Figure out way to improve stack trace when fails to connect to Chromedriver

[paulcsmith]

Right now there are lots of loops so the stacktrace is huge. Maybe using a while would shorten that up

[hanneskaeufler]

Is that starting with a

[1529857725.524][SEVERE]: bind() returned an error, errno=48: Address already in use (48) for you as well?

(Almost full stacktrace, omitted the hundreds of repetitions):

``` .................................................[1529857725.524][SEVERE]: bind() returned an error, errno=48: Address already in use (48) E

Failures:

1. Publishing a post results in a post on the frontpage

   Expected to find element on page, but it was not found.

   ▸ looking for: [name='post:title']
   

   lib/lucky_flow/src/lucky_flow/find_element.cr:48:5 in 'raise_element_not_found_error'
   lib/lucky_flow/src/lucky_flow/find_element.cr:22:7 in 'retry'
   lib/lucky_flow/src/lucky_flow/find_element.cr:14:33 in 'run'
   lib/lucky_flow/src/lucky_flow/find_element.cr:32:5 in 'retry_after_delay'
   lib/lucky_flow/src/lucky_flow/find_element.cr:20:7 in 'retry'
   lib/lucky_flow/src/lucky_flow/find_element.cr:14:33 in 'run'
   lib/lucky_flow/src/lucky_flow/find_element.cr:32:5 in 'retry_after_delay'

</details>

[paulcsmith]

Yeah, that huge stacktrace would be nice to shorten up. I think using a while or loop in

lucky_flow/src/lucky_flow/find_element.cr

Lines 13 to 24 in 324116d

	def run
	matching_elements.first? || retry
	end
	private def retry
	self.tries += 1
	if has_retries_left?
	retry_after_delay
	else
	raise_element_not_found_error
	end
	end

might work better. But I haven't tried it yet so it may still have the same issue

And no I don't see that error. That means you probably already have a chromedriver instance running. Maybe for some other app?

[johansmitsnl]

I have the same issue on my Debian laptop (testing)

crystal spec
Already created node_development
Did nothing. No pending migrations.
Unhandled exception in spawn:
bind: Address already in use (Errno)
  from /usr/share/crystal/src/socket/tcp_server.cr:73:15 in 'initialize'
  from /usr/share/crystal/src/socket/tcp_server.cr:32:3 in 'initialize:reuse_port'
  from /usr/share/crystal/src/socket/tcp_server.cr:32:3 in 'new:reuse_port'
  from /usr/share/crystal/src/http/server.cr:139:5 in 'bind_tcp'
  from src/app.cr:48:5 in 'listen'
  from spec/setup/start_app_server.cr:4:3 in '->'
  from /usr/share/crystal/src/fiber.cr:255:3 in 'run'
  from /usr/share/crystal/src/fiber.cr:29:34 in '->'
  from ???
[1540129866,250][SEVERE]: bind() returned an error, errno=98: Adres is al in gebruik (98)

It only works again when I kill: 6315 pts/3 Sl 0:00 lib/lucky_flow/src/lucky_flow/../../vendor/chromedriver-2.40-linux64 --port=4444 --url-base=/wd/hub

[jwoertink]

Just came across this. We had pointed to Brave as the browser_binary, but it would throw the Already in use error on this repo's specs. Running the lucky flow specs with crystal spec -DDEBUG let us see that the chromedriver didn't support anything other than Chrome 78.

This would be a great message to be able to display. As for how? no clue 😆

/cc @russ