?? We don't have Autocorrect.
If the software recognizes a URL, then the URL will be encoded. But perhaps the software cannot determine when a URL is
improperly formatted.
Inside the code box are some URLs. The last three are badly formatted. The software cannot recognize all of the bad formats. When placed inside a code box, the forum software will not modify the URLs or attempt encoding.
OK www.nytimes.com
OK http://nytimes.com
OK nytimes.com BUT INCOMPLETE
BAD Http://nytimes.com CAPITAL LETTER
BAD htp://garbage.com MISSPELLED
BAD www.nytimes .com DISCONNECTED
Here is how the software handles those URLs. The two properly formatted, complete URLs are recognized and encoded. The incomplete URL is ignored (but will work in a browser). The badly formatted "Http" URL and the disconnected URL are
partially recognized, but improper encoding results. The misspelled URL is not recognized.
OK
www.nytimes.comOK
http://nytimes.comOK nytimes.com BUT INCOMPLETE
BAD
http://Http://nytimes.com CAPITAL LETTER
BAD htp://garbage.com MISSPELLED
BAD
www.nytimes .com DISCONNECTED
The URL recognition software most likely parses phrases without regard to capitalization becaused web adresses are not sensitive to that. Thus the software did recognize the "Http" URL and tried to encode it. I agree in theory that this is bug, but only so much recognition can be built into software for recognizing badly formatted phrases.
To illustrate the parsing comment -- if you type the following into a browser adress bar, you will be taken to the New York Times website.
Http://nytimes.com
HTTP://NYTIMES.COM