solve issue/192 (the case when <br> tag exists within <pre> tag)

[kawacchu]

In no.750 and no.751, the author have written sample case using <br> tag within <pre> tag, which is not a user concern but we might not have expected such cases.

For example, no.750 サンプル1 入力1
<pre>6<br />1 1<br />7 4<br />0 5<br />1 3<br />-8 9<br />5 1</pre>
is given for online-judge-tools.

Beautiful Soup documents says that

If a tag contains more than one thing, then it’s not clear what .string should refer to, so .string is defined to be None:

So we should fix
s = tag.string or '' # tag.string for the tag "<pre></pre>" returns None
(in _parse_sample_tag function, onlinejudge/service/yukicoder.py file).

In this pull request, I use .strings generator and iterate to make what the user expects.

But the unexpected behavior for the user will remain in some problems. (although I have never seen them in a real environment.)

Let me show a tricky problem.

問題文
文字列Sが1行で与えられる. Sは小文字のアルファベットと<と>と/を用いて構成される.
Sに含まれるrの個数を1行で出力しなさい.
制約
1 <= S <= 100
サンプル
入力1
yukicode
r
出力1
4

... because this problem's author thinks 入力1 as <strong>yuki</strong>code<br>r

In my opinion, user's expectation for oj d is to download what they see and we have not to handle with such tricky problems.

[kmyk]

LGTM
It's perfect!

[kmyk]

In terms of problems whose samples freely use HTML tags, I think we ideally should just abandon entire parsing or ignore non-<br> tags (i.e. read as yukicode\nr). Using emphasis in samples is considerable, but, it seems enough rare to just ignore them for now.

(Note: If the problem author think the string <strong>yuki</strong>code<br>r itself as an input, it will be escaped and appears as >strong<yuki;>/strong<yuki;code>br<r in HTML.)

[kawacchu]

Thank you for your kind review !
I will consider what you replied in #553 .