Adjacent challenge block quotes are parsed as one

For example, in episode 3 of the git lesson there are two challenge blocks adjacent to each other: http://swcarpentry.github.io/git-novice/03-create/index.html, but the `get_blocks()` function is only returning one continuous block quote.

Note that I can get around this with the graph representation or maybe inserting a break after the tag, but it's still annoying nonetheless.

``` r
library("up2code")
git <- get_lesson("swcarpentry/git-novice")
#> cloning into '/tmp/Rtmp6aov4N/swcarpentry--git-novice'...
#> Receiving objects: 1% (84/8345), 42 kb
#> Receiving objects: 11% (918/8345), 394 kb
#> Receiving objects: 21% (1753/8345), 590 kb
#> Receiving objects: 31% (2587/8345), 2427 kb
#> Receiving objects: 41% (3422/8345), 6381 kb
#> Receiving objects: 51% (4256/8345), 7406 kb
#> Receiving objects: 61% (5091/8345), 8102 kb
#> Receiving objects: 71% (5925/8345), 8407 kb
#> Receiving objects: 81% (6760/8345), 8647 kb
#> Receiving objects: 91% (7594/8345), 8871 kb
#> Receiving objects: 100% (8345/8345), 10134 kb, done.
git$episodes$`03-create.md`$challenges
#> {xml_nodeset (1)}
#> [1] \n \n Plac ...
```

^{Created on 2020-05-12 by the [reprex package](https://reprex.tidyverse.org) (v0.3.0)}

Start by reproducing the issue with get_blocks() on episode 3 of the linked git-novice lesson and inspect how adjacent challenge block quotes are grouped. The work is done when the two adjacent challenge blocks are returned separately without requiring a graph workaround or inserted break.

Written by the indexing model from the issue text.