Tiny Tiny RSS: Community

Strange Feeds Issue


#1

Describe the problem you’re having:
Installed latest version and imported OPML. Able to run update_daemon2 and see it working, but some feeds post keep floating to the top, showing as latest… one example is iclarified com. I have about 30 feeds, but all I see in all articles is same posts from the same feed and they are out of order and not latest.

If possible include steps to reproduce the problem:
not really should how to reproduce but here is one of the feeds iclarified com

tt-rss version (including git commit id):
[Tiny Tiny RSS]) v15.7

Platform (i.e. Linux distro, PHP, PostgreSQL, etc) versions:
Ubuntu 14.4
mysql Ver 14.14 Distrib 5.5.62, for debian-linux-gnu (x86_64) using readline 6.3
PHP 5.5.9-1ubuntu4.27

Please provide any additional information below:
Seems that maybe TTR is not picking up the date and sorting it as new ? IS there anyway to have it just mark feeds as it gets them ? I see other feeds are updated but I dont see any posts from them, just same posts over and over from same feed, I tried to change sorting and changed to all items, but still same… Please let me know what Im doing wrong…


#2

how about you update to latest version first? i’m not going to dig up something from 2015 because of you.

also if you’re posting about feed-related issues it’s probably a good idea to post problematic feed URLs.

gonna need to upgrade first though


#3

Wait thats what I pulled from git I will check again. Also discourse is not letting put full urls, if you look in text it is there … I will upgrade first tho


#4

that’s new user anti-spam limitation, just post without http:// or something.


#5

what you mean iclarified.com? because that’s not a valid feed URL.

e: look here My feed doesn't parse properly, or does it?


#6

Its not letting me post the full URL thats what I said befor, thats why if you go to the site you can see there feed, I will try to post it again Iclarified com / rss/ rss.xml I did the parser thing, and it shows up fine. Let me update and Ill get back to you.


#7

Ok, Updated everything to the latest versions. [Tiny Tiny RSS] v19.2 php -v
PHP 5.6.40 Server version: 10.1.38-MariaDB MariaDB Server

But feeds from Iclarified com / rss/ rss.xml are killing the feeds, they are the only once floating up, I`m trying to do newest first…

Please help!


#8

Also noticed basicly the same thing on www. buzzfeed. com/ index.xml its like every time it pulls them it brings them to the top, and it doesnt matter how I sort, with Newest or default.


#9

if you look at myfeedsucks output you’ll see that items of iclarified feed have no dates, in this case tt-rss assigns it (current timestamp at the time of import) - so new articles from this feed are always going to be on “top” and not sorted correctly.

i’m not sure i understand your problem though. you should be able to just mark this feed as read and go on to your other feeds.


#10

This is not the only feed this its doing this too, So you saying it keepa repulling the same posts from same feeds over and over and adding them to the main feed ? Why isnt it marking it as already gotten and leave it in time line as it first downloaded.

I usually have read and unread toghether as I go back to them or thats how I know where Im.

I did look at myfeedsucks, and while the output does have pubdate I dont see your tool picking it up. Im sure they formatted it wrong or something, but other service always put them in the right order and doesnt bring them up. I have another feed that I posted yesterday that shows date in myfeedsucks, but TT-RSS still keeps repulling them and putting them on top, www. buzzfeed. com index.xml

Basically in feeds I need a time line with Newest First (based on pub date ? Or when post first came in to TT-RSS, otherways I just keep seeing same feeds over and over in different order and not just from iclarify.


#11

no, i’m not saying that. articles are not identified by their date (or lack thereof).

i still don’t understand what’s going on with your tt-rss install. do duplicate articles appear or do they stay unread or do they become unread again or does something else happen. maybe you should explain this in more detail, with screenshots or something.

check for any third party plugins being enabled, maybe reset tt-rss settings. who know what might be wrong with your configuration.

try using feed debugger (f D) on iclarified feed to see if articles are properly detected as read, output should look similar to this:

[12:43:08/20340] guid 1,https://www.iClarified.com/69785/castro-podcast-app-gets-discovery-section-live-search-faster-subscribing / SHA1:4ddb3410e751fc13af977e37bd22bd25972acdef
[12:43:08/20340] orig date: 
[12:43:08/20340] date 1552048988 [2019/03/08 12:43:08]
[12:43:08/20340] title Castro Podcast App Gets Discovery Section, Live Search, Faster Subscribing
[12:43:08/20340] link https://www.iClarified.com/69785/castro-podcast-app-gets-discovery-section-live-search-faster-subscribing
[12:43:08/20340] language 
[12:43:08/20340] author 
[12:43:08/20340] num_comments: 0
[12:43:08/20340] looking for tags...
[12:43:08/20340] tags found: 
[12:43:08/20340] done collecting data.
[12:43:08/20340] article hash: c54e483b7e9ea33843543282043edfc845d3bf71 [stored=c54e483b7e9ea33843543282043edfc845d3bf71]
[12:43:08/20340] stored article seems up to date [IID: 954028], updating timestamp only

“stored article seems up to date” is the key part. don’t check “force rehash” in feed debugger.

on a semi-related note i still don’t understand how a first time user of tt-rss managed to install code from 2015 on an ubuntu server from 2014. unless you’re time traveling or something.

point is, you’re reporting weird issues (which i wasn’t able to quickly replicate) on a weird configuration after making some weird decisions. unless someone else can corroborate any of this, i’m not going to waste a lot of time figuring out what’s wrong with your install.

<pubDate>Fri, 08 Mar 2019 03:51:28 UT</pubDate>

that isn’t a valid timestamp, because UT is not a valid timezone (their feed generator likely ate last character of UTC). these iclarified people need to fix their feed.

tt-rss is, out of the box, quite strict about feed parsing and data formats it accepts as valid. if that is unacceptable to you, use something else. :man_shrugging:


#12

Hello,

I would like to address the 2015 and old version on Ubuntu. I have a old VPS which is running old LAMP stack on 14.04, it was a quick way for me to try TT-RSS, and I cloned directly from git which I guess is no longer maintained, I would disregard the original setup, as I now deployed this on a brand new VPS with updated fresh LEMP stack on 16,04 and this is the only thing that I have on this server, nothing else. TT-RSS is a fresh clone from your repository… No settings are changed, except for the password, view set to all, and newest first, Im looking in "All Articles", Im manually running update_daemon2.php from command line. And no extra plugins are enabled, I`m viewing in incognito mode with out any chrome plugins. I imported about 30 feeds that I have running on other services…

Here is whats happening, basically all I see in the feed is iclarified and they are out of order as well, like right now on top of the feed is a post from March 4th.


(its not letting me put second picture in post, but it from the site showing date of Posted March 4, 2019 at 6:14pm

I still need to see posts in the feed that are marked as read and also the out of order posts from the feeds shouldnt happen, even if I remove iclarifed.

I`m not saying I want to use other service, I want to make this work, I understand its strict, but as posts from feeds come in, there is no way to put a time stamp on the posts from the feed as they come in(meaning when TT-RSS sees the post it time stamps it, or puts in the proper order ?)

here is debug output

Blockquote
[13:37:36/22721] guid 1,hXXps:// www .iClarified. com /69724 /new-airpods-to-feature-super-fast-wireless-charging-new-surface-coating-bigger-case / SHA1:174e04f97068082dd533201aca467a1e4c6bd1a7
[13:37:36/22721] orig date:
[13:37:36/22721] date 1552052256 [2019/03/08 13:37:36]
[13:37:36/22721] title New AirPods to Feature Super Fast Wireless Charging, New Surface Coating, Bigger Case?
[13:37:36/22721] link www. iClarified. com /69724 /new-airpods-to-feature-super-fast-wireless-charging-new-surface-coating-bigger-case
[13:37:36/22721] language
[13:37:36/22721] author
[13:37:36/22721] num_comments: 0
[13:37:36/22721] looking for tags…
[13:37:36/22721] tags found:
[13:37:36/22721] done collecting data.
[13:37:36/22721] article hash: 89d78cc28f4f497679c7b4855b33223743151330 [stored=89d78cc28f4f497679c7b4855b33223743151330]
[13:37:36/22721] stored article seems up to date [IID: 2906], updating timestamp only
[13:37:36/22721] purging feed…
[13:37:36/22721] Purged feed 32 (60): deleted 0 articles
[13:37:36/22721] done

My goal, is to have all feeds be in Chronological, ideally and again, I`m new to this, I would want TT-RSS to mark the feeds with internal time stamp as it got it and present it in the feed that way, if the feed stamp cant be used, if I can modify something, or need to create something to make it so please let me know !

Also, @fox the work you have done on this project is outstanding, ti prosto molodec ! Its an amazing piece of work and a brilliant setup.


#13

@fox all this might be useless, I htink I know whats happening… Its the cloudflare cache… I`m pretty sure…


#14

alright with your explanation the vps thing makes a lot more sense

unfortunately if timestamp is invalid there’s no way to ensure sorting because the order in which elements appear (and are iterated through) in feed XML is not guaranteed to match logical post order.

tt-rss tries its best to provide logical sorting but it’s not always possible, also it doesn’t work very well on initial pull when you add many articles at the same time. it gets better from then on because batch timestamps guarantee that articles added in subsequent feed updates never appear before previous ones.

try searching “batch timestamp” here and on old forums, i remember explaining in detail how this works multiple times.

well it seems that this works as it should

since you’re using All articles (instead of adaptive) you’re going to see articles from broken feeds at the top at least for some time until several more feed updates happen and new articles replace them.

i suggest using adaptive so you can run through actual unread content and then switch to all articles if you need to see something older.