refactor: use smol_str for template strings

[sunng87]

This patch uses smol_str library to hold template and path strings. Thanks to the inline and stack based string, we gain some performance improvements with this:

     Running benches/bench.rs (target/release/deps/bench-b84c4a320f8d765c)
Gnuplot not found, using plotters backend
parse_template          time:   [15.024 µs 15.292 µs 15.701 µs]
                        change: [+1.1207% +2.2612% +3.7369%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high severe

render_template         time:   [1.6720 µs 1.6745 µs 1.6774 µs]
                        change: [−12.666% −12.469% −12.246%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 14 outliers among 100 measurements (14.00%)
  2 (2.00%) high mild
  12 (12.00%) high severe

large_loop_helper       time:   [111.68 µs 111.86 µs 112.06 µs]
                        change: [−13.894% −13.667% −13.431%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  3 (3.00%) high mild
  1 (1.00%) high severe

large_loop_helper_with_context_creation
                        time:   [290.04 µs 290.43 µs 290.75 µs]
                        change: [−7.6980% −7.4180% −7.1419%] (p = 0.00 < 0.05)
                        Performance has improved.

large_nested_loop       time:   [117.20 µs 117.41 µs 117.66 µs]
                        change: [−20.581% −20.093% −19.617%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 8 outliers among 100 measurements (8.00%)
  7 (7.00%) high mild
  1 (1.00%) high severe

deeply_nested_partial   time:   [185.19 µs 185.42 µs 185.72 µs]
                        change: [+1.6880% +1.8193% +1.9683%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild

For the large_nested_loop where we have a lot of context/path string manipulations, we get up to 20% speed up.

However, this patch can be considered as breaking change for users who access to TemplateElement directly.

[coveralls]

coverage: 83.49% (+0.04%) from 83.449%
when pulling 3fc6688 on feature/smol-str
into b1b39cd on master.