QCBOREncode_Tell and QCBOREncode_SubString

[laurencelundblade]

Address #247 and #248

(initial PR; testing still needed)

[BrianSipos]

Even with my comment above, the addition of QCBOREncode_RetrieveOutputStorage() and QCBOREncode_Tell() fits my needs.

[BrianSipos]

Since this is the encoder side and the original buffer is mutable anyway, could this return a non-const UsefulBuf?

[laurencelundblade]

That would be counter to the UsefulBuf convention of UsefulBuf's with data always be const and empty buffers not const. I think that convention is upheld everywhere (including t_cose).

I don't mean to be pedantic about it, but I'd like to keep it as for consistency with the convention.

[BrianSipos]

That's fair and my need can use overall Tell offsets where modifying the encoded form is needed for efficiency.

[BrianSipos]

This doesn't fully address #248. For the decoder there could be a public API similar to QCBORDecode_RetrieveInputStorage() to access the input UsefulBufC.

[laurencelundblade]

This doesn't fully address #248. For the encoder there could be a public API similar to QCBORDecode_RetrieveInputStorage() to access the input UsefulBufC.

You want the thus-far-encoded CBOR, right? You can get that by passing 0 as the start to SubString(). Finish() can also be called multiple times, but it errors if there are open arrays or maps, where SubString doesn't. (If you look inside the implementation of Finish and SubString are almost the same too)

Thanks for your feedback.

[laurencelundblade]

This doesn't fully address #248. For the encoder there could be a public API similar to QCBORDecode_RetrieveInputStorage() to access the input UsefulBufC.

You want the thus-far-encoded CBOR, right? You can get that by passing 0 as the start to SubString(). Finish() can also be called multiple times, but it errors if there are open arrays or maps, where SubString doesn't. (If you look inside the implementation of Finish and SubString are almost the same too)

Thanks for your feedback.

Ahh, but you are talking about QCBORDecode() aren't you. That's fine. Got too focused on encode.

[BrianSipos]

Ahh, but you are talking about QCBORDecode() aren't you. That's fine. Got too focused on encode.

Yes it didn't help that I had a typo in my comment. This is about a similar interface for the decoder.