lint(track_config): optimize concept cycle checking slightly (#489)

We were concatenating a sequence with a one-item sequence, but it's
better to use the `&` overload that appends one item [1][2]. The
Nim compiler deliberately doesn't optimize the former into the latter.

The concept cycle checking is a hot spot on tracks with a larger number
of Concept Exercises. For example, with the csharp track, about a third
of the `configlet lint` execution time was spent in `checkForCycle`.
This commit removes some unnecessary allocations within that hot spot,
and thereby produces a small (but consistent) `configlet lint` speedup
of 3-5% on such tracks.

[1] https://github.com/nim-lang/Nim/blob/9084d9bc02bc/lib/system.nim#L1672-L1686
[2] https://github.com/nim-lang/Nim/blob/9084d9bc02bc/lib/system.nim#L1688-L1701

src/lint/track_config.nim

@@ -427,7 +427,7 @@ proc checkForCycle(prerequisitesByConcept: Table[string, seq[string]];
   if hadCycle:
     return
 
-  let updatedPrereqPath = prereqPath & @[currentConcept]
+  let updatedPrereqPath = prereqPath & currentConcept
   if currentConcept in prereqPath:
     var formattedCycle = &"{q updatedPrereqPath[0]} depends on {q updatedPrereqPath[1]}"
     for i in 1..prereqPath.high: