itch.io

Results: Failure.   No game produced.

Didn't get to dedicate too much time this year, but I learned a lot about emscripten as a build target for C projects.  The plan was to simply take any previous 7DRL entry, build it with emscripten, and publish the results.   Taking C and turning it into javascript has a little extra shimmy and shake to do.   The first time I tried this I thought it'd be as easy and switching one and and firing it off. A nice little feature on the side. But no.   The tough part here is every library you want to use needs WASM binaries.  It's a different build target. It needs all the code.   SDL comes pre-packaged with the emsdk, which is great. But libtcod and ncurses do not and so you have to build them with emcc.   

A couple of days digging into what's previously been done with ncurses on emscripten tells me there's still work to do there. I'd love to dive in and help debug this thing, but a 7DRL challenge isn't the best place. Switching over to a different project that used SDL, I learned a lot about what needs to be done porting over code.   There's a couple ways of bringing over files to load, and I got one working. There's also extra shenanigans you have to pull when dealing with big loops, which... well I never got to the bottom of. 

In the end I got one screenshot of the initial screen before I had to kill it as it was eating 1GB/sec of ram.

If you want to make something, use tools you know. If you want to grow, use tools you don't.

Yep, that was a bug in the makefile introduced in the last moment. Really embarrassing. libtcod likewise needs to be built. And you need base-dev for pkg-config to build libtcod.   Bit of a mess really. ALWAYS check your release.  Really, I should have used the laptop and walked through the whole process.   Lesson learned.

Well that's no fun.  Hmm... It's explicitly c99...

Ah, not a standard thing. It shows with -Wpedantic... but it's only a warning.
    There's a risk of ambiguity for what the compiler jumps to when you declare something at the start of a case, but the default behavior of gcc works well enough.   I think you've got some sneaky means of inserting -Werror into your build tools.

None of this was every planning on being -Wpedantic. It's just game, not life-critical code. Remember kids, the MISRA standard makes you MISRAble.

the game is really great,maybe the best one so far. The idea of fishy legs is funny. ALso the game machanics are fabulous. Overall I think it's an marvelous game.

Really cute :>

I love the design and the feature that you can pick your cat. There is a bug that your stuck at level 1 though the enemies and the cats are cute. so it's an awesome game.

cool idea but the AI seems a little funky. (one time one of the red guys just disappeared for no reason)

Good job making the game I really like it. the color in the game is fantastic. Keep making more awesome roguelikes.

A lovely exploration of command queues

Very pretty. Not super intuitive, but it's learnable.  Little things like "seg faults" means you can't run it while it re-loads. Before re-reading what shift-click does, not knowing what any programs do made me feel like a true script kiddie. Very authentic.  4.5 different types of health is a bit much. All in all, very enjoyable 7DRL.

I like the mechanic of the enemies being unseen. Having to shoot in random directions had me in a panic.   It's  a huge space to search, and the crystal economy makes it painful to see totems you just can't use. Once you get all the stones they should really point you towards the obelisks.

DEADLINE!!!

ok, I've got about 18 minutes. First off SUCCESS! WOOO! We have a traditional roguelike with map, monsters, items, and inventory. The twist I added was a heavy pet-centric theme with what I consider a neat smooth gradient elemental system three-way circle of death.

Downsides:

-Leashes got cut. too many bugs too close to launch. Meaning psycho monsters won't eat you if you give them too short of a leash. DOGs aren't all that special for being domesticated. You can't store your monsters in inventory capsules/eggs. You can't pull them out to battle other tower monsters. No leashing weakened and exhusted tower monsters and making them pets.

-No eggshells. This should have been easy given the framework around it. You were supposed to be able to move your pets around the elements

-Swords and shields are just fluff. potions and magic items don't exist.

-While I set up everything to dump that big idea list into monNames.c... I never got around to it. So the only special monster is DOG. That's it.

-The town is only one static screen. All the townies are identical.

-No shortcut merchant. I have them trying and dying instead. oof.

-No plot with the endgame EXE, the weav cult, the basement, scavengers. Content takes TIME.

-oof, and no traits or actions. So gameplay is restricted to bump-attacks, which is a far-shot from all the interactions and depth that pokemon games provide.

I'm happy with what I got, but I was hoping for so much more. As I get older and become a better dev, there are some important lessons that have REALLY helped out: Always have a working build. All production halts if you have a game-breaker. Bugs come first. Just writing down the plan of what's coming next is really helpful. Wrapping up bugs and getting to a commit so often left me wondering around thinking about TODOs and improvements. It's easy to get sucked into brainstorming when you really ought to be getting to the next feature. And don't be afraid to tackle big features. It'll be changes and bugs and you might rip out some code, some you even just wrote an hour ago, but work through it and move the game forward. Don't have sacred cows, be okay with change. Some cleanup can save you a ton of debugging later, but oh man, some cleanup efforts are just a waste of time.

Spotted a minor bug. One of the ghosts didn't end the game for me.  Left side. It looks like the blue ghost was painted right over the top, but I could keep moving

It's against the spirit of the challenge. Far too easy to "just put that off till after release".    

hooooooo boy. So this was a hearty 3 days. 

A dreary amount of time dealing with how to rotate things around in a grid. Eventually I just accepted that the rotation math was "more realistic" than what I had original planned.   "it's a feature!"

A late addition (as of about 3 hours ago) of adding pathing and mid-action collisions makes it far less ridiculous... but the real-time clock and delay makes it more annoying.   And there's no time-dilation between the player's speed and the guard's chasing. 

There's sort of a goal, obstacles, (the guards are surprisingly annoying),  and jumping over stuff. Most important there's a pretty fleshed out system of grid-based turn-based inertia. Which was my main goal.  So I'm calling this one a success.  
Time is out. So the remaining bugs with collisions and the complete lack of deadly-falls between rooftops will just have to be ignored.  All in all, for three days effort (6 if you count the last two attempts) I'm pretty happy with what I got.