kelson

macmanx, I guess this didn’t come out right (chalk one up to incomplete revision), but I’m on your side. FAQs, searches — they save everyone time. The idea that saying “you can find this by searching for xyz” or “this is in the FAQ at xyz” is somehow unhelpful and even rude (in and of itself) seems shortsighted at best.
But then we do live in the era of instant gratification.

I’ve seen the PCUser plague as well, and seriously considered blocking on the UA string. (Fortunately, installing a slightly modified three strikes plugin has blocked the current run.) A search through my site’s logs for October for “PCUser” — nearly all were (attempted) posts from this scumbag.
Nearly all.
I did find three visits — all early last week — that looked like normal web use patterns: each came in from a Google search, two were single-page visits while the third looked around at a half dozen pages, and they hit all the associated scripts, images, stylesheets, etc. They were also all hits to other, non-WP areas of my site. So it looks like *someone* has that UA, but it’s very rare.
I’ll be renaming wp-comments-post.php to be a bit more proactive. I was going to suggest an option to randomly rename it as part of the install process, but realized that would be trivially defeated since the bots are obviously pre-harvesting the URLs (based on the presence of referrers in the hits). They could as easily harvest the script location as part of the process.
(As for recommending that people search the archives… what do you think FAQs are for? As long as it’s suggested politely, it saves everyone more time in the long run, including the person asking the question. Give a man a fish vs. teach a man to fish, as the old proverb goes.)

It’s not just a nit: SharpReader refuses to process feeds with the invalid header! Instead it reports “Error reading URL: The underlying connection was closed: The server committed an HTTP protocol violation. ” (I just spent a ridiculous amount of time tracking the problem down to this typo.)
This appears to be fixed in CVS, but broken in 1.2.1.

This is also an accessibility issue. When there is a primary language for a document (be it HTML, RSS, or anything else), it’s advisable to indicate that for the benefit of screen readers, search engines, aggregators, etc.
OK, so it’s an easy one-line addition to each of several files. But most people publishing with WP probably have never even read the RSS spec, probably don’t even know that this tag exists. They aren’t going to add it without a massive education campaign like the Web Standards Project or Mozilla evangelism projects. Even then, it’ll only be those comfortable with hacking PHP and/or XML. For many, the whole purpose of using a publishing tool like WordPress is so they don’t have to worry about coding!
Alternatively, a small, easy change can be made in the authoring tool that will make their feeds easier to index and thus easier to find. IMO it doesn’t even need a UI, as long as the language is stored somewhere. It would be one of those sensible defaults.
As for “bad or false” metadata — I don’t think anyone here is advocating that. If the feed contains multiple languages, it seems to me that would be the case for leaving out the tag.
The only reason I can see that this feature would not be easy to add is if WordPress does not know what language it is being used in. I have been assuming that, with the localization push, WP would have this stored somewhere. If not, please correct me.

On a side-note, it seems the RSS generator for this forum isn’t escaping HTML/XML characters in the title element, breaking the feed for this thread. (everything looks like <title>www.ads-software.com thread: <language> tag required in rss</title>)

s/required/recommended/
It may not be required, but it would certainly be useful, particularly for feed discovery, accessibility, etc. If WP already stores the (default) language somewhere, it should be easy to add the tag. If not…, well, it probably should.
As for multi-language blogs, the RSS spec states that <language> only applies to the channel element, not to individual items, so that would be a case where it would be better to leave the language tag out or to pick the more-frequently-used language (if there is one).

See also bug 367, which I submitted last night.
Looking at the code for WP 1.3 — which removes the rss_language setting from the database as part of the upgrade process — I suspect the setting was left over from an earlier version, with the GUI removed in 1.2 and the option itself removed in 1.3.
I’m hoping, though I couldn’t find it, that there’s a general language setting somewhere in the current code which can be used in its place.
(Disclaimer: I’m not a WP developer, I just reported the bug and attempted a patch.)

Thanks. That looks a bit too high-traffic for me right now — I’m not heavily into development, it’s just a few issues I’m chasing down — but I’ll keep it in mind the next time I have this kind of problem.
(On the specific issue, it looks like Matt spotted the patch and added it to CVS, so I guess the delay really was the fact that the fix was in “paste this here” format instead of a patch.)

Thanks. That seems to be working!

There is one circumstance in which “paged” pages do *not* change: monthly archives, once the month is over. For instance, I have two pages worth of posts for August 2003, and those are never going to change (barring typo corrections, extra comments, etc.)

TechGnome – sure, trackbacks and pingbacks came up partway through… but the original question hadn’t been answered completely (and it looked like pingbacks/trackbacks came up due to someone misunderstanding the question.)
I’m not in the mood to install a nightly just now, though. If someone in the know could confirm that the old weblogs.com and blog.gs options have been generalized to allow pinging of multiple (configurable) sites, I’d appreciate it! Thanks!

The original question wasn’t about pingbacks or trackbacks, it was about “pinging” a directory like weblogs.com to let it know your site has been updated.
weblogs.com and blo.gs are hard-coded into WP 1.0, but there are other sites (rootblog.com, for instance) that work on the same principle, often via the same mechanism. The feature Sushubh mentioned sounds like it, assuming you can set multiple URIs and not just one.

Sounds fair. When I have a chance I’ll see what I can do.

The bounces I got earlier referred me to https://www.pobox.com/bounce-badfrom.mhtml which cites RFC 2822 as their justification.
Here’s what I did. Just before the infamous line 111 I added:
$comment_from_email = $comment_author_email;
if( $comment_author_email == ”) {
$comment_from_email = $admin_email;
}
Then I replaced $comment_author_email on the next line with $comment_from_email
This may have nothing to do with the problems ginadapooh is having, but at least it will help in some cases.

Thanks! At first I was worried that this might break HTML in titles (I’ve used italics from time to time), but it looks OK – and now it both validates and shows up in Netscape 4! (Don’t laugh, I have friends who are still stuck on it because their computers are too old to run anything newer.)