Improve parsing performance

[Gabriella439]

This uses takeWhile{,1}P as a faster alternative to many . satisfy to
speed up parsing large labels

This also uses the many re-exported by megaparsec which is more efficient
for types that implement MonadPlus

[ocharles]

Any numbers to go with this?

[Gabriella439]

@ocharles: I added two new benchmarks:

• The long variable name benchmark shows a performance improvement related to label parsing
• The large number of arguments benchmark shows a (smaller) performance improvement related to using the improved many

Before:

benchmarking Issue #108
time                 344.5 ms   (332.5 ms .. 353.1 ms)
                     1.000 R²   (1.000 R² .. 1.000 R²)
mean                 349.2 ms   (345.6 ms .. 354.7 ms)
std dev              5.533 ms   (1.035 ms .. 7.169 ms)
variance introduced by outliers: 19% (moderately inflated)

benchmarking Long variable names
time                 585.9 ms   (NaN s .. 637.3 ms)
                     0.998 R²   (0.998 R² .. 1.000 R²)
mean                 597.7 ms   (586.5 ms .. 609.0 ms)
std dev              13.25 ms   (10.58 ms .. 15.33 ms)
variance introduced by outliers: 19% (moderately inflated)

benchmarking Large number of function arguments
time                 152.4 ms   (143.2 ms .. 168.1 ms)
                     0.993 R²   (0.976 R² .. 1.000 R²)
mean                 153.8 ms   (143.6 ms .. 170.1 ms)
std dev              19.33 ms   (9.145 ms .. 31.83 ms)
variance introduced by outliers: 40% (moderately inflated)

After:

benchmarking Issue #108
time                 344.1 ms   (323.8 ms .. 357.6 ms)
                     1.000 R²   (0.999 R² .. 1.000 R²)
mean                 343.1 ms   (339.4 ms .. 346.5 ms)
std dev              4.249 ms   (1.928 ms .. 5.340 ms)
variance introduced by outliers: 19% (moderately inflated)

benchmarking Long variable names
time                 56.60 ms   (52.82 ms .. 59.34 ms)
                     0.994 R²   (0.987 R² .. 0.999 R²)
mean                 57.57 ms   (56.35 ms .. 59.46 ms)
std dev              2.711 ms   (1.607 ms .. 4.033 ms)
variance introduced by outliers: 15% (moderately inflated)

benchmarking Large number of function arguments
time                 141.6 ms   (134.1 ms .. 147.4 ms)
                     0.997 R²   (0.991 R² .. 1.000 R²)
mean                 159.9 ms   (153.3 ms .. 171.7 ms)
std dev              12.59 ms   (7.641 ms .. 17.65 ms)
variance introduced by outliers: 14% (moderately inflated)