Exerro, on 26 November 2017 - 07:07 PM, said:

Hmm. That does sound interesting.

Exerro, on 26 November 2017 - 07:07 PM, said:

I could add some half-baked animation support with some sort of CSS, also I could allow scripts to be loaded as the page is rendered. I'll think about that.

Exerro, on 26 November 2017 - 07:07 PM, said:

Sure! I'll have to do some hacky things, like having only 4 or so colors available to add custom values (luckily the browser only uses a gray/white theme, so it might work out well)

Exerro, on 26 November 2017 - 07:07 PM, said:

I was doing a thing a while ago with OneOS and pkg to install apps, so I might possibly be able to add that.

Exerro, on 26 November 2017 - 07:07 PM, said:

That was something I plan to add. In some things, I have the align variable so I could use something with that.

Exerro, on 26 November 2017 - 07:07 PM, said:

Probably could do that with messing with Y. I'll think about it.

Exerro, on 26 November 2017 - 07:07 PM, said:

I was thinking about adding some lua-table configuration or something halfway through writing this post, so I could probably add something like a "lst" (lua style table) tag or something.

I also noticed you had mentioned protocols. I do have protocols in Mercury, with local:/// (the extra / is not a typo, there is just no host), and rdnt://, and pages built into the browser as mercury:// domains, like mercury://home.

Wilma456, on 26 November 2017 - 07:40 PM, said:

No, it doesn't. It uses its own subset of XML with HTML like tags, as much as I can.

Following the release of Firefox Quantum, I decided to make a web browser, a kinda-rendering engine, along with my own subset of HTML using a url parser, and a XML parser. The browser is more hacked-together than anything, but it currently features modem support, local file support, scripts, <text> with x,y and b(ack)g(round) and f(ore)g(ground) support, <a> but only href and x/y/fg/bg, <br>, <hr> using teletext characters, which makes it nice, buttons which invokes the scripts, and probably other things I don't remember.

The script part is very sandboxed. No print, no write, no nothing. You got math and string APIs, then ask/alert functions, a copy of the XML parsed page to edit it, and other things. It is specifically made to be really restricted and secure. Also, scripts can ONLY be invoked by using buttons. You define the scripts, give it an id, then you create a button with the script id as a tag. And only then will scripts work.

There is also no scrolling support, either. You can probably implement this in your page using buttons and scripts.


So, I'm not entirely sure if I explained everything well (probably not), but I want your thoughts and suggestions for this. It'd be nice.

Thanks for your responses in advance.

Exerro, on 12 November 2017 - 03:25 PM, said:

It is a table holding objects. Keys are IDs, values are OOP classes/tables.

Anyways, I think I've fixed it. It was more of an error on my part, but I think the problem was the names on the for loop and (i)pairs, maybe even my misunderstanding of this stuff. after further testing, it was. also, i didn't edit because of my internet being bad n stuff.
But, thanks anyways.

SquidDev, on 12 November 2017 - 02:34 PM, said:

I can't just use pairs. If I throw another button into the mix, I need to use ipairs.
Basically, I have 2 buttons. They both work perfectly fine using pairs. Can't use ipairs though. (probably because of the way i'm storing them: id, then OOP class)
If I throw another button into the mix, making it 3 buttons, I can use ipairs, but not pairs. If I use pairs, then nothing works.

Here's a little more code, this is from the API

elseif e[1] == "mouse_click" then
            --find what it clicked on
            local x = e[3]
            local y = e[4]
            local button = e[2]
            for id,t in pairs(self.Scene.View) do
                if t.X <= x and t.Width + t.X >= x then --x matches
                    if t.Y <= y and (t.Height or 1) + t.Y >= y then --y matches, this is the right object
                        if t["OnClick"] ~= nil and type(t["OnClick"]) == "function" then
                            t:OnClick(x,y,button)
                        end
                        break
                    end
                end
            end

I look through all the objects in the scene using pairs, see if the X and Y matches up (and they do, but not in ipairs with 2 or pairs with 3), then call the OnClick function.
It works with pairs using this code:

local me = QuickDraw.Init() --initing stuff
QuickDraw.AddObject(button) --defined earlier
me.Scene.BG = colors.black
me.Scene:Add("button", "b1")
me.Scene:Add("button", "b2")
--defining properties and stuff

...but if I through in another button, (me.Scene:Add("button","b3")), then pairs doesn't work and I have to switch to ipairs.

That's the best that I can explain it, anyway.
my internet sucks at the moment too, so its hard to post things and it takes a long time for it to go through

KingofGamesYami, on 12 November 2017 - 02:03 PM, said:

I have a program that tests the various functions of my API, and when I have 2 buttons, I can click neither of them unless I'm using pairs.

KingofGamesYami, on 12 November 2017 - 02:03 PM, said:

I probably could, but most likely not.

I have a GUI API that allows me to create, well, GUIs. Problem is that I need to switch between ipairs and pairs in order for clicking to work.

Here's where clicking is handled (in a for loop), and sorry for my cruddy coding and OOP.

...
for k,v in ipairs(self.Scene.View) do
...

Thanks in advance.

QuickPaint is a simple paint program. And when I say simple, I mean its a WYSIWYG (what you see is what you get) program.
The files are also compatible with Paint (from the default ROM). Although it doesn't like loading files that were originally made by Paint, you can load files that were made with QuickPaint and also touched by Paint. (loading directly from paint = bad; making picture in quickpaint, then editing a little in paint = ok) Why? Paint has support for transparency. QuickPaint doesn't.

What makes it Quick?
It isn't. If anything, its a little slower due to the way I program my GUIs!
The "Quick" prefix comes from programs I make, because they're simple. (and no, you can't find these on the web)

What are its features?

• Makes pictures
  
• Somewhat cross-compatibility with Paint
  
• Clears (by pressing C)
  
• Saves/loads
  
• Click the red circle at the upper right hand corner of the screen to quit.

There really isn't much to say. It's a paint program!

How to use:
You use it like a Paint program.
Keybinds: (S)ave, (L)oad, ©lear
Press ? to get help from the program.

Where can I download this?
You can get it on PasteBin - pastebin get 1KAd2e20

Licensing
Under GPLv3. Just PM me if you want to do something like include it in your OS.

Dave-ee Jones, on 22 August 2017 - 01:10 AM, said:

Yep, you got it right. I made this to make it easier for prototyping (and doors) but it could be used for simple modem projects.
The side and the reply channel aren't used because of my standards of modem thingys (throwaway the side, and assume reply channel is sender channel + 1) but in a way it kinda makes it easier, I guess.

adore the not so amazing

SimpleAdore

wireless made kinda simple, i guess

So... what is this again?
This is a rapid prototyping wireless library thingy that makes wireless easier.
You only need to call about 3 or so functions to get on the air, plus transmitting and receiving is a lot easier, considering you just call the Recieve function or the Transmit function and it does all the work for you.

Why is SimpleAdore named SimpleAdore?
tl;dr recycling code
I was developing a Door Lock API named Adore, and nothing worked so I rewrote it but 90% of it was non-door-lock code stuff so I took that out, cleaned it up a bit and here we are.

What can this be used for?
Anything that uses modems/rednet. For testing, I used this to make a what-you-see-is-what-you-get wireless door lock (which will be released soon).
And guess what? It is 78 lines smaller than my door lock without the library on the server side, and it is 61 lines smaller on the client side.

Also, for the license: this is under a GPLv3 license.
You may use this in your program, just attribute me and you're in the clear.

Where can I download this?
From PasteBin - pastebin get mXGfG7Kw SimpleAdore


Current issues:

• No check for if you actually setup the modem by running WrapModem
  
• Automatic modem discovery thingy is touchy/doesn't work

SquidDev's right. If you want to make a multitasking OS, and you aren't that good at Lua: you are going to be in for a very long time.
Things you are going to have to learn: sandboxing, courotines, the event system, filesystem, a ton of Lua syntax, and much much more.

A good idea is to browse the CC wiki if your new to Lua because it will provide you with a lot of information and tutorials and such.

TL;DR: learn lots of lua, browse the wiki, use tutorials for help, and use Ask A Pro if your really stumped.

WattanaMC, on 06 April 2017 - 07:36 PM, said:

OneOS isn't being developed anymore.
Which, means it isn't being fixed anymore.

KingofGamesYami, on 14 February 2017 - 03:44 PM, said:

Ah, my bad. I'll write a little notice now.

Bomb Bloke, on 14 February 2017 - 12:03 AM, said:

There was none, so I just ported it. But, I still gave credit.

Bomb Bloke, on 14 February 2017 - 12:03 AM, said:

Yeah, I could. I'll update it later.

LZW is a sorta not really kinda compression algorithm thingy created by Abraham Lempel, Jacob Ziv, and Terry Welch hence the name, Lempel Ziv Welch. It is used in the *nix program compress, and GIF images.

I have ported somebody's implementation of it in Lua, you can find the original here.

It isn't really that bad, but it has it's quirks. (Shaves 1 character off of "This is a test.", your mileage may vary.)
It is also fast, less than a second for "This is a test."


You can download it with: pastebin get WZKTR9q4 lzw
Yeah, I know about the typos. I originally thought it was LWZ.
Also, sorry about the sad mush of a post I made. Couldn't upload much because were dealing with a storm at the moment

The API is licensed under MIT because of the original file.
You can find a layman term page for MIT here.

TechnicalCoding, on 02 February 2017 - 10:14 PM, said:

Just because your linting something doesn't mean its completely crash proof.
The least you could do is loadstring and xpcall it, but that means when it crashes, instead of getting sad it will do a function instead of erroring out.

Piorjade, on 02 February 2017 - 10:09 PM, said:

In past experiences when testing Unity 8 it worked fine for me.
Haven't tested XFCE yet, I'll do it later tonight.

The only thing I could think of that might be slowing it down is the rendering.
I can't recall any time I've had to wait a few seconds for a Java program to show up to be completely honest.

Anavrins, on 02 February 2017 - 09:57 PM, said:

No, wasn't referencing your post, and yes I know that method has more holes than account password management in Windows.

[offtopic]No offence, but your program reminds me of the OneHalf virus of the MS-DOS days.[/offtopic]

This is a project I've been working on for the past few days. It is a EFI-like bootloader (Kinda like the Mac bootloader)

It works like the Mac bootloader (or so I think), basically it gives you a second to put in a boot modifier and then loads Mac OS.
Basically, I have a front end (shown below) which is basically a looping loading screen. It may look like it is doing nothing but in reality it has a script running in the background. The script sets up the environment, then says "Hey, I'm finished, you can run this function now!". The function in question (ChameleonDone) MUST be defined in the script, or else the bootloader gets sad.



So far, I've been able to make OneOS able to run under Chameleon.

(You don't get to see the animation since it's in a emulator, it's that fast.)
(Another note, for some reason OneOS boots faster in Chameleon than it's native bootloader. It might be that I stripped a lot of code that was not needed though)

So, this is the tl;dr version of what Chameleon does:

• waits for boot modifier (z for shell, x for custom startup script)
  
• launches /boot
  
• while the animation is playing, the script is doing things to setup the environment (?)
  
• once the script is done it will launch the ChameleonDone function
  
• ???
  
• Profit!

There is a log function too, but it is mainly used for scripts.

I am going to release the source sometime tonight, along with the OneOS script.
Chameleon
OneOS start script (save to /boot)

Anyways, I'd like your thoughts on this and on how I could improve it.
Thanks for any feedback.

I'm not entirely sure if I'm/anybodys going to use this, but I decided to make it anyway.

I've been running Plasma 5.8 on Kubuntu 16.04 and haven't been experiencing problems.
yet.
The funny thing is that I'm on a wayyyy lower end card and computer (AMD A10 with built in Radeon (Linux says its a AMD Gallium 0.4) on a laptop, sigh)
At least I think it's lower end.

tl;dr such working wow no problem much fast

You could do some encryption (not recommended though)
Maybe have the user input a password to decrypt the system files on startup, so if somebody puts in a disk they wouldn't be able to modify the files without knowing the users password.
To be honest, I find it really flawed, but its worth a try.

Another idea if your testing security is to publish it on GitHub and get a few people to break the security.
One more idea is to go do something like Google does when it finds out that Chrome OS has been tampered with. Still, flawed.

The only other piece of advice I can give is to think like an attacker:
Think of attack vector. Try attack vector. Fix if required.
Rinse and repeat.

A good idea if your making a program is to use [[ and ]] for multi line strings(?).
If you need to supply some custom variables (maybe iteration of program) a good idea is to concatenate your variable to the start.
Here's a sample program you could hack on:

local code = [[
print("This is program number " .. n .. "!") --Notice the absence of n being defined.
]]
write("How many programs do you want to make? ")
inp = read()
amount = tonumber(inp)
if inp == nil then --Check if its actually a number
  error("Not a number")
end
for i=1,amount do --basically for(int i = 0; i < amount; i++) for some languages
  print("Program " .. i)
  local h = fs.open("program" .. i, "w") --open file in write mode
  h.write("local n = " .. i .. "\n") --add the n variable
  h.write(code) --write our code
  h.close() --close the file
end

If you run it, and give it say 5 programs to make, run one of them and it should output "This is program number <number>!"

Lupus590, on 30 December 2016 - 03:41 PM, said:

Yeah, I'd go for a state changes clause. I don't mind people making changes, in fact it's more of a "state me and it's ok" kind of thing.
And I know Creative Commons isn't really for code, more of media. I was starting to regret putting that in as my default license, I'll change it later on today.

This was more of a minor project to be honest, I love FOSS and all of my minor programs are under some sort of Creative Commons (if I remember correctly) because I thought it was kinda fair, well, its more of what you see is what you get and if your unsure/need permissions PM me.

I'll update my licenses to maybe something more friendlier, like GPL or MIT.

emerald, on 30 December 2016 - 03:10 PM, said:

Yeah, go ahead.

Bomb Bloke, on 02 December 2016 - 03:34 AM, said:

Stekeblad, on 02 December 2016 - 03:14 PM, said:

Oh, sorry for my mistake then. Well, technically it could by using string.dump on a function combined with loadstring, but I'm not exactly sure if that'd work or not.

Couldn't you just use textutils, instead of writing some functions? (But I do admit, nice code)
Something like

function save(table, file)
local h = fs.open(file, "w")
h.write(textutils.serialize(table))
h.close()
end

--fyi, i did not make this code really nice and functional, its a "do as i say, and dont complain" kinda thing

function load(file)
local h = fs.open(file,"r")
local r = h.readAll()
h.close()
return textutils.unserialize(r)
end

Not trying to criticise you, but CC already has an API for it. You did do a good job of coding it though.
[offtopic]Now I'm having flashbacks to creating the pkg encryption function, bad times.[/offtopic]