[CR][WIP] Foundation for implementing medieval mod shields

[chaosvolt]

This is one of those ideas I'd had on the backburner for a while. Basically, there are two main things preventing the inclusion of shields:

• The big one, my original hackish solution does not allow for the equipped shield to preclude use of two-handed weapons.
• Second, an inability to actively block with them.

The main suggestion that has come up as a way to support these features has been the idea of an off-hand to hold a shield or secondary weapon in. However, this would be a big change, and has been an oft-suggested idea generally dismissed as not worth the effort.

The main reason is that while it could be used for some refactoring of picking items up and holding them, the other main uses would be:

• Going all John Woo with two pistols, which has been flatly dismissed as impractical.
• Doing the same with weapons weapons, which is also impractical. Barring situational uses like rapier-and-dagger, which would STILL be of limited use because how many zombie swordsmen do you expect to encounter?

Additionally, one thing lost when changing shields in the way originally proposed is a loss of passive protection from ranged attacks, which is a major element of historical shield usage. The ability to actively block with them PARTIALLY exists via arm blocks, but one part of this PR will eventually address that.

---

Onto the actual content of the PR. Barring a pressing need to go through the effort of adding offhand support (which is regarded as low-priority and low-utility), adapting the existing method to work in a more plausible manner might be better for the moment.

• First on the coding list, allowing worn items to preclude the use of both hands. The "you can't wield this thing" check should additionally print a message if one of the causes for a failed attempt to wear a two-handed weapon was due to a worn item.
• Ported over shields from the dropbox'd version, with the flag added.
• Added mention of the flag to item.cpp and the JSON documentation.

The other half of the required code is in #14741 it seems. Thanks to Coolthulhu for that, as my attempt at it did not go well. x_x

[chaosvolt]

Success on the first commit? Glorious.

Additionally pondering other uses for the implemented flag, via any other items that might warrant it.

[Malkeus]

Needs new flag added? and your other shields too...

[chaosvolt]

Ack. I could've sworn I added the flag to those. @_@

[kevingranade]

I have no idea where you get the idea that off-hand support is "low-priority and low-utility". It is harder to do, but it is the correct way to handle shields and off-hand weapons (and off-hand tools). See http://smf.cataclysmdda.com/index.php?topic=10444.msg247535#msg247535

[chaosvolt]

Hmm. Part of was that it hasn't been discussed much in a while. Given the complexity it would add, it likely won't be taken up by anyone competent enough to try it for quite some time.

EDIT: That reminds me. This PR would at least have the added use of adding the relevant flag to briefcases and other items.

[Coolthulhu]

If we are ever to consider the hacky shields, they have to do at least those two things:

• Prevent using weapons AND shields with the hand they cover. Since unarmed strike is a weapon, this leaves us with just one option: only one shield may ever be worn and the wielded weapon must be 1-handed if the shield is to be worn
• The shield must be usable to execute blocks with. As in - it must enable the block technique to be used AND must correctly redirect the damage to the shield.

The former is simple - just check one-handedness on equip and disallow one-handed characters from wearing shields (if they want to use shields as weapons, they can wield them as weapons).

The latter wouldn't be hard, but it's not totally straightforward.
You'd have to:

• Write a function that locates a shield in player's worn items and returns it
• Use it in player::block_hit in melee.cpp to replace weapon if an equipped shield exists
• Update player::handle_melee_wear to take an item argument so that things other than current weapon can be damaged

[chaosvolt]

Hmm. The latter was something I had planned, yes. Might need to look for the equip code to see how to implement the former.

[chaosvolt]

And the latest entry in the "What was I thinking?" chronicles, started to add a check in the wear functions for "are we wearing a briefcase/shield OR do we lack hands" before remembering that I already added "are we wearing a briefcase/shield" to the function for defining whether the player has only one arm free.

[Malkeus]

Welcome to the department of redundancy department, you are welcome in this place!

[chaosvolt]

Then again, if I could think up a way to say "You don't have a hand free to wear this." in two differently distinct ways, I could split that check so the message returned is more clear as to why the player can't wear it. Though then I'd need to further refine the has_two_arms check or split "wearing a briefcase/shield" in a separate check.

[chaosvolt]

First I copied from the check that had a count (which, as you said, wasn't initialized that far up), then I wound up placing the intended replacement in the wrong place, by overwriting the part in the "you can't wear more than 2 X" by mistake. >.<

[Malkeus]

No worries man, this shit is confusing. At least you're trying ;)

[chaosvolt]

...and then something still failed. Whee.

EDIT: Extraneous ) due to yet more lazy copy-paste.

[Malkeus]

Parentheses, brackets, braces and commas can all go to hell...

[chaosvolt]

That would be it, yeah.

This is what happens when you start off basing a check of something that doesn't make use of the %s addon, then change your mind afterward. :V

[Malkeus]

I'm using codeblocks (bc free) and I couldn't get by at all without it's ability to tell me where something is declared and/or implemented. What are you using?

[chaosvolt]

I seriously need to install MSYS2 to do compile testing. Usually it's more Notepad++ along with trial and error. And even then I only recently went from doing source code edits with plain Notepad or Wordpad. >.<

[Malkeus]

Oh man, I can't recommend something like codeblocks enough then. You can easily search the entire source code for things, hovering over a function name will tell you where it came from, and you can right click and find declaration/implementation and it takes you there. It took a little effort but I got it to compile for me too. What OS are you using, maybe I can assist you with getting compiling working. At the least I can send you what I used to make it happen.

[chaosvolt]

Ah. Never tried using Codeblocks for editing, I'd only thought to use it for compiling. Er, trying to at least. >.>

And hopefully it'll work alright, just kept forgetting to try it when I had the chance. Mostly because the few times I go poking around in the source code, I don't find out whether it's faster and more efficient to compile-test until after I've thrown up half a dozen failed commits. ._.

[Malkeus]

If I compile in curses it only takes about 10 minutes. Curses is good enough for most testing.

[chaosvolt]

Hmm. Never tried that. And predictably I yet again hit glorious inconsistencies to compiling.md's guide to using MSYS2. It starts with it saying one of the updates the guide says to grab is outdated and thus skipped, then pretty much unravels from there. >.>

[Malkeus]

I sent you a link on the forums. It has mingw64 as the compiler. That guide is a clusterfuck. After you get the libraries installed, I had to change the Project -> Build Options -> linker settings tab to include the right libraries. Also make sure it's using the compiler you want, it defaults to some kind of default craziness.
This is my linker settings tab, I had to add several of those libraries: http://puu.sh/mksKj/477674d812.png

[chaosvolt]

Just...augh. I suck at actually handling the source, I suck at compile-testing, and so far all I've done is commit fuckups and get griped at for going the lazy route and using Jently to test this shit.

I'm starting to think I should just stop bothering with source-editing PRs entirely, if it's just going to be another string of stupid mistakes.

[Rivet-the-Zombie]

I'm starting to think I should just stop bothering with source-editing PRs entirely, if it's just going to be another string of stupid mistakes.

There's nothing wrong with making a mistake so long as you learn from it.

[chaosvolt]

Trying to. And failing, it seems. ._.

[Coolthulhu]

Setting up a compiler helps a lot with learning because with a compiler you can just throw bad code, correct it a bit and repeat until it stops being so bad that it doesn't compile.

I checked the tutorial and I think the dead link you were talking about is now replaced by a proper one. At least all the links I checked were actually working.
The tutorial isn't perfect because it has extra steps (who needs tiles anyway?), but it should be working now.

[chaosvolt]

Hmm. I should try it again, yes. Not sure which one to go with, since so far it's been three different compilers tried and just as many failures.

Still...sorry. Just got kinda set off by that. I know it's been nothing but fuckups on my end, and you don't really have time to constantly correct my mistakes. But when the first response to a question about what I'm doing wrong is to berate me for not being able to compile-test it, and call it "lack of respect" as if I'm intentionally trying to make things as much hassle as possible, is just...disheartening.

[Coolthulhu]

But when the first response to a question about what I'm doing wrong is to berate me for not being able to compile-test it

It wasn't the first response.
First I wrote what has to be done. Then I wrote a lot of the code for you.
Then I explained what and where is wrong. Maybe not directly, but that's because you should be able to figure some things out on your own by reading compiler's complaints.

Only then I added a note about you not compiling things.

as if I'm intentionally trying to make things as much hassle as possible

Not intentionally, but it looks like you're trying to skip a part of the development because it's annoying to set up, hoping that there is a way around it.

This causes problems for everyone. Including you, because you can't compile it fast AND because even if someone tested it, it would most likely end up having bugs you'd have to fix before merging.

If you have problems with installing the compiler, ask. I bet you could even get someone on IRC to explain the entire tutorial step by step.
But don't ask how to do stuff after skipping setting up the environment. The best answer to that is "go back and set up the environment".

I'm not going to close the PR or anything like that, but criticism is a part of the review.

[chaosvolt]

Sorry. Trying at the moment to follow the instructions for compiling (again), will likely get on the IRC once I'm done with the initial first steps.

And so far, I've found no evidence that compile-testing will actually be easier and faster than the way I've been doing it for, what, nearly a year by now (correction, about 10 months)? Even now I'm seeing a big potential issue in compiling: the need to add the required addons to the actual source directory.

My options are either to copy the entire source folder (which takes about as long as you'd expect 1.45 GB to) so that Github won't be trying to commit the added stuff to the pull request, or I'd have to move the needed files into it, then remember to move them elsewhere before commiting.

[chaosvolt]

And then there's the other issue. Given all the jokes I crack about the source code melting when I so much as look at it, it should be apparent that I inevitably tend to make some sort of mistake, and wind up at a complete loss for what I'm even doing wrong. If it's committed to the pull request, at least then others can look at it and find the predictable amateur fuckup that I completely failed to notice.

If I'm instead compile-testing it and not committing it, others can't actually look at my work, meaning I have to go and ask for help, then provide code excerpts of what all I'm doing when I eventually hit a problem and have no idea what I'm doing.

And even then, there's the possibility that when I post what I'm doing, I might miss important information in other files that might be the cause of the problem. And if I'm at the point where I don't know WHERE the problem is, it's rather obviously harder to provide those code excerpts, specifically excerpts that would actually reveal what part I fucked up.

[Coolthulhu]

Committing unfinished PRs is fine (as long as you keep the [WiP] in title).
Do what you can, commit when you can't (or when you could and did).

I'm seeing a big potential issue in compiling: the need to add the required addons to the actual source directory.

If you compile without tiles, sound and localization, you don't need those.
Instead of make TILES=1 NATIVE=win32 LOCALIZE=1 it's make NATIVE=win32 LOCALIZE=0
You can add RELEASE=1 for faster compilation (but less helpful messages on crash).

The second windows tutorial is more complicated, but it doesn't meddle with source directory.

[chaosvolt]

Hmm. Will be trying to get it set up then. For now though, I'm still at a loss for what to do now.

Just...sorry about this.

EDIT: Compiling in a nutshell:

[Malkeus]

http://stackoverflow.com/questions/29450016/o1-2-3-with-std-c1y-11-98-if-cmath-is-included-im-getting-error-hypo
However...this isn't an error I get when compiling which leads me to think you might be using an outdated mingw. What version do you have?

[chaosvolt]

However...this isn't an error I get when compiling which leads me to think you might be using an outdated mingw. What version do you have?

The one linked to by COMPILING.md. >_>

[Coolthulhu]

Try the tutorial under "Rough guide to building with only MSYS2".
It is newer and thus less likely to cause any problems.

[chaosvolt]

Well, time to repeat the mistakes I made when I tried that last time, I guess.

EDIT:

pacman --needed -Sy bash pacman pacman-mirrors msys2-runtime

I'm assuming that's supposed to be "bash pacman --needed -Sy bash pacman pacman-mirrors msys2-runtime" instead?

EDIT 2: Apparently Github format ate the "bash" and new line in that quote.

[Malkeus]

I'm a rebel and I always get the newest version of everything, it very very seldom is the wrong thing to do.

AngbandTK is an old ass game and I shouldn't have been surprised it wouldn't work with the brand new versions of it's requirements...

[chaosvolt]

Hmm. Straight-up copying all 3 lines into the command does jack shit.

[Malkeus]

My ancient talisman protects me from your druidic chants!
Malkeus gestures frantically with an old eyball on a string. It menaces with spikes of cuteness! Carvings depict a hexecution of beetles in human skin.

[chaosvolt]

Druidic chants? Something something tekeli-li. :V

Well, I'll see about compiling later on, if I get any farther in the attempt,

For now I'm still at that part where I'm stumped and would need someone more competent to figure out what I fucked up. >.<

[Night-Pryanik]

I can just upload somewhere my entire codeblocks folder with my working settings. Not sure if it'd wrote something in the registry, but you can try to download and use it nevertheless.

[chaosvolt]

That might help, unsure if it writes registry or appdata files, but is worth a try. ^^"

[Night-Pryanik]

@chaosvolt https://drive.google.com/file/d/0BwEAK44XlpItYXhTbGtKQ0s0eTQ/view?usp=sharing
I made some special version of source code for CodeBlocks to work properly - source code from github didn't worked for me. So if you want to try my version - just say, I'll upload it too.

[chaosvolt]

Ah, will look at it once I'm able to get back on the laptop, as will be out soon. Thank you.

[chaosvolt]

Well shit. Now merge conflicts. Greeeat.

[Coolthulhu]

This function doesn't return a string. String is text. A reference to an item isn't a string of text, it's item &.

item &player::best_shield()

[chaosvolt]

Ack. Is that all this needed to work? Scheisse.

Well...I'm actually already working on splitting this PR up. >.<

[chaosvolt]

Closing this now that the important parts have been split into separate PRs.