General Discussion

[Xela]

I need help with timed progress bar in the last push. I made it based on this example, and yet it doesn't work like it supposed to 

Oki, gonna take check it out.

[DarkTl]

It fixes the stats going red... remove your variable and check it yourself. Or don't waste your time... two fixes is better than one

Of course I did it before pushing. Nami has 0 vitality by default thanks to the testing file. If you go to her equipment screen and select an  item from MC inventory, all her stats will become red without the $ tempc = "#F5F5DC" fix.

[Xela]

How do you want diving to work?

[DarkTl]

For a start I wanted the bar to work. Then I planned to activate random matrix at the same time with the bar, and once the bar is over, matrix is over too. The random matrix will use items with "diving" location, almost like the fishing screen.
So it will be random matrix with small invisible areas and limited time to find them.

[Xela]

Just pushed, scope your variables, there is no reason to pollute global scope with every little thing, it's already in not the best of shapes. Screens, functions, classes all have their own namespaces you can use. They is now even new way to scope variables to labels, it should get really good with the next release.

[Xela]

http://python.net/~goodger/projects/pycon/2007/idiomatic/handout.html
https://www.python.org/dev/peps/pep-0008/

https://youtu.be/OSGv2VnC0go

Some good stuff is in there.

[DarkTl]

That has nothing to do with python because it doesn't have screens like renpy. And the screen CTDed until I made the variables global like in the wiki example 

Just wonders of screen language again, when you cannot pass variables to a screen like you can to a function, not without the "default" statement.
 

[Xela]

Just wonders of screen language again, when you cannot pass variables to a screen like you can to a function, not without the "default" statement.

Yeah... this feels like a particularity bad design. You can access them nut you cannot mod them using default Actions.

[Xela]

BTW, I am getting more and more issues with Git, after some (small bit of) research, this was one of the first suggestions:

Code: [Select]

$ git gc --prune=now
$ git remote prune origin
(you open git extensions and click Tools ==> Git Bash to type in commands). You might want to do the same, especially if you haven't freshly cloned in a while.

[DarkTl]

I wrote a whole mess of logic where you can use 4 different labels that the game will autocall. You can reuse them, create new once and etc. Options here are limitless and very interesting, not to mention, potentially the most important towards gameplay.

I don't understand what are you talking about here. There is nothing unusual in your recent pushes.

[Xela]

I don't understand what are you talking about here. There is nothing unusual in your recent pushes.

I've coded that system over an year ago, it's quite powerful and adaptive and it calls a chain of four labels, if they exist. The concept behind the design is that you can modify the MC as you wish in those labels, create unique items, add traits and etc. Those labels are also reusable so you can sort of build MC out of different "blocks". I don't remember if the system allows to add icons to MC screen or not but we can take care of that later.

Just like with any other system, it'll take some trial and error but it doesn't mean that we cannot start messing with it.

[Xela]

Take a look/play around with mc_choices.rpy file, there are some comments there as well. It's in the mc_setup folder, I'll be back in the evening (switching to Android).

[DarkTl]

I wrote short testing code which should use the hidden matrix to hide 5 objects at the screen.

Either I don't understand something, or the hidden matrix is broken. By default there are zero areas at the screen. I know it for sure, because for testing I created them with 500x500 size, impossible to miss. But once I changed alpha of hidden matrix buttons to 1, there is always one single area at the top left corner, with zero randomness. Even though the corner is always empty with the old alpha = 0.001.

I dunno why your hidden matrix acts like that, even if I gave it incorrect dict somehow, it acts differently with different alpha.

I had the very same issues when tried to use it for fishing screen before making different screen with imagebuttons.

[DarkTl]

Take a look/play around with mc_choices.rpy file, there are some comments there as well. It's in the mc_setup folder

Sadly, the setup is absolutely incompatible with dual class system. I'd prefer a dual menu where you select traits/classes for MC's mother and father at the same time, which in turn give certain traits/classes for MC. Including dual classes if parents classes are different.

Also not all options should be hierarchical - in fact, most shouldn't. Strict hierarchy complicates things a lot.
For example, MC's mother could have any race, giving MC certain race bonuses. But race cannot be a part of big, strict hierarchy, it's not tied to class for example.

Basically, the setup screen is so old compared to the current concept that it would be easier to use multiple renpy menu statements to setup MC than use the setup screen.

[picobyte]

BTW, I am getting more and more issues with Git, after some (small bit of) research, this was one of the first suggestions:

Code: [Select]

$ git gc --prune=now
$ git remote prune origin

With this you can check your git history.

Code: [Select]

$ git fsck --full
 It may show dangling blobs. They are harmless, usually. Git tries to keep almost anything in history, even if you staged something once. They idea is that you can get back to the state of the code at that time. I believe after a while you can get some that lost their reference. The prune will indeed remove those but you may also lose some history.