Quote:
Originally Posted by JimmXinu
First, great job finding that debug code. I didn't know trick and it gives us some important details. 
When you two do it, you are getting redirected by the web server to the domain name 'unicorn_story' which clearly does not exist (no TLD).
When I do it from my home network (and from my server in the cloud), I get redirected to the same URL path, but at archiveofourown.org. Also, your redirects are to https and mine to http--I don't know why that would be.
Also, I'm seeing some different headers:
Code:
header: Server: nginx/1.6.2
header: Location: http://archiveofourown.org/works/11656197/chapters/26225727
header: X-Aooo-Debug: The Lion
header: X-Hostname: ao3-front02
header: X-Page-Speed: 1.9.32.3-4448
While you are seeing:
Code:
header: Server: nginx/1.13.7
header: Location: https://unicorn_story/works/11656197/chapters/26225727
header: X-Aooo-Debug1: Archive Unicorn
header: X-Hostname: ao3-front04
header: X-Page-Speed: 1.13.35.1-0
So I think you are hitting a different set of servers that are misconfigured. Maybe because of network optimization that sends us each to a 'closer' server, maybe because they are upgrading--your debug is showing higher version numbers.
What I can't explain as easily is why it would continue to work in your browser. They may be differentiating by user-agent, but another possibility is they are deliberately sending you two in particular in a different direction. That could happen if you've been doing a ton of downloads and tripped some blocking behavior.
Here are a few suggestions: - You could report the problem to AO3 and see if they can correct it. If it were me, I'd send them the python debug output as proof to start the conversation. It's much more informative than if you start with 'FanFicFare isn't working for me on AO3'.
- Your browser may continue to work because of user-agent specific handling. You could try configuring FFF to use the same user-agent as your browser. In personal.ini set user_agent:--sites like http://getright.com/useragent.html can tell your browser's user-agent.
- You could download using the FFF web service until the problem is resolved. That has its own issues as the web service has daily resource limits that can be exhausted and can't populate your Calibre columns.
- AO3 is one of the sites that offers ebook formats directly for download. You can download their EPUB files and use those instead of using FFF.
|
Hahah, I did change the user-agent and it worked! Yay! And then I tried running the python again and it worked too. So I'm thinking your idea about us getting routed to some broken hosts might be correct. Either way, it's working now, and maybe the user-agent will keep it working in the future.
Thanks so much for looking into this and responding with suggestions.