Add functions to load and dump hex string

[aiotter]

These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).

SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later

Closes #1287.

[aiotter]

Thank you for your review, I updated my code

[pguyot]

I am also concerned CI didn't run because of the merge conflict, even if it LGTM.

[bettio]

Everything looks good to me too.

However, if you feel confident with git rebase -i release-0.6 or any other tool, would you mind rewriting history so:

• Fix binary_to_int/2 to work in the same way with OTP can be fixed up in Add support for binary_to_integer/2
• Fix binary:decode_hex/1 to work in the same way with OTP can be fixed up in Add support for binary:decode_hex/1 and binary:encode_hex/1,2

Also by doing so you would allow GH running again the CI.

That would be super appreciated.

[aiotter]

Okay, I'll rebase the branch onto release-0.6 to clean up the history and to resolve conflict.

[aiotter]

On libs/estdlib/src/binary.erl, I copied and pasted only the docs and specs of the functions from OTP. Specs are just a piece of information and have no copyright, but the docs can. Do you think it can cause a copyright problem?

The docs of libs/exavmlib/lib/Base.ex is also copied from Elixir, and it is sure that the Elixir contributors have copyright for them. So I made that clear at the top of the file.

[aiotter]

btw, it seems every file of the source code has Apache 2.0 at the top of it. Which part of AtomVM is licensed under LGPL?

[aiotter]

I had forgot that a language server I use behaves differently from erlfmt. So I fixed it.
https://github.com/atomvm/AtomVM/actions/runs/11171766326/job/31064782745?pr=1304

[bettio]

btw, it seems every file of the source code has Apache 2.0 at the top of it. Which part of AtomVM is licensed under LGPL?

The project used to be LGPL then we switched to Apache-2.0 OR LGPL, so for end users means pick your favorite license.
It's ok to have some module and libraries that are Apache-2.0 only, while LGPL only is not allowed.

[bettio]

I apologize for noticing this late, but there is a small detail that must be fixed: bases lower than 10 requires more room in the buffer. So we must update it accordingly to avoid potential buffer overflows.
Just make this change as a fixup, there is no need to add an additional commit.

I also suggest converting a 62 bit number to binary as an additional test case.

Everything else is good.

[aiotter]

I just appended @realglobe-Inc at the end of the copyright.

[bettio]

I just appended @realglobe-Inc at the end of the copyright.

@aiotter I'm ok with this, still there is a problem with buffer size: when dealing with a base like 2, for which the maximum number has more digits than the 23 used for base 10.
Take a look to the comments inside the changes to see what I mean.

[aiotter]

@bettio Oh I already sent a reply for your comment. Didn't you missed it, or am I missing something?
#1304 (comment)

[bettio]

I think something happened and your change / your comment somehow is not visible or at least I cannot open it. Can you double check it too? Part of that comment is visible only through preview.

[aiotter]

I don't know why, but my old comment is regarded as a new code review in draft. Maybe a bug of GitHub? I abandoned the code review and resent a comment, so now you can check it ;-)
#1304 (comment)

[bettio]

I don't know why, but my old comment is regarded as a new code review in draft. Maybe a bug of GitHub? I abandoned the code review and resent a comment, so now you can check it ;-) #1304 (comment)

yeah, looks likely a GitHub bug, it was very odd. Anyway I left a reply comment.

[aiotter]

I increased the buffer size. It seems ready to be merged now!

[bettio]

I increased the buffer size. It seems ready to be merged now!

Thanks for the changes, I'm going to merge it right now.