fix incorrect slicing by length detection cases

[Techatrix]

I initially just wanted to simplify the code that handles the slicing syntax in AstGen.zig but then found two bugs related to the special Slicing By Length syntax:

This example will incorrectly trigger slicing by length even though no end index has been provided. It will crash the compiler.

test {
    var foo = "";
    _ = foo[0..][0.. :0];
}

In this example, the provided sentinel expression is never analyzed and no sentinel mismatch safety check is reported.

test {
    var foo: [2:0]u8 = .{ 1, 2 };
    _ = foo[0.. :1][0..2];
}

The PR that added the slicing by length special case has explicitly mentioned support for s[start..:sentinel][0..len] and s[start..:sentinal_1][0..len:sentinel_2] but appears to have never tested what happens if the first sentinel is invalid.

The diff as a whole is not very readable. I would recommend to look at the commits individually.

Here is a Benchmark of running ast-check on Sema.zig:

poop "build/stage4-fast/bin/zig ast-check src/Sema.zig" "build/stage4-sos-fast/bin/zig ast-check src/Sema.zig" -d 20000
Benchmark 1 (192 runs): build/stage4-fast/bin/zig ast-check src/Sema.zig
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           104ms ± 1.76ms     101ms …  109ms          0 ( 0%)        0%
  peak_rss           77.7MB ± 48.1KB    77.4MB … 77.8MB         24 (13%)        0%
  cpu_cycles          340M  ± 5.09M      332M  …  357M           1 ( 1%)        0%
  instructions        516M  ± 11.3       516M  …  516M           0 ( 0%)        0%
  cache_references   1.41M  ± 63.2K     1.24M  … 1.56M           7 ( 4%)        0%
  cache_misses        354K  ± 34.8K      298K  …  416K           0 ( 0%)        0%
  branch_misses      4.75M  ± 3.17K     4.75M  … 4.76M           9 ( 5%)        0%
Benchmark 2 (194 runs): build/stage4-sos-fast/bin/zig ast-check src/Sema.zig
  measurement          mean ± σ            min … max           outliers         delta
  wall_time           103ms ± 2.37ms     101ms …  129ms          2 ( 1%)          -  0.7% ±  0.4%
  peak_rss           77.8MB ± 67.5KB    77.6MB … 77.8MB          0 ( 0%)          +  0.1% ±  0.0%
  cpu_cycles          337M  ± 6.85M      332M  …  411M           3 ( 2%)          -  0.7% ±  0.4%
  instructions        517M  ± 9.75       517M  …  517M           2 ( 1%)          +  0.1% ±  0.0%
  cache_references   1.42M  ± 58.4K     1.25M  … 1.86M           8 ( 4%)          +  0.6% ±  0.9%
  cache_misses        341K  ± 30.2K      294K  …  484K           1 ( 1%)        ⚡-  3.8% ±  1.8%
  branch_misses      4.73M  ± 3.05K     4.73M  … 4.76M           1 ( 1%)          -  0.4% ±  0.0%

[andrewrk]

Thanks!