Add ShangMi (SM2/SM3/SM4/SM9) algorithm families

[Mehrn0ush]

Fixes #811

Adds SM2/SM3/SM4/SM9 algorithm families to the CycloneDX cryptography registry and updates algorithmFamiliesEnum accordingly.

Note: some official ShangMi specification mirrors appear region-blocked or unreliable from multiple locations.
To avoid fragile links in the registry, this PR relies on globally accessible references (RFC Editor + ISO)

[Mehrn0ush]

Hi @bhess — if you have time, would you mind taking a look at this one as well?

This adds ShangMi (SM2/SM3/SM4/SM9) families and updates the schema enum.
I used RFC/ISO references where the official mirrors weren’t reliably accessible.

Thanks a lot — always appreciate your feedback.

[stevespringett]

cc: @bhess

[bhess]

Thanks @Mehrn0ush for the additions! See the comments inline.

[Mehrn0ush]

One question on SM2 naming conventions for the registry: I currently model SM2 as SM2-256 (and ...-256 for ENC/KEX) to be explicit about the common curve size. Would you prefer making the suffix optional (SM2[-256]) to align with the registry’s broader matching style, or keeping it strict as SM2-256? I can adjust either way — just want to follow the convention you think is best for downstream matching.
And,
Related detail: for SM4 placeholders I followed the existing conventions in cryptography-defs.json, where block cipher modes use {ivlen}, while ae modes use {ivLength} together with {tagLength} (consistent with how AES-GCM/CCM are expressed in the registry). If you’d rather normalize these placeholders to a single naming form across the file, I’m happy to align SM4 accordingly.

Thanks again @bhess

[Mehrn0ush]

Made SM2’s -256 suffix optional to accept both SM2 and SM2-256, following the registry’s [] convention. Updated patterns accordingly.

[bhess]

Thanks for the updates, LGTM!