liferea
liferea copied to clipboard
lwindolf
Enhance options for feed downloads
https://blogs.oracle.com/linux/rss does not work in Liferea 1.12.2 (on Linux Mint-19.3). Sometimes you get various parsing errors but other times it can totally hang Liferea and you have to kill it! It's not just Liferea, neither curl nor wget just work either.
You can get the feed to work using command curl --cookie-jar /tmp/oracle-cookies --user-agent 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:84.0) Gecko/20100101 Firefox/84.0' --max-time 20 https://blogs.oracle.com/linux/rss. I found both --cookie-jar and --user-agent to be required, but --max-time was optional. I expected to need --referer but didn't.
So I propose adding "use cookies", "max time", and "change user agent" feed download options (see above curl examples). "follow redirections" (curl --location) and "referer" (curl --referer) options would also be nice to have.
Obviously you can already do all of those things using curl but not everyone will know that or think of it, and having handy GUI options would be nice for handling over-complicated and badly designed RSS servers like Oracle's.
To be honest I'm a bit biased against fixing Liferea to be able to access Oracle feeds. I feel with you having to read what Oracle posts :-)
As a workaround please change your feed subscription to a command running exactly your curl statement!
Yeah, I debated about even creating this bug. It took 5 minutes to write the bug and 20 minutes to remove all the OT rants about how much Oracle (the company & the DB) and the Unusable Enterprise Kernel (UEK) suck... I ended up creating this because I figured:
- Publishing the hack-around was not a bad idea, esp. if it might help for something else equally stupid but NOT-Oracle
- Adding some more flexibility to the feed downloader might be worth at least talking about
I'd certainly label this as a low-priority maybe-nice-to-have.
I'm not sure if it's the same issue as OP has/had, but I've noticed Liferea sometimes seemingly not downloading a valid feed, only to confirm that curl does. I'm using version 1.13.5 on Ubuntu 21.04 64-bit.
Is there some debugging option or logging that can help root cause this issue?
@marciof you can run debugging using the command line option --debug-net this provides connection details and possible error codes.
@lwindolf, thank you. Here's what I found out. So Liferea actually does check the feed's URL, but it seems that when it gets a 304 (what I'm seeing now) it still runs the conversion filter, but with no data on stdin. Is this the expected behavior though?
@marciof While it might not be useful to run the filter on a HTTP 304 result, it should not hurt either as a 304 should have no content anyway. I guess if your feed doesn't parse it's caused in an earlier fetch request. If you could share the feed I could try reproducing. Could you add a new ticket for this?
@lwindolf, the feed does parse. I think I was just surprised to see the conversion filter still being invoked on empty content, as I didn't see any mention of this in the documentation.