feat(encoding/unstable): decodeBase32Hex() and encodeBase32Hex()

[BenMcLean981]

As requested in #4917 I have added base32hex encoding and decoding.

I first created new tests from the base32 test cases and using this tool:
https://tomeko.net/online_tools/base32hex.php?lang=en

Once those tests passed, I extracted the common base32 and base32 hex into a _utilsmodule with a new parameter for a lookup table.

This is my first time contributing to an open source project and using deno. I am having trouble running the task ok. It is failing on my documentation examples with the error below. I think I'm missing a step though, because it's executing with the published module from jsr rather than the local version. This means it makes sense that my module is not there, because it won't be in version 0.224.3. If there's any step I'm missing here I'd really appreciate any feedback.

[error] Failed to execute snippet: 
import { decodeBase32Hex } from "@std/encoding/base32hex";
import { assertEquals } from "@std/assert/assert-equals";

assertEquals(
  decodeBase32Hex("GZRTMMDDGA======"),
  new TextEncoder().encode("6c60c0"),
);
error: Unknown export './base32hex' for '@std/encoding@0.224.3'.
  Package exports:
 * .
 * ./ascii85
 * ./base32
 * ./base58
 * ./base64
 * ./base64url
 * ./hex
 * ./varint
 at file:///home/benmclean/projects/deno/deno_std/encoding/base32hex.ts:54
[error] Failed to execute snippet: 
import { encodeBase32Hex } from "@std/encoding/base32hex";
import { assertEquals } from "@std/assert/assert-equals";

assertEquals(encodeBase32Hex("6c60c0"), "GZRTMMDDGA======");
error: Unknown export './base32hex' for '@std/encoding@0.224.3'.
  Package exports:
 * .
 * ./ascii85
 * ./base32
 * ./base58
 * ./base64
 * ./base64url
 * ./hex
 * ./varint
 at file:///home/benmclean/projects/deno/deno_std/encoding/base32hex.ts:74
2 errors found

Closes #4931

[CLAassistant]

All committers have signed the CLA.

[iuioiua]

Thank you for this. @std/encoding is about to be in its release candidate phase, so it'll be in a change freeze (except for documentation improvements and bug fixes). This change will be merged once that's complete.

[iuioiua]

Can you please add the new module to the exports in encoding/deno.json?

[BenMcLean981]

Can you please add the new module to the exports in encoding/deno.json?

I actually already did, I didn't add the _utils.ts module though because it should be private.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.88%. Comparing base (714fccb) to head (6cccaaf).
Report is 60 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4931      +/-   ##
==========================================
- Coverage   96.37%   95.88%   -0.49%     
==========================================
  Files         465      475      +10     
  Lines       37506    38464     +958     
  Branches     5527     5553      +26     
==========================================
+ Hits        36147    36883     +736     
- Misses       1317     1539     +222     
  Partials       42       42              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[BlackAsLight]

The only error left seems to be that you forgot to update the assertEquals in the jsdoc examples when you copied it over.

[BenMcLean981]

The only error left seems to be that you forgot to update the assertEquals in the jsdoc examples when you copied it over.

Good catch, this has been fixed, thanks.

[iuioiua]

Hey there. Sorry for the delay. On second thought, with the only difference being the alphabet used, I think we should add an options argument to the base32 functions, instead of creating new symbols and modules. We can then just use a simple ternary operator in the function bodies to choose the appropriate alphabet. It could look like this.

interface Base32Options {
  /**
   * Base32 format to use.
   *
   * @default {"standard"}
   */
  format?: "standard" | "hex";
}

What do y'all think?

[iuioiua]

Actually, we can have separate functions but have them all in the existing base32 module, since there's so much overlap.

[BlackAsLight]

I think the best way to do it would be the same way base64 and base64url are done.

[iuioiua]

base32 and base32Hex functions have almost complete overlap - there's no benefit to separating them. Arguably, the same could be said for base32Url, but that's a separate conversation.

[BlackAsLight]

I don't have an opinion one way or the other on how it should be done, but I think consistency is best.

[iuioiua]

I chatted with Yoshiya, and on second thought, we're happy to proceed with this format. I also made a couple of tweaks. Now, LGTM! Thank you very much.