search

nextcloud / news

📰 RSS/Atom feed reader
https://apps.nextcloud.com/apps/news
GNU Affero General Public License v3.0
870 stars 185 forks source link

Article search does not work / not pops up #1613

Closed gsantner closed 2 years ago

gsantner commented 2 years ago

IMPORTANT

Read and tick the following checkbox after you have created the issue or place an x inside the brackets ;)

Explain the Problem

The keyboard shortcuts help explicitly mentions the keyboard shortcut q to pop up some kind of article search.
When I press the key q, nothing happens. At all.

Screenshot_20211212-133142

General nextcloud search doesn't find anything either.

Screenshot_20211212-133228

Steps to Reproduce

Explain what you did to encounter the issue

  1. Open nextcloud news in browser
  2. Press the key q according to documentation
  3. nothing

System Information

SMillerDev commented 2 years ago

Yeah, the documentation is wrong here. Article search doesn't exist anymore.

gsantner commented 2 years ago

Article search doesn't exist anymore.

Why that? That's one core feature of news readers

SMillerDev commented 2 years ago

Because it doesn't work well within the nextcloud system. And you're only the second user to mention this, so clearly there is not as much demand as you'd think.

Grotax commented 2 years ago

Nextcloud changed how searching works and moved to this "unifed search" which required the apps to reimplement search. So far we only have search for folders and feeds but not for the feed items

gsantner commented 2 years ago

Generally, is there some active work going on in the project?
It feels to me at most issues/PR's I look at nextcloud (be it server "app" or android app) - low interest in making it better, fix it on your own, use something else, or ignore it.

Know that behavior well from other projects, but when that kind of maintenance & routine mode kicks in it's often waste of time for self, for users hoping that it works as expected, and for users reporting the issues.

As far I can tell from a few days of using this nextcloud news:

Web frontend

web frontend is buggy and slow -

https://user-images.githubusercontent.com/6735650/145715219-9be1b83a-916f-4469-89f8-331374f97235.mp4

I see the intentions and PR for replacing the framework, but seems rather incomplete and not like it happens any time soon.

API & Clients

The API is vendor-specific just so it fits the project own needs. There is APIv1 - OK, the old and outdated one. But APIv2 is a draft since many years and still not implemented - the PR looks like only stub code to me.

Why not get rid of the v2 draft completely as it anyway won't happen - instead make a (i.e. nearly-fever compatible) v3 and actually implement it? There are a few de-facto-standard protocols for RSS Readers, most notably Fever & Open/Google_Reader.
What I get is that neither you maintainers/core members nor your contributors want to implement a vendor-specific ~5 year old API draft that didn't made it so far.
It doesn't require years of work to implement some GET article-details, GET article-array and export some database values as json/xml..

grafik

I read through most of the comments here and on the android app, and my main findings are :::authentication not compatible with nextcloud:::: or ::they should implement our brew:::. However clients can more easy adapt to some special authentication method, rather than implementing a whole own API. There are not much applications that are compatible to NextCloud News API - even less which do work & do sync. For OpenReader and Fever there is broad support.

I tried all NC-News Android & Linux-desktop clients I could find, all of them either cannot login, or don't sync properly (not even the official nextcloud news android app). I thought about helping some well-working rss-reader application with sync capabilities to make it work with nextcloud - but neither want them to waste time with it, nor mine to implement&maintain a whole new API which might be anyway replaced & is unique to nextcloud. (i.e. news flash)

TL;DR: Am I wasting my time here with reporting issues-bugs & requesting (re-addition) of important features - because anyway nothing will happen/improve? And don't be mad about this comment - I want to help you, and I want to know if you even care about getting help.

Grotax commented 2 years ago

Well I understand your frustration, thing is this is an open source project with not much resources.

I for example am the maintainer and the holder of the key to release the app. But I didn't take this project because I'm a PHP pro or a fan of JavaScript in fact I'm not a big fan of either.

But I wanted to keep news alive to keep my favourite app going.

I have been using news for many years mostly together with the android app and I'm very happy to have this solution, instead of paying some company to analyse my reading behaviour.

Developing a Nextcloud app is not easy or comfortable in my view, till this day my dev setup is annoying and regarding some of the issues there is simply no solution. Not understanding the full architecture doesn't help either.

So yea if you expect things to change because you report them, I have to say that will probably not happen. It's simply not motivating enough to do something, at least for me.

The few things I contributed inside of the news app were always difficult and annoying to do. And no I'm not a professional full stack web dev.

Grotax commented 2 years ago

Why not get rid of the v2 draft completely as it anyway won't happen - instead make a (i.e. nearly-fever compatible) v3 and actually implement it?

I would love to do something like that but it's just too difficult and frustrating for me, I don't have the energy and knowledge to implement such thing.

Grotax commented 2 years ago

What I get is that neither you maintainers/core members nor your contributors want to implement a vendor-specific ~5 year old API draft that didn't made it so far.

Yea the problem is that there are in fact not many people helping here, there is me who got assigned to be the key owner. Then there is another core member, the rest comes and goes. Which is totally fine but doesn't make this project very effective.

I try my best to reply here and answer people in a polite way, help with git issues and so on. And I hope that some day somone that is motivated and skilled shows up and fixes lots of the open points.

SMillerDev commented 2 years ago

Generally, is there some active work going on in the project?

Yes, but since nobody is getting paid everyone works on whatever they want.

web frontend is buggy and slow

Yes it is, but the only PR that managed to improve that broke before we could release it.

The API is vendor-specific just so it fits the project own needs. There is APIv1 - OK, the old and outdated one. But APIv2 is a draft since many years and still not implemented - the PR looks like only stub code to me.

The folder part of the implementation is already done, but again in people's free time. So we can't put a timeline on that.

Why not get rid of the v2 draft completely as it anyway won't happen - instead make a (i.e. nearly-fever compatible) v3 and actually implement it?

You're welcome to do that. I'm happy with the current implementation but pull requests are welcome.

There are a few de-facto-standard protocols for RSS Readers, most notably Fever & Open/Google_Reader.

That unfortunately does not work in the Nextcloud API framework.

However clients can more easy adapt to some special authentication method, rather than implementing a whole own API.

In my experience if you use a standard API but change some key component it actually becomes harder to implement than a whole new API.

I tried all NC-News Android & Linux-desktop clients I could find, all of them either cannot login, or don't sync properly (not even the official nextcloud news android app).

That sounds like a curiosity in your server setup because all apps that are listed in the documentation are supposed to work just fine.

SMillerDev commented 2 years ago

TL;DR: Am I wasting my time here with reporting issues-bugs & requesting (re-addition) of important features - because anyway nothing will happen/improve?

I'd say yes. Unless you can motivate someone else to implement your requests they will go stale and will be closed.

If you find actual bugs in the API I'll look at them, but anything else is likely not going to be fixed or implemented.

gsantner commented 2 years ago

Thank your for your responses, much appreciated.

Well I understand your frustration, thing is this is an open source project with not much resources. Yes, but since nobody is getting paid everyone works on whatever they want.

Know very well how it works - I'm developing (&maintaining) quite a bunch open source projects in my free time as well.

I have been using news for many years mostly together with the android app and I'm very happy to have this solution, instead of paying some company to analyse my reading behaviour.

That also one of my points - just used a different self-hosted solution before (TT-RSS) - for about 10 years now.

But I wanted to keep news alive to keep my favourite app going. I try my best to reply here and answer people in a polite way, help with git issues and so on. And I hope that some day somone that is motivated and skilled shows up and fixes lots of the open points.

I really appreciated your feedback and understand why you are doing it. The best personal advice I can give is to not waste time further with and move on. I got also into such projects where the initial creator plus some other people where, and I helped with quite some things. Now I'm alone there.

That sounds like a curiosity in your server setup because all apps that are listed in the documentation are supposed to work just fine.

Everything works pretty OK (performance wise "as slow" (or "as fast") as any own/nextcloud instance I used so far. Other syncs like contacts & calendars don't have such sync issues. Nextcloud is mostly in the same network, so no connection issue.
Lets say I scroll through a feed in the android app and mark all of the articles read. Then go in with the next 3 or so feeds. Then refresh (pull-up, or it autosynced by time interval) to see new articles. And then the ~50 articles of the first mentioned feed show up again - as unread with unread counter. Same with rssguard - I mark some things read on rssguard. Click sync, also keep it open in background. Then later somewhen open the web or mobile application on another device, just to see the articles are still there & marked unread.

Grotax commented 2 years ago

Yea thanks for the advice, for me it's actually not that big of an issue that the app is not working for everybody and not perfect.

I wanted this solution to keep going and to stay in the Nextcloud community. All free software projects I know that are not backed up by big money have these kind of issues. And news is really a harmless problem if I compare it to something like openssl or gpg.

And I think there are actually quite a lot of people that are happy that we keep publishing new versions and improve the app over time.

Because else they would need to host another service, which maybe they would not even be capable of, which would leed them to feedly I guess.

If this doesn't fit somones expectations, they are free to use another app, after all this is free software.

SMillerDev commented 2 years ago

My question though is - why are you both "riding a dead horse" then, wasting your time with it?

Because it's not wasting time. We're keeping it up to date and it's dependencies secure. And whenever I feel like adding a feature I add it, it just generally isn't a feature that users requested because I am very rarely motivated to do those.

If you're unhappy with the current state of the app, you likely won't be any happier over the next couple of years. But judging by the downloads a lot of people are happy and will probably stay happy over the years because news will keep working.

As for the syncing issues, unless you can reproduce them in the web app, I don't think they are problems that we can fix. As a frequent user of both a 3rd party (fiery feeds) and the web UI I have never seen this issue occur.