Let C code handle string conversion

[Bisaloo]

No description provided.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 7d8bea6 is merged into main:

• ❗🐌pkg_load: 4.85ms -> 4.95ms [+0.38%, +4.06%]
• ✔️read_bigendian: 1.42ms -> 1.4ms [-3.43%, +0.55%]
• ✔️read_bool: 1.34ms -> 1.32ms [-3.95%, +0.57%]
• 🚀read_empty: 1.39ms -> 1.35ms [-4.54%, -0.14%]
• ✔️read_float32: 1.34ms -> 1.38ms [-3.56%, +8.28%]
• 🚀read_float64: 1.37ms -> 1.34ms [-3.68%, -0.84%]
• 🚀read_int16: 1.37ms -> 1.34ms [-3.58%, -0.39%]
• ✔️read_int32: 1.39ms -> 1.36ms [-3.46%, +0.3%]
• ✔️read_int8: 1.38ms -> 1.36ms [-4.47%, +1.41%]
• ✔️read_npz: 3.37ms -> 3.35ms [-1.8%, +0.79%]
• 🚀read_string: 1.64ms -> 1.37ms [-19.27%, -14.46%]
• ✔️read_uint16: 1.36ms -> 1.35ms [-1.79%, +0.2%]
• 🚀read_uint32: 1.72ms -> 1.65ms [-7.56%, -0.68%]
• ✔️read_uint64: 1.88ms -> 1.93ms [-3.59%, +8.88%]
• ✔️read_uint8: 1.39ms -> 1.36ms [-5%, +0.05%]
• ✔️read_unicode: 1.69ms -> 1.66ms [-4.39%, +0.83%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 516b0b6 is merged into main:

• ❗🐌pkg_load: 4.49ms -> 4.6ms [+0.28%, +4.95%]
• 🚀read_bigendian: 1.24ms -> 1.2ms [-4.31%, -1.52%]
• 🚀read_bool: 1.15ms -> 1.13ms [-3.5%, -0.79%]
• 🚀read_empty: 1.23ms -> 1.19ms [-4.5%, -2.27%]
• 🚀read_float32: 1.14ms -> 1.1ms [-3.57%, -2.28%]
• 🚀read_float64: 1.14ms -> 1.11ms [-3.63%, -1.93%]
• 🚀read_int16: 1.17ms -> 1.14ms [-4.03%, -1.72%]
• 🚀read_int32: 1.19ms -> 1.13ms [-5.99%, -3.13%]
• 🚀read_int8: 1.16ms -> 1.12ms [-5.26%, -2.13%]
• ✔️read_npz: 3.05ms -> 3.02ms [-2.01%, +0.19%]
• 🚀read_string: 1.44ms -> 1.15ms [-21.15%, -19.1%]
• ✔️read_uint16: 1.25ms -> 1.24ms [-3.37%, +1.17%]
• ✔️read_uint32: 1.51ms -> 1.5ms [-2.6%, +0.36%]
• 🚀read_uint64: 1.69ms -> 1.65ms [-3.58%, -1.54%]
• ✔️read_uint8: 1.22ms -> 1.2ms [-3.91%, +0.55%]
• 🚀read_unicode: 1.49ms -> 1.3ms [-13.83%, -10.92%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if f1b8341 is merged into main:

• ❗🐌pkg_load: 4.4ms -> 4.6ms [+2.19%, +6.74%]
• 🚀read_bigendian: 1.23ms -> 1.2ms [-3.36%, -1.46%]
• 🚀read_bool: 1.16ms -> 1.12ms [-4.95%, -2.14%]
• 🚀read_empty: 1.18ms -> 1.14ms [-3.79%, -1.71%]
• 🚀read_float32: 1.19ms -> 1.15ms [-4.28%, -2.23%]
• 🚀read_float64: 1.19ms -> 1.16ms [-4.26%, -0.11%]
• 🚀read_int16: 1.16ms -> 1.12ms [-3.85%, -2.19%]
• 🚀read_int32: 1.15ms -> 1.11ms [-3.77%, -2.37%]
• 🚀read_int8: 1.16ms -> 1.11ms [-4.97%, -2.45%]
• ✔️read_npz: 3.07ms -> 3.06ms [-1.52%, +1%]
• 🚀read_string: 1.43ms -> 1.14ms [-21.37%, -19.46%]
• 🚀read_uint16: 1.22ms -> 1.18ms [-5.16%, -0.6%]
• 🚀read_uint32: 1.5ms -> 1.46ms [-4.9%, -0.86%]
• 🚀read_uint64: 1.72ms -> 1.68ms [-3.71%, -1.14%]
• 🚀read_uint8: 1.2ms -> 1.16ms [-4.17%, -2.2%]
• 🚀read_unicode: 1.53ms -> 1.36ms [-13.26%, -10.05%]
  Further explanation regarding interpretation and methodology can be found in the documentation.