Error: Failed to load url - has no information

[kirillrogovoy]

My original problem description is here.

What I found is that it's possible to detect the exact error using the onResourceError callback which gets fired before the onLoadFinished. So we can remember the last error and show it if onLoadFinished returns fail status.

I didn't do it myself because I feel that using such callbacks in the core might violate some rules or good practices of the project. Please, let me know what you think about the idea, so I can make a PR or just do it yourself if you feel like it.

[awlayton]

Sounds feasible maybe. You would need to make sure you don't mix up the onLoadFinished of one page and the onResourceError of another page.

I doubt I will have time to do it myself any time soon.

Any thoughts @johntitus?

[nevmerzhitsky]

I also have this problem, please push solution!

[awlayton]

What problem do you have @nevmerzhitsky? This issue is a feature request, not a bug report.

[johntitus]

@nevmerzhitsky and @flashhhh, can you try setting the horseman timeout to something big, like 30 seconds, and retrying this?

[kirillrogovoy]

Hi, @johntitus!

Will try this a bit later. Is upgrading the package to the latest version required?
If not, I'm sure to say we always used timeouts greater then 30 seconds.

Our current working version is 3.0.1.

[johntitus]

If you don't set the timeout in the Horseman constructor, it defaults to just 5 seconds.

It looks like that if a timeout occurs, the first thing that happens is the resource requested fails, which is the error that shows up, instead of a timeout.

I do recommend upgrading before testing, just to ensure we're all on the same page.

[kirillrogovoy]

Rechecked. As I said, we already set a large timeout value (30 sec currently). I tested with greater values (like 60) and nothing changed.

[Its-its]

Still persists for me even when settings timeout to 30 seconds. Has to be something with loading the page, it error's within the first couple seconds of it trying to load a page for me over https with the 30 second timeout. So the large timeout is more-so redundant than anything with this bug.