In node_buffer.h, why do all the calls accept (and return) char * instead of unsigned char *?

[sgerace]

I'm writing a node addon and I was just curious if there was a reason why all of the call in node::Buffer accept and return char * instead of the (arguably) more representative unsigned char *. I'm analyzing buffers as they are passed into my function, so it is certainly trivial for me to do:

unsigned char *data = (unsigned char *)node::Buffer::Data(args[0]);

I was just wondering if there was a specific reason why char * was chosen since buffers (by definition) are intended to contain binary data.

[mscdex]

/cc @trevnorris

[bnoordhuis]

It's pretty much an accident of history. I'd have picked uint8_t* myself.

[trevnorris]

Probably has something to do with the old v8 api working with char*. But it is for legacy reasons.

[trevnorris]

@bnoordhuis Would we consider changing the native API again, or is this something we'll be leaving alone? char* may have been chosen to work most easily with the V8 API at the time, but that isn't so much the case now.

Personally I'd say we leave it as is unless there's a compelling reason to change it.

[bnoordhuis]

I'm -0 on changing it now. It would cause a lot of breakage in the ecosystem for comparatively little benefit.

[smartmouse]

totally agree, pls do not break things.