Improve code chunk handling in base .R files

[kylebutts]

This PR fixes #1453 and fixes some bugs in handling of code chunks in .R and .Rmd files.

The first change is with getChunks(). Previously, for .R files, the scan would find the first # %% and mark it as the start line. Then, it would go until it found the next # %% and incorrectly mark this as the end line. The line just before the next # %% needs to be the end line (see #1453). This PR fixes this:

The second change is with getCurrentChunk(). I have simplified this code greatly and do much less work. Basically, I check for three edge cases (no chunks, line < first chunk, line < last chunk) and return early. Otherwise, I select the first chunk with .endLine > line. This works both for cursor within chunk and cursor between chunks.

There is one change worth noting. Before, getCurrentChunk() would return undefined when line was after the last chunk. I instead return the last chunk and fix bugs with any commands this impacted.

Testing

To test this, I created these two files (temp.R and temp.Rmd) and tried out the commands for running chunks, selecting current chunk, and going to next and previous chunk.

# %% Chunk 1
print("Chunk 1")

# %% Chunk 2

print("Chunk 2")

# %% Chunk 3
print("Chunk 3")

---
---

```{r}

print("Chunk 1")

```


```{r}
print("Chunk 2")
```

```{r}
print("Chunk 3")
```

[renkun-ken]

LGTM Thanks for working on this!