Add pending logging #5
Conversation
…_idle' fails and raises the error
…ait_for_idle' hit's TimeoutError
| "because something took a very long time (for example a page load " \ | ||
| "was slow). If so, setting the :timeout option to a higher value might " \ | ||
| "help." | ||
| message = "#{message}\n Connections still pending:\n #{pending_urls.join(', ')}" unless pending_urls.empty? |
There was a problem hiding this comment.
This isn't a critique of your code here (because I use unless a lot) - but I read this the other day:
Read This Post 'Unless' You're Not A Ruby Developer
There was a problem hiding this comment.
Do you want to override the whole message string, or just append to it if there are pending urls? (it's replacing it atm)
There was a problem hiding this comment.
Maybe something like: (I think that's how you add to a string.. )
| message = "#{message}\n Connections still pending:\n #{pending_urls.join(', ')}" unless pending_urls.empty? | |
| if !pending_urls.empty? | |
| message =+ "#{message}\n Connections still pending:\n #{pending_urls.join(', ')}" unless pending_urls.empty? | |
| end |
| if Utils::ElapsedTime.timeout?(start, timeout) | ||
| if @page.browser.options.pending_connection_errors | ||
| pendings = traffic.select(&:pending?) | ||
| pending_urls = pendings.map(&:url).compact |
There was a problem hiding this comment.
Does the pending connection always have a url here? I thought we saw that sometimes it was empty, because the pendings are a collection of Network Exchange objects, from which the url method is request&.url. I could be wrong here - and it could also be that when we tested last time the exchange didn't have a request (and therefore no url) but when it's run normally they do.
| if @page.browser.options.pending_connection_errors | ||
| pendings = traffic.select(&:pending?) | ||
| pending_urls = pendings.map(&:url).compact | ||
| puts("[DEBUG] 'wait_for_idle' pending connections:\n#{pendings.map(&:inspect)}") |
|
In terms of shipping this PR though - would recommend just keeping it on a branch, then referencing that branch in freeagent, keeping the main/master branch the same as upstream 👍 |
|
Pointing this to |
|
Example of this being triggered (stress test run): Backtrace doesn't show the In this example the failure comes down to |
What
After reverting this fork back to the upstream
0.13.0tag, we want to explicitly emit pending connections when we raise aTimeoutErrorfromwait_for_idle.There are three places that
TimeoutErrorgets raised, each has it's own context, and so will need custom error handling if we want to get specific useful information out:How
Deliberately triggering a failure in a spec:
Example output:
We're also dumping as much of the pending context as possible: