tld: incorrectly parsed Wikipedia table

[dgw]

One table of the Wikipedia article we use for TLD details in tld gets parsed wrong, and makes things like this happen:

<dgw> ,tld xn--q7ce6a
<SopelTest> [tld] : Lao | Bulgaria: .ລາວ | Bulgarian: Laos | Cyrillic: Lao | bg: Lao | .bg: Not in use | No: .la
<dgw> ,tld ລາວ
<SopelTest> [tld] : Lao | Bulgaria: .ລາວ | Bulgarian: Laos | Cyrillic: Lao | bg: Lao | .bg: Not in use | No: .la

A quick look at the HTML didn't reveal any obvious structural differences between this table and the correctly-parsed others, but something is obviously tripping up my rudimentary HTMLParser-derived class. I'll probably need to spend some quality time with pdb, trying to figure out where in the parsing routine the data gets mangled.

Follow-up to #1939 (comment)

[dgw]

Peppered debug logging through the parser, spun up my test bot, issued a bunch of TLD commands, and… nothing. Can't reproduce this any more. Will leave in the 7.1 milestone for historical purposes, but seems this likely wasn't our problem.