Make Vienna more flexible in searching articles within certain feeds

[tijmenvangulik]

Describe the bug
I add the following rss feed which seems to be a valid rss feed but vienna can not find any articles

To Reproduce
Add the following feed:
https://www.deingenieur.nl/nieuws/rss/all

In the log you see:

Connecting to https://www.deingenieur.nl/nieuws/rss/all
124 KB received
And message: No articles in feed
When you open the feed in a browser you see that er many news items in de feed

Screenshots
Add screenshots to help explain your problem, if possible.

Please complete the following information:

• Vienna version n 3.8.7 :15608906: (8038)
• macOS 13.2.1

[Eitot]

It is actually not a valid RSS feed. The <item> elements (these are the feed's entries) are supposed to be child elements of the <channel> element. This feed here places the <item> elements outside of the <channel> element, which is why Vienna does not find them.

<rss version="2.0">
    <channel>
        <title>de-ingenieur</title>
        <pubDate>Sun, 23 Apr 2023 13:02:58 +0200</pubDate>
        <copyright>Copyright (c) 2023, De Ingenieur</copyright>
        <language>nl</language>
        <description>De Ingenieur Rich Site Summary</description>
    </channel>
    <link>https://www.deingenieur.nl/nieuws/rss/all</link>
    <item>
        <title>Waterstof tanken bij mobiele pomp</title>
...

[tijmenvangulik]

Thank you for the quick reply. I will take it up with the rss provider. Sorry to bother you with this wrong error report. Tijmen

…

On 23 Apr 2023, at 13:04, Eitot ***@***.***> wrote: It is actually not a valid RSS feed <https://validator.w3.org/feed/check.cgi?url=https%3A%2F%2Fwww.deingenieur.nl%2Fnieuws%2Frss%2Fall>. The <item> elements (these are the feed's entries) are supposed to be child elements of the <channel> element. This feed here places the <item> elements outside of the <channel> element, which is why Vienna does not find them. — Reply to this email directly, view it on GitHub <#1676 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAVG7NRIFZBW6MPJNTLWE2LXCUEDDANCNFSM6AAAAAAXIODX6A>. You are receiving this because you authored the thread.

[Eitot]

Interestingly, Reeder does work with this feed. This raises the question whether Vienna could have more lenient parsing of RSS feeds here, e.g. looking for <item> elements regardless where they are located.

@barijaona What do you think? Feature request?

[barijaona]

I am all in favor of more flexibility.

But there are various popular specifications of RSS which are in fact unrelated, so we'll have to be careful not to break feeds conforming to a specific specification.

[barijaona]

Examples:

• in the example listed for RSS 0.9,<item>s are outside <channel>
• in the example listed for RSS 1.0, <item>s are listed inside <channel> and detailed outside <channel>

Good to know: work on RSS 1.0 and RSS 2.0 were done separately. The later is not the successor of the former.

[Eitot]

It is dubious. All of the sample files on that webpage have <item> elements as child elements of <channel>. The specifications of 0.90, 0.91 and 0.92 do not explicitly say where <item> should be placed, but imply that it is a child element of <channel>: "An item that is associated with a channel" and "There can be no more than 15 in a 0.91 ". Regardless, RSS 2.0 – which is what this feed claims to follow – does not allow this.

However, would it be harmful to look for <item> elements under the <rss> element if no items are found under <channel>? If a particular RSS feed was extended with custom elements, it would be using namespaces (which could be ignored when looking for <item> elements outside of <channel>).