🔡 Unsigned integer 4,8,12,16,20,24 to hexadecimal

[CalebSerafin]

What type of PR is this.

• Bug
• Change

3. Enhancement

What have you changed and why?

Formats a scalar as the specified length hexadecimal string.
These lengths are divided into multiple functions for 4, 8, 12, 16, 20, 24 bits.
This is needed for UUID generation and output. May help other areas.

Please verify the following and ensure all checks are completed.

1. Have you loaded the mission in LAN host?
2. Have you loaded the mission on a dedicated server?

Is further testing or are further changes required?

1. No

• Yes (Please provide further detail below.)

How can the changes be tested?

[
    A3A_base16LookupTable   # 0x0        isEqualTo "0",
    A3A_base16e2LookupTable # 0xb4       isEqualTo "b4",
    0xaaf       call A3A_fnc_uint12ToHex isEqualTo "aaf",
    0xdead      call A3A_fnc_uint16ToHex isEqualTo "dead",
    0xca1eb     call A3A_fnc_uint20ToHex isEqualTo "ca1eb",
    0xffffff    call A3A_fnc_uint24ToHex isEqualTo "ffffff"
] isEqualTo [true,true,true,true,true,true];

[jaj22]

What part of UUID generation would this be necessary for? I can only see the utility for log output, where it's maybe a bit much given that decimal output would mostly work. I also struggle to see a case where you'd use most of the functions, so one 256-entry lookup table and 16+24 bit output functions seems like plenty of functionality. I don't think this is worth a first-level folder on its own either (think of the poor newbies trying to navigate the codebase), maybe a subfolder under utility using the file = CfgFunctions option.

According to https://community.bistudio.com/wiki/Initialization_Order, preinit is unscheduled anyway so there's no point in the isNil wrapper here.

[CalebSerafin]

For UUIDs longer than the 24bits or made from multiple pieces of information, working with them as strings makes it far easier to stick pieces together. Yeah, I could cut the 4096 entry lookup table. However, the 16 entry one will still be needed to avoid needing to trim zero's afterwards. These are good points.

[CalebSerafin]

Okay, the changes have been made but I'm unable to test as Arma dropped a big S.O.G. update and I don't have the internet for it.

[jaj22]

For UUIDs longer than the 24bits or made from multiple pieces of information, working with them as strings makes it far easier to stick pieces together. Yeah, I could cut the 4096 entry lookup table. However, the 16 entry one will still be needed to avoid needing to trim zero's afterwards. These are good points.

Ah, so you're intending to generate with random (maybe in multiple parts) but then convert immediately to string for transfer & storage? That makes sense with SQF.

Are you sure you need all the output functions though? Struggling to see what you'd ever need the short ones for.

[CalebSerafin]

They might be handy. I think that's a fair justification for 3-4 lines per function.

[jaj22]

The mini-functions don't actually benefit from the #include, do they? Seems that they're redundant with the preinit.

[CalebSerafin]

It's in case it gets called by another pre-init function.
Hmm, okay, I'll move pre-init to the antistasi init and remove the #include.
Edit: Ah no, it should initialise before JIP, just in case it might be used.

[CalebSerafin]

Edit: Ah no, it should initialise before JIP, just in case it might be used.
Maybe if we add a centralised preinit function (pretty much for preJIP) to Antistasi, the order can be enforced there. What do you think @jaj22?

[jaj22]

That might be necessary in the future, but I think I'm ok with unordered pre-init at the moment, and then create a central function if we need it later.

[CalebSerafin]

@jaj22 pull request approved? Required for merge.

[jaj22]

Weren't you going to remove the #includes? If you're giving me a hard choice between the includes and a centralized ordered preinit then I'll go for the centralized pre-init. For now though I feel like it's fine to have a condition that pre-init doesn't have dependencies on other pre-init.

[CalebSerafin]

@jaj22 Are the new files to your liking. The removal of #include has allowed me to remove functions for uint4 and 8, as they are single lookups.

[jaj22]

Doesn't this one have to be relative in case someone is running the mission unzipped?

[jaj22]

Oh, probably needs ..\..\ regardless. initServer.sqf and init.sqf don't need all the double-dot stuff because they're actually in the root.

[CalebSerafin]

Ah copy-paste caught.