Stringer should use a more informative user-agent on feed fetching #278
3 participants
Sounds good to me. Let's just go with Stringer and a link to the project for now.
In the long run it would be nice for part of the user-agent to be configured for each install as people who read their logs want to know not just what software is doing the fetching but who is running it.
I'm not sure about that. e.g. when I use Chrome, it doesn't include "localuser=spraints" in the user agent.
That's a good point. In thinking about it more I agree; Stringer is much more like a browser or standalone feed reader app than it is like, say, a web spider built with some general crawling software (which should identify the party responsible for its operation). So simply identifying the software is both good enough and the right approach.
Yeah I agree about the local user thing - that seems unnecessary. "Stringer (https://github.com/swanson/stringer)" is what I would recommend.
Right now Stringer fetches feeds with the totally uninformative user-agent of 'Stringer' (cf the code in app/tasks/fetch_feed.rb). This is less than ideal and may raise eyebrows of people paying attention to their logs. I think that the feed fetching user-agent should at least include the URL to here.
(A version number of some sort is also traditional, eg 'Stringer/0.x (https://github.com/swanson/stringer)'.)
In the long run it would be nice for part of the user-agent to be configured for each install as people who read their logs want to know not just what software is doing the fetching but who is running it.