Fix < in code blocks being escaped twice in safe mode

[LuisFigueroaG]

Fixes #240

In safe mode, a < that looks like the start of a tag (a<b, i<3) is replaced with < early on, before the block gamut runs. When that text is inside a code block, the code block encoding escapes the & again, so the example from the issue renders as:

<pre><code>for (i=0; i&amp;lt;3; i++) {

The blockquote from the original report isn't swallowed any more, but the double escaping is still there with the same input. It also happens in fenced code blocks (without pygments), in code blocks nested in lists, and in Markdown Documentation - Syntax.text when converted in safe mode.

What I changed: the escaped < is now hashed, the same way complete tags are already hashed in safe mode, so code blocks leave it alone and it's restored as < at the end. Header ids unescape it before slugifying, so # Comparing a<b keeps its current id.

Tests:

• New case tm-cases/safe_mode_incomplete_tags_in_code_blocks covers indented, fenced and list-nested code blocks, a heading with a<b, and a literal < in a code block (which should still come out as <). It fails on master.
• test.py -- -knownfailure: OK
• test/test_redos.py: 10 tests, OK
• pycodestyle lib/markdown2.py: same warning count as master

I also ran every .text file under test/ through master and this branch, with no safe mode, escape and replace, and with and without a large set of extras. The only difference outside email obfuscation (random on every run) was the intended one in the Syntax document.