This task does not explain the expected outcome and why pressing Ctrl+U should allow entering Unicode characters (somehow).

Found while updating the FreeBSD port.

Are you sure this doesn't happen without it being enabled?

https://en.wikipedia.org/

On which wiki are you experiencing this?

On main Wikipedia site.

Firefox 55.0.1

Firefox 55.0.1 on FreeBSD 11.1

Here is a recording/playback demo: https://higuma.github.io/web-audio-recorder-js/

What you are referring to is HTTP pipelining. The server doesn't have to read pipelined headers once it determined that there are too many requests. The TCP decongestion procedure takes care of the rest.

The spec seems to suggest failing requests is a valid response

Failing requests isn't a valid response. If I would run this from some company behind the firewall, nobody else will be able to use wikipedia from that IP address because it is flagged as issuing too many requests? This doesn't make sense.

The attached python program runs many requests in many connections concurrently. It saves all responses into the file result.txt
After about 30 seconds this file gets the lines like this:

I use MW API.
I run requests from several threads. How do you define "a lot"?

Always reproducible for me in both mentioned browsers.
New filters beta feature is enabled by default for me for some reason.
It appears that the "Back" button just reloads the page.

Clicking "Link to other maps" isn't persistent there. So it is going to require many clicks to go to non-OSM map just like it does now.

I think so too.
Are you able to substitute the background color when image has transparent background?

I think nobody wants to close the image, hit "back" and see that image again. This isn't the expected or intuitive behavior.

For example, the link is in the top right corner of https://en.wikipedia.org/wiki/San_Francisco_Bay

I don't think wikipedia does user tracking. But the situation when after the login/logout cycle site still looks the same but the cookies are now present, begs the question "what is different now?" IMO there is no need to have this preserved state, since it doesn't serve any obvious purpose, so the cleanest way is to delete the cookies.

It also seems to be related to the size of the response. 32kB is the critical size. The response size is 32766=0x7FFE bytes.

It is 100% reproducible.

This problem is a recent regression. Didn't happen a month ago.

This is still a problem in Chrome as of version 48.0.2564.97.

Either the "changed" field, or the hash of the original text.

It is also red here when I am logged in.

I don't think that the browser is to blame here, because the problem doesn't happen when one clicks out of the page edit screen. It only happens on the preview screen which is a variation of the edit screen.

Observing the problem in the classic editor, with options "Show the edit toolbar", "Enable enhanced editing toolbar" enabled.

No, your assumption is wrong. I have been seeing this for many months, and I am sure this will happen again.

I wasn't seeing the same failure. Will update when it logs again.

For example:

https://api.wiktionary.org?action=query&prop=revisions&titles=<...>&rvprop=timestamp|user|comment|content&format=json

This may have been fixed for the new articles, but I spotted this ones created on 2015-06-29.

Yes, just always print an access key. There is enough space generally.

So what is the suggested way then how Wiktionary can validate submitted data?

After looking into this, I believe now that Lua functions that handle wiktionary templates don't have access to the section information, as it has been parsed. Particularly, to the location of template in the document hierarchy isn't accessible within Lua script.

Some JavaScript should be able to easily achieve this.

And maybe the whole irc dependency need to be removed since only irc-related imports are of ircbot?

If you see that this has been fixed in the newer version then please close it. I don't have 46 available yet.

Last stable release: 44.0.2403.128 (July 29, 2015; 30 days ago[2][3])

Wikipedia says 46 is only dev preview version for Android. Are you on Android?

• Open this URL: https://en.wiktionary.org/w/index.php?title=zebra&action=edit
• Add several empty new lines in the middle
• Open this URL: https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=wiki&tbm=bks
• Copy text there: "A free encyclopedia built collaboratively using wiki software."
• Paste it in the middle of empty lines that you added in wiktionary

How to reproduce:

Classic editor, not Visual, and chrome browser.

What is the project corresponding to basic Wiki software (php/html/js/etc)? This is what should be assigned here, but I don't find anything like that in the list.

I know this isn't too difficult to do with JavaScript. This behavior can be made optional.

This turned out to be a bug in chrome that doesn't percent-encode "quotation" character when copied to clipboard.

I was under impression that there is some Lua code that runs behind the scenes? I don't know why that language is even needed to be there.

Problem is that people copy/paste wrong language a lot.

In Russian they add stress mark to tell where the stress is, but the canonic form has no stress mark.

{{...}} should only substitute it for the string "[...]". If it interprets "[...]" and treats the trailing ']' there as a link closure symbol this is very wrong because link closing bracket hasn't been written at this location. Also there is technically the opening bracket in "[...]" too.

When you have the link [http://www.xxx.com/ Blah blah ... end of blah] - whole text "Blah blah ... end of blah" should be the link.