nextcloud / news

:newspaper: RSS/Atom feed reader
https://apps.nextcloud.com/apps/news
GNU Affero General Public License v3.0
855 stars 182 forks source link

Read articles haunting me #2717

Open erebion opened 1 month ago

erebion commented 1 month ago

IMPORTANT

As suggested by @Grotax in #1908, I'm opening a new issue.

While typing the issue, I realised I am on an alpha version. But maybe you will still find the report helpful. If not, please just close. But maybe it helps somehow. :)

Explain the Problem

What problem did you encounter?

Steps to Reproduce

Explain what you did to encounter the issue

  1. Mark article as read
  2. After a while (no specific amount of time) it will be back as unread

System Information

Contents of nextcloud/data/nextcloud.log ```json Paste output here ```
Contents of Browser Error Console Read http://ggnome.com/wiki/Using_The_Browser_Error_Console if you are unsure what to put here ``` Paste output here ```

don't have access right now, can add this later

SMillerDev commented 1 month ago

What do you have your cleanup threshold set to in the admin settings?

erebion commented 1 month ago

Set to 200.

Also:

Deleting unread articles automatically: off Max Redirects: 10 Timeout: 60s Interval for updating: 3600s

SMillerDev commented 1 month ago

And roughly how many articles do you have?

erebion commented 1 month ago

And roughly how many articles do you have?

Differs per feed. None are extremely large, the largest have a number in the lower hundreds. Most have less than a hundred posts. I use aroudn twenty feeds.

Articles come back in many feeds. For example the Debian project's blog. Suddenly an article from 14 years ago is shown as "new".

SMillerDev commented 1 month ago

Most have less than a hundred posts. I use aroudn twenty feeds.

So the total amount of articles is more than 200, which means that after every cleanup it'll pull in the old articles again, since only the newest 200 remain.

erebion commented 1 month ago

So the total amount of articles is more than 200, which means that after every cleanup it'll pull in the old articles again, since only the newest 200 remain.

If that's what the setting does, then the wording is not clear on what the implications are.

It reads like "We will delete every articles more than 200".

It does not read like "We will haunt you with every arcticle more than 200 and show old ones as new".

If that is what this setting is meant to do, then I'd like to suggest changing the wording to something that makes this clear.

Either way, I find it weird that it would get articles older then the most current 200 articles, from years ago, and then show them as new. Doesn't it just save "Everything older than this is too old"?

erebion commented 1 month ago

Also, I just realised that the German version states the opposite:

Setting:

Defines the maximum amount of articles that can be read per feed which will not be deleted by the cleanup job; if old articles reappear after being read, increase this value; negative values such as -1 will turn this feature off.

It therefore states:

if old articles reappear after being read

German version states:

wenn ältere Artikel nach dem Lesen erhalten bleiben sollen

Translated back into English:

if older articles are to be kept after reading

Which changes the meaning completely and explains my utter confusion.

erebion commented 1 month ago

Where can I work on the translation? Transifex seems to only get used for Nextcloud itself, but not Nextcloud News, right?

Grotax commented 1 month ago

The algorithm for cleanup works actually a bit differently than just keeping 200 read items or whatever the value is.

News fetches the feed and then checks

If the feed has less than 200 items in the feed file then News will keep 200 read items and whatever amount of unread items.

If a feed has more than 200 items then we take whatever that number is as limit, if I'm not wrong that is also stored in the DB.

But there are cases where this might still not work. Also some feeds behave a bit strange because of the way they are generated.

erebion commented 1 month ago

Sure, though I'd still like to help out with the translation. Having a translation that pretty much states the opposite does not help.

Grotax commented 1 month ago

In our documentation this is also explained https://nextcloud.github.io/news/admin/

Transifex is also used for the news app.

It might seem logical to also filter items based on date like when we see an item that is much older than the oldest we have in the db then we don't add it to the db. But the issue with that is that it's quite common for feeds to have the wildest date combinations. Or feeds that work based on most viewed or other metrics where an old item can suddenly come back to the feed.

In your case I would just adjust the value to a higher number and see if that works for you.

erebion commented 1 month ago

Transifex is also used for the news app.

Would you be so kind to let me know where I can find the project?

In your case I would just adjust the value to a higher number and see if that works for you.

I've just disabled cleanup completely. It's just a bit of text anyway.

erebion commented 1 month ago

Setting it to -1 does not resolve the issue, articles still come back.

erebion commented 1 month ago

Will I have to wait for all articles to come back, mark them as read and then be done, not just by setting this to -1?

erebion commented 1 month ago

Setting it to -1 does not resolve the issue, articles still come back.

It seems that Nextcloud News does not store changes in the settings.

I've now changed it with

# sudo -u www-data php -f occ config:app:set --value "\-1" --update-only  news autoPurgeCount
Config value 'autoPurgeCount' for app 'news' is now set to '\-1', stored as mixed in fast cache

Now the UI just shows it set to \-1, but due to https://github.com/nextcloud/server/issues/26109, I cannot set it to -1 without a backslash. But maybe it interprets something that looks invalid as having that disabled..?

Grotax commented 3 weeks ago

Yea the issue is that the Admin UI is not able to update the fields anymore: https://github.com/nextcloud/news/issues/2533

-1 will probably lead to an error but I'm not sure.

I'm sorry that you are facing these issues, the problem is that news has not nearly enough (developer) support for what it tries to be...

erebion commented 3 weeks ago

Can I manually set that somewhere? Maybe in the database?

Grotax commented 2 weeks ago

Yes you could update the setting in the database. Nextcloud has an app config table.

But I don't know the details at this time so you have to check.