)
	defer dialCancel()
	// Make connection
	dialer, err := t.Dialer(dialCtx, nil)
	if err != nil {
		return err
	}
	httpClient := &http.Client{Transport: &http.Transport{DialContext: dialer.DialContext}}
	// Get /
	resp, err := httpClient.Get("https://example.com")
	if err != nil {
		return err
	}
	defer resp.Body.Close()
	if err != nil {
		return err
	}
	bodyBytes, err := ioutil.ReadAll(resp.Body)
	fmt.Println(string(bodyBytes))
	return nil
}

A slight variation on the example client.

if I revisit this, I think I may accept a port-mappings structure that will be mutually exclusive with the existing local/remote port pieces. This will give the most flexibility.

Closing. Please comment if this is still an issue.

Hrmm, I suppose the fix here is to have enable network check bootstrapped even when already enabled. May still have event part ahead of time set per #40. Will investigate.

Will look to moving event capture asynchronous and before the sending of the conf. Same at #40.

Will look to moving event capture asynchronous and before the sending of the conf.