Problem with Tim Bray's ongoing blog

[abatkin]

I have been using tt-rss for the past few versions (I'm on 1.8 now) but have found that it seems to have trouble correctly parsing Tim Bray's ongoing blog which is in Atom format.

There are actually two issues, which I suspect are unrelated. I may be able to take a look into this at some point, but figured I'd see if anyone else already knew what was wrong or had looked in to it first.

The first issue is that the links back to posts are all wrong. For example, the post "Two Factor, Twice" should link to https://www.tbray.org/ongoing/When/201x ... Two-Factor but in in tt-rss the link points to http://www.tbray.org/ongoing/Two-Factor. I think the problem is that the xml:base attribute (which in the case of this post contains "When/201x/2013/06/24/") is not being honored.

The second issue is that the body is never rendered properly. Actually, looking back, it seems that now the body of each post isn't rendered at all, at least for the past few months. Maybe this is some sort of HTML sanitization issue? His posts are all "content type='xhtml'"

Ideas? Thanks!

[fox]

Best part is that you never bothered to post the actual feed url, nevermind trying on latest trunk.

[abatkin]

Sorry, minor oversight (I thought he was notable enough that you would know what I was talking about). His blog is https://www.tbray.org/ongoing/ and the feed URL is https://www.tbray.org/ongoing/ongoing.atom

I haven't tried on the latest trunk because the bug has been there for the past few tt-rss releases so I have no reason to believe that it would have suddenly been fixed in trunk. But of course I could be wrong, and if that's the case I can set up a test site to try it out and report back.

[fox]

>Sorry, minor oversight (I thought he was notable enough that you would know what I was talking about).

Yes obviously everyone on the planet knows and cares about stupid blogs you read. Get over yourself.

>I haven't tried on the latest trunk because the bug has been there for the past few tt-rss releases so I have no reason to believe that it would have suddenly been fixed in trunk.

Are you seriously so socially inept as to not realize that it is basic politeness to check if your problem may be already fixed before reporting it? The underlying logic of "it hasn't been fixed before therefore it won't be fixed ever" is completely beyond me either.

[sleeper_service]

Sorry, minor oversight (I thought he was notable enough that you would know what I was talking about). His blog is https://www.tbray.org/ongoing/ and the feed URL is https://www.tbray.org/ongoing/ongoing.atom

I haven't tried on the latest trunk because the bug has been there for the past few tt-rss releases so I have no reason to believe that it would have suddenly been fixed in trunk. But of course I could be wrong, and if that's the case I can set up a test site to try it out and report back.

what's a tbray? some rabid horse? and why should I care.... scratch that, I don't care.

and, forget wasting time setting up a 'test site', just upgrade to trunk and enjoy the newly fixed things that you're doing without now.

[timbray]

I’m pretty sure the feed is correct, but I’m also aware that it causes problems for lots of readers. It makes heavy use of relative URIs, check out all the xml:base values and so on. Since I helped design the Atom format, I deliberately used the constructs in the feed to explore corner cases.

[abatkin]

Awesome celebrity cameo appearance.

Running from the master branch appears to resolve the xhtml content issues.

The links are all still broken though (this affects both links back to posts themselves, as well as to any relative URLs within the (x)html content, like img or a tags). I've been looking through the source, but I think fixing this may be more difficult than I can handle myself. The issue is that (from what I have read) xml:base can appear almost anywhere, and they "stack", so for every link, you would need to walk up the DOM, pushing any xml:bases onto a stack, then popping them back off once you are done, to assemble the "final" URL. Even for the actual (x)html content, you would at least have to (potentially) build up an xml:base for the root of that content before displaying it (at least there you don't have to actually parse the markup, since you would expect the browser to handle anything in there).

Thoughts?

[fox]

Yeah xml:base is not supported, as far as I'm concerned this is too much absurd atom fuckery to deal with. I'm definitely not losing any sleep over it and sperglords who want to save a few bytes so much.

[ICE]

Hi Fox,
RSS of rutracker.org also uses xml:base.
So I'm also interested about this functionality.
As far as I understood function Article::create_published_article is responsible for creation articles.
It uses $url as variable with article URL.
I've found occurrences of this method few times through the code (shareToPublished and sharepopup).
But Its not easy to found out the right place that must be corrected.
I know PHP so-so ans I want to help you with this issue.

Could you please give me some details: where exactly is it good to correct the problem and where can I get raw atom content (I can try to found out it myself my I believe it is better to ask author as you are the one who better then others knows architecture).

Thank you.

[fox]

Please post in English. If you don't know English, stop posting.

[ICE]

1) What is the problem with my English?
2) Don't you need some help?

[fox]

>1) What is the problem with my English?

The problem is that you can't speak it.

>2) Don't you need some help?

See above.

[blainemono]

hello niko this is your brother roman i need help with downloading american imperialist torrents also bowling

[ICE]

>>The problem is that you can't speak it.
Sure, if you prefer we can talk about grammar here, but anyway please comment your complain.

And sure it is very rough to talk like you did (see above).
Please be more polite, it is not an english classes here.

[blainemono]

hello nico roman is getting upset