Initial NodeJS string wrapper

[MarkReedZ]

I've created a Str object wrapper for Node.js

    const a = new stringzilla.Str('hello world');
    const b = new stringzilla.Str('hi');

    const result_1 = a.count(b);

And implemented a few functions with tests.

        {"indexOf", NULL, str_find, NULL, NULL, NULL, napi_default, NULL},
        {"find", NULL, str_find, NULL, NULL, NULL, napi_default, NULL},
        {"rfind", NULL, str_rfind, NULL, NULL, NULL, napi_default, NULL},
        {"startswith", NULL, str_startswith, NULL, NULL, NULL, napi_default, NULL},
        {"endswith", NULL, str_endswith, NULL, NULL, NULL, napi_default, NULL},
        {"count", NULL, str_count, NULL, NULL, NULL, napi_default, NULL}

[recurseml]

Missing error handling for napi_get_cb_info call. The function returns napi_status which should be checked. If it fails, the function could proceed with invalid args/jsthis values. Should check return value and handle error cases.

📚 Relevant Docs

---

React with 👍 to tell me that this comment was useful, or 👎 if not (and I'll stop posting more comments like this in the future)

[recurseml]

No NULL check after malloc. If allocation fails, the function continues with an invalid pointer which could cause a crash. Should check if obj is NULL and handle the error case appropriately.

---

React with 👍 to tell me that this comment was useful, or 👎 if not (and I'll stop posting more comments like this in the future)

[recurseml]

Using a relative path ('../../build/Release/stringzilla') with the 'bindings' module is incorrect and could cause module resolution failures. The 'bindings' module is designed to automatically locate native addons and should be used with just the module name. The original code using bindings('stringzilla') was correct. The change to a relative path breaks the standard module resolution mechanism and could fail when the package is installed in different locations.

📚 Relevant Docs

---

React with 👍 to tell me that this comment was useful, or 👎 if not (and I'll stop posting more comments like this in the future)

[ashvardanian]

Thank for laying all the groundwork, @MarkReedZ! Even though we can't access string contents without copy, we still have a zero-copy NAPI for Buffer-s! So I'm finally proceeding forward with a more functional JS binding!