LyqydOS Beta/Development

procman:217: attempt to index ? (a boolean value)

Spoiler

inventor2514, on 29 October 2012 - 01:20 AM said:

Excellent OS! I did notice that the mouse_scroll event x and y values are not adjusted like they are with mouse_click when I was messing around with events.

BigSHinyToys, on 29 October 2012 - 02:23 AM said:

This is very well done. I have found one bug.
upon attempting to resize the Run Program box

procman:217: attempt to index ? (a boolean value)

Also can I have permission to start new windows and processes from [Advanced Computer] Mouse File Browser

And one idea maybe unbordered window.

Spoiler

My program currently is not expecting the value from term.getSize to change so doesn't resize very well maximize then restart it fixes this until I fix that.

elseif sEvent == "window_resize" then
  w, h = term.getSize()
  redrawText()
  redrawMenu()

BigSHinyToys, on 29 October 2012 - 07:56 AM said:

I have modified Mouse File Browser to be compatible with LyqydOS so no more flickering wildly and resize no longer crashes lol.
This can run stand alone or from LyqydOS still as well. [Advanced Computer] Mouse File Browser


@Lyqyd

have you found a way to make paint run without cashing ?

also will the API have access to a function for creating a new process and window if so can I integrate it into Mouse file browser ?

awsumben13, on 29 October 2012 - 09:59 AM said:

wow this looks really good, i'm jealous :P/>/>. does it support multi-tasking and a suggestion: instead of using ascii characters to border the windows, have the window a certain background colour so it doesnt need a border at all

ElvishJerricco, on 29 October 2012 - 09:52 PM said:

Very cool OS. You might consider making use of the mouse_drag even to allow people to drag windows to move them or drag their corners to resize. Also it'd be nice if windows stayed open until you closed them even if the process finished. That way if i, for example, misstype a pastebin command i can still see the errors.

inventor2514, on 29 October 2012 - 01:20 AM said:

I did notice that the mouse_scroll event x and y values are not adjusted like they are with mouse_click when I was messing around with events.

ElvishJerricco, on 29 October 2012 - 09:52 PM said:

You might consider making use of the mouse_drag even to allow people to drag windows to move them or drag their corners to resize.

rm .lyqwin

1,1,19,9,true,rom/programs/color/paint

procman:541: attempt to index ? (a nil value)

540 elseif event[1] == "mouse_scroll" then
541  event[3] = event[3] - processTable[process.focusWindow].window.x
542  event[4] = event[4] - processTable[process.focusWindow].window.y
543  if event[3] > 0 and event[3] < processTable[process.focusWindow].window.w - 1 and event[4] > 0 and event[4] < processTable[process.focusWindow].window.h - 1 then
544   resumeProcess(process.focusWindow, event)
545  end
546 else

1. Not entirely sure what you mean by this, or how it would be a useful thing.
2. Not entirely sure what you mean by this.
3. Apps designed for LyqydOS are in the pipeline, but all portions of it are planned to have backwards compatibility with non-advanced computers.

Leonardoas26, on 11 December 2012 - 08:47 AM said:

I think I found some glitches.
1. Resizing windows is difficult.
2. It crashes if you open the startup/start program.
3. If stops behaving properly when you type "cd (something)" in the Run Program box.

Lyqyd, on 11 December 2012 - 11:38 AM said:

Leonardoas26, on 11 December 2012 - 08:47 AM said:

I think I found some glitches.
1. Resizing windows is difficult.
2. It crashes if you open the startup/start program.
3. If stops behaving properly when you type "cd (something)" in the Run Program box.

I'm not really sure what you mean by difficult. It's a little quirky right now as far as the window border is concerned, but it's pretty easy for me. Can you provide more details? :-)

Hmm. Are you trying to run the start program from the run program box? If not, I'm not sure what you meant. The start program isn't meant to be used that way, but I'll take a look at making it not break things at least.

What does it do when you run cd from the run program box? I'm guessing it has then changed directory (as would be expected), but "behaving properly" could go either way at that point. Does it do something besides just act like it's set to whatever directory you cd'd to?

RunasSudo, on 16 January 2013 - 01:24 PM said:

I just looked at the source code… I think I'm gonna have a lie down now……….

urielsalis, on 17 January 2013 - 09:28 AM said:

use pixels instead of text is more intuitive

Lyqyd, on 17 January 2013 - 07:55 AM said:

Heh. It takes a fair bit of code to get proper windowing behavior put together!

urielsalis, on 17 January 2013 - 09:28 AM said:

use pixels instead of text is more intuitive

RunasSudo, on 17 January 2013 - 01:25 PM said:

Lyqyd, on 17 January 2013 - 07:55 AM said:

Heh. It takes a fair bit of code to get proper windowing behavior put together!

What's with all that crazy "create a new windowcompositor and then rawset some of its variables to ones I've defined" stuff?
My PoC windowing "OS" (see sig) does that with a fair bit less complexity. I think. Maybe.

Lyqyd, on 21 January 2013 - 04:44 PM said:

LyqydOS has been updated with some slightly better multi-tasking (fixes a couple minor issues), and an optimization in the compositing system that should significantly reduce flicker in games. There are further updates I am working on that finally add a window manipulation API so that programs can set their own window caption, size, position, minimum and maximum size, and a few other things. A feature that's a bit further out is a "fullscreen" mode that games written with LyqydOS support will be able to take advantage of, that sacrifices certain windowing features to gain faster screen manipulation.

anonimo182, on 30 January 2013 - 12:49 PM said:

Bug: When you minimize a window, if the cursor was blinking it will still blink after minimizing
When you resize the run box, it will read until its default limit, not to its resized version

Lyqyd, on 11 March 2013 - 11:57 AM said:

I expanded the window decoration functionality today to include text color and background color, so this sort of thing is possible now:

Spoiler

This change isn't ready for release, but I hope to push it out soon. :)/>

Sora the Hedgehog, on 14 March 2013 - 04:34 PM said:

Wow, I really like this, very well built…

Quick feature request. Could you make it so that left clicking on the "Lyqyd Menu" lists "commonly used" programs stored in..
I dunno, /LyqydOS/common-programs? And having right click do the current "run program" behavior? That would have me sold for sure.

And a question : Interface API? I'd love to write programs for LyqydOS.

Anyways, I like it, and I'll be sure to install it on my Adv. Terms. :)/>

~Sora

oeed, on 14 March 2013 - 07:06 PM said:

Lyqyd, on 11 March 2013 - 11:57 AM said:

I expanded the window decoration functionality today to include text color and background color, so this sort of thing is possible now:

Spoiler

This change isn't ready for release, but I hope to push it out soon. :)/>

Hmmmm :P/>

Spongy141, on 15 March 2013 - 04:48 PM said:

Lol, nice OS, but how the hell do you have almost 300 fps!!! Dude do you got like 32 gbs of RAM?

Lyqyd, on 30 October 2012 - 12:24 AM said:

I don't think leaving windows of dead processes open would be a good thing overall. There are certain cases (like your example) where it would be nice to see the output after the program ends, but I think that the annoyance of other programs not closing when they should outweighs it. For instance, it would be very annoying to exit the edit program using its internal menu and then also have to close the window manually. It would also be a departure from the way most window managers work, in my experience.

pcall(shell.run, table.concat({...}, " "))
os.pullEvent("key")

 
if not term.isColor() then
  nonColorMode
else
 colorMode()
end

EthanAG, on 25 May 2013 - 07:57 AM said:

Also lyqyd i had a idea maybe if you have a color specific code for advanced computers and non advanced computers:

if not term.isColor() then
  nonColorMode
else
colorMode()
end

Thats just what i had in mind

if process.windowCompositor.target.isColor() then
    pTable.window:setType("standard")
    pTable.window:setDecorations("pretty")
end

Lyqyd, on 30 June 2013 - 02:06 AM said:

I just pushed out an update that refactors a number of things in the code to make it easier to add interesting features down the road. All of the window interactions involving temporary windows (such as the action menu for a window, or the run program dialog box) are now done with a single service that's started when the process manager starts. This means that the system won't eat an extra process ID for every process the user starts.

Also new in this release is the debug utility that superaxander suggested. The usage is simply to prepend "debug " to the command/program you want to try to run. This will hold open the window it runs in until after a key is pressed once the program has stopped running. That allows a developer to view error messages their program throws under LyqydOS.

I've updated the other two screenshots in the original post to reflect the new appearance and features, including a shot of the "ps" program in action. You can see the process ID and the program's path for each running process. Note that not every process has a window (lws, for example) and that the lua window in that shot is minimized.

Most users who have the installer file still should be able to update simply by re-running it, as it will fetch a new copy of the installer before attempting to install.

Lyqyd, on 12 September 2013 - 11:00 AM said:

If anything, it's written imperatively, not functionally. "Functional" programming has a specific definition that doesn't simply mean, "uses functions". OOP styling isn't always the right tool. I try to use the correct tool for the job wherever possible. A couple of my APIs could be somewhat described as "OOP", at least as much as Lua can be OOP. I am not someone who tries to use tools where they don't make sense simply because they're "cooler". I'm curious as to where improvements could be made by changing the way the code is written. Also, if you have any questions about the code, I'd be glad to answer them.

startup:49: attemp to index ? (a nil value)

logsys, on 17 November 2013 - 08:32 AM said:

I would love to be there a big advanced monitor version

Systemater, on 18 November 2013 - 03:57 PM said:

I guess this is the best programming going, but mostly like the NDF-Jays but now its finished but you can add abit of your own programs in this just for tips. I'm sorry if my Engish is bad.

Timendainum, on 01 December 2013 - 06:05 PM said:

Any plans for multi-user/login support. Integration with LyqydNet?

How about something like this. You have a login server, which acts like a domain controller of sorts. You can login to LyqydOS with a network un/pw. Then add in a file server, email client, and such.

logsys, on 12 December 2013 - 04:26 PM said:

1. Your GUI is not that good in terms of folder and file explorer. A little change to the GUI and it would get done 2. You would need time to search and think about this Now it's in your hands to think if this is good enough to make and will make you have a better OS. Have a good time!

Lyqyd, on 13 December 2013 - 10:52 AM said:

LyqydOS is just starting to come into its own with some recently introduced changes that haven't been pushed to the main release yet (such as integrated modem management and better event hooking for services). Features such as an included alternate shell, a file browser, and quite a few other features are planned for eventual inclusion, either by default or easily added on in the installation process or via an included package manager. You can see the current state of the code on the develop branch on my github. The package manager is in a repository called CC-packman if I recall correctly.

It is meant to be a fully featured system, I am simply taking my time to make a good foundation to base it on before I try to go for all the shiny stuff.

logsys, on 14 December 2013 - 10:46 AM said:

Also, try get CC-Forums admins to help u in it. Again, awesome job!

Lyqyd, on 15 December 2013 - 02:47 PM said:

Yep! That allows me to optimize screen writes by finding the largest contiguously colored chunk of the current line and write it all at once. Between that and not redrawing unchanged portions of the screen, the upshot is that there are usually very few writes that need to be made on each yield, which saves bandwidth and helps to reduce flicker and slowness.

Lyqyd, on 15 December 2013 - 03:43 PM said:

That would either require a change to tables of tables of characters (as opposed to the current table of lines), or it would require a lot more work on the parts of the code that put the tables together, since I couldn't then overwrite the strings on a character-by-character replacement basis, I'd have to parse out where each color identifier was and handle it separately. Exponentiation doesn't seem to me to be terribly slow, so doing two of those operations per write isn't too hard on things. I think it would be slower to try to walk the table to find the length of the largest contiguous chunk, and I think it would be slower to have to parse out a string of numbers to try to find where to overwrite it. If I were simply overwriting the entire screen on every yield, it might be easier to use a table of tables of characters/numbers, of course.

Symmetryc, on 15 December 2013 - 02:57 PM said:

That way you wouldn't lose efficiency when you have to convert the hex code into decimal and then raise 2 to the power of it (which, correct me if I'm wrong, takes a good bit of processing power, especially if 2 is being raised to 10, 11, 12, etc.) because you could just use tonumber(color).

Edit: So like you would have "512 512 512 8 1 " rather than "fffab23"

GravityScore, on 23 December 2013 - 04:33 AM said:

Symmetryc, on 15 December 2013 - 02:57 PM said:

That way you wouldn't lose efficiency when you have to convert the hex code into decimal and then raise 2 to the power of it (which, correct me if I'm wrong, takes a good bit of processing power, especially if 2 is being raised to 10, 11, 12, etc.) because you could just use tonumber(color).

Edit: So like you would have "512 512 512 8 1 " rather than "fffab23"

Raising something to a power is easy for a computer. It's just a simple bit shift, which is probably one of the quickest function a computer could perform. You do need a basic understanding of binary to get this though. Tonnes of stuff on the web for that.

Say you want to raise 2 to the power of 8:
2 in binary is 0010. When you perform a bit shift, all you're doing is moving all the 1s in the binary number to the left a certain number of places. So if we do a bit shift of 1 on the number 2, we get 0100 (see how the 1 moved 1 place to the left?), which in decimal happens to be 4. A bit shift of 2 results in 1000, which happens to be 8. Bit shift of 3 is 0001 0000, which is 16, 0010 0000 is 64, 0100 0000 is 128. See the pattern? So to raise 2 to the power of 8, we just perform a bit shift of 7, giving us 0001 0000 0000, or 256. You can perform a bit shift in lua using the operator << for a left bit shift (moving the 1s to the left), and >> for a right bit shift (moving the 1s to the right). You can use a bit shift in CC Lua using the bit API. Just bit.blshift(2, 7) (which gives 256) or bit.brshift(256, 7) (which gives 2) for left and right shifts respectively.

GravityScore, on 23 December 2013 - 04:33 AM said:

Symmetryc, on 15 December 2013 - 02:57 PM said:

That way you wouldn't lose efficiency when you have to convert the hex code into decimal and then raise 2 to the power of it (which, correct me if I'm wrong, takes a good bit of processing power, especially if 2 is being raised to 10, 11, 12, etc.) because you could just use tonumber(color).

Edit: So like you would have "512 512 512 8 1 " rather than "fffab23"

Raising something to a power is easy for a computer. It's just a simple bit shift, which is probably one of the quickest function a computer could perform. You do need a basic understanding of binary to get this though. Tonnes of stuff on the web for that.

Say you want to raise 2 to the power of 8:
2 in binary is 0010. When you perform a bit shift, all you're doing is moving all the 1s in the binary number to the left a certain number of places. So if we do a bit shift of 1 on the number 2, we get 0100 (see how the 1 moved 1 place to the left?), which in decimal happens to be 4. A bit shift of 2 results in 1000, which happens to be 8. Bit shift of 3 is 0001 0000, which is 16, 0010 0000 is 64, 0100 0000 is 128. See the pattern? So to raise 2 to the power of 8, we just perform a bit shift of 7, giving us 0001 0000 0000, or 256. You can perform a bit shift in lua using the operator << for a left bit shift (moving the 1s to the left), and >> for a right bit shift (moving the 1s to the right). You can use a bit shift in CC Lua using the bit API. Just bit.blshift(2, 7) (which gives 256) or bit.brshift(256, 7) (which gives 2) for left and right shifts respectively.

MudkipTheEpic, on 23 December 2013 - 12:59 PM said:

Don't you mean Java?

————-
| Progname |
————-

bjornir90, on 18 April 2014 - 07:53 AM said:

Hello ! I've got a question for you, lyqyd : I want to study how you did all the multitasks things bit i'm a bit lost on the github ^^. Do you mind to guide me ? :)/>/>

apemanzilla, on 18 April 2014 - 12:57 PM said:

bjornir90, on 18 April 2014 - 07:53 AM said:

Hello ! I've got a question for you, lyqyd : I want to study how you did all the multitasks things bit i'm a bit lost on the github ^^. Do you mind to guide me ? :)/>/>

He used the coroutine API to let two different threads run in parallel.

Symmetryc, on 18 April 2014 - 03:54 PM said:

apemanzilla, on 18 April 2014 - 12:57 PM said:

bjornir90, on 18 April 2014 - 07:53 AM said:

Hello ! I've got a question for you, lyqyd : I want to study how you did all the multitasks things bit i'm a bit lost on the github ^^. Do you mind to guide me ? :)/>/>/>

He used the coroutine API to let two different threads run in parallel.

Except for the fact that they don't actually run in parallel.

bjornir90, on 18 April 2014 - 07:53 AM said:

Hello ! I've got a question for you, lyqyd : I want to study how you did all the multitasks things bit i'm a bit lost on the github ^^. Do you mind to guide me ? :)/>

Spoiler

3141592651, on 30 July 2014 - 08:52 PM said:

Hello Lyqyd,

I seem to have discovered a bug. Sometimes when you resize shell to be smaller [from the corner button] and it is in the top left corner of the screen, it causes this error:

window:199: vm error: java.lang.NegativeArraySizeException
Press any key to continue

Begin Edit 1
Also, do you have any plans to create an API for LyqydOS program/application development? Example:

--Sample Program

thewindow = lyqydOS.createWindow()
thewindow.setTitle("Test Program")
thewindow.setSize(5 ,5)
thewindow.setPos(3, 3)
thewindow.setResizable(false)
thewindow.setTextColor(colors.green)
thewindow.setBackgroudColor(colors.white)
thewindow.write("Hello!")
sleep(3)
thewindow.close()
-- The code can continue, but the window will close.

Etc..

Or if there is such an API, can someone tell me where to find it? I want said API because I believe this OS has potential and I want to use it as the OS for all the workstations in my CC office building, but I would need to make custom programs to do so as the office building has existing CC infrastructure; for example, print servers, backup severs, a file server, and a router controlling all the sub-domains in the wired network.
End Edit 1
Begin Edit 2
When and if you do press a key, the computer shuts down.
End Edit 2
Begin Edit 3
And if you are not going to do the API any time soon, would it be OK with you if I made a derivative of LyqydOS for networks that had said API and post it in a separate topic and say it's a derivative of LyqydOS/give you credit for the base?
End Edit 3

DoomRater, on 26 November 2014 - 03:08 PM said:

That's pretty funny because OneOS handles this beautifully and pretty much the way I expect. I'm not talking about malicious programs, by the way, I'm talking about programs that have no knowledge they're being run on an OS that previously assumed they are the only thing running on the computer. If a program has knowledge of an OS, then SURE, it could get around redirection of os.shutdown() just fine. I don't care about such programs because they were either made to be malicious or intend to tell the OS that it needs the entire computer to reboot and the OS can handle the closing of all other programs first. My point is that the OS should always have more control over shutdown and restart than the programs running on it do. If that's too much work for you, or for whatever reason doesn't fit your philosophy, though, well, I can just keep using OneOS.

DoomRater, on 26 November 2014 - 03:08 PM said:

That's pretty funny because OneOS handles this beautifully and pretty much the way I expect. I'm not talking about malicious programs, by the way, I'm talking about programs that have no knowledge they're being run on an OS that previously assumed they are the only thing running on the computer. If a program has knowledge of an OS, then SURE, it could get around redirection of os.shutdown() just fine. I don't care about such programs because they were either made to be malicious or intend to tell the OS that it needs the entire computer to reboot and the OS can handle the closing of all other programs first. My point is that the OS should always have more control over shutdown and restart than the programs running on it do. If that's too much work for you, or for whatever reason doesn't fit your philosophy, though, well, I can just keep using OneOS.

Jiloacom, on 26 November 2014 - 03:27 PM said:

I've seen soooooo many install programs on this forum, that if it finds something unexpected (Like the computer isn't advanced), nope, no error() for it, it just shuts off the computer. So I agree with you on the fact that the Operating System (regardless which one) needs to at least help prevent against those kinds of things. Maybe just having a dialog box saying:

This program wants to use os.shutdown() . Allow it?
Yes : No

Lyqyd, on 26 November 2014 - 04:41 PM said:

Jiloacom, on 26 November 2014 - 03:27 PM said:

I've seen soooooo many install programs on this forum, that if it finds something unexpected (Like the computer isn't advanced), nope, no error() for it, it just shuts off the computer. So I agree with you on the fact that the Operating System (regardless which one) needs to at least help prevent against those kinds of things. Maybe just having a dialog box saying:

This program wants to use os.shutdown() . Allow it?
Yes : No

That sounds like incredibly poor design, which should probably be corrected on their end. :P/>

GopherAtl, on 22 February 2015 - 09:33 PM said:

–snip–

ashnwill, on 24 February 2015 - 04:38 AM said:

GopherAtl, on 22 February 2015 - 09:33 PM said:

–snip–

I can't tell whether or not this counts as necroing…

ashnwill, on 24 February 2015 - 04:38 AM said:

GopherAtl, on 22 February 2015 - 09:33 PM said:

–snip–

I can't tell whether or not this counts as necroing…

Lyqyd, on 24 February 2015 - 06:07 PM said:

Anyway, the calc and minesweeper programs look great! Thanks for the contributions. I think I'll add Calc to the main LyqydOS repository soon, and I'm looking forward to your other project being ready for prime time! :)/>

shell.setPath(shell.path()..":/usr/bin")

Lyqyd, on 27 February 2015 - 07:46 PM said:

Packman works fine without LyqydOS installed, to the best of my knowledge. It want to be installed at /usr/bin/packman, so you'd either want to start it with the whole path each time or add the /usr/bin path to your shell; this should do it, if I recall correctly:

shell.setPath(shell.path()..":/usr/bin")

If you can't get it to work, let me know. The best way to report issues with packman is on its github repo (lyqyd/cc-packman, will link later).

GopherAtl, on 27 February 2015 - 09:29 PM said:

quickest and easiest way is to go to the packman github repo, make a github account if you don't already have one, and then use the edit button on the repolist file to add your own repo, like I did. The edit will become a pull request, which Lyqyd will accept and merge at some point. It should point to your own packlist file, either on github or pastebin or anywhere else you can get a raw file from a constant url that you can still edit (you'll need a pastebin account, if you don't have one already, to be able to edit a file without changing the url). By using your own repo, you won't have to wait for lyqyd to pull changes to your own packlist before they'll show up.

> packman install compositor
PackMan
Multiple packages have been found:
	Lyqyd:compositor
	Bedrock:compositor
Please specify which package you would like to install.
> packman install Lyqyd:compositor
PackMan
...

type = none

comp500, on 28 February 2015 - 07:24 AM said:

Namespaces

comp500, on 28 February 2015 - 07:24 AM said:

WePackageItForYou (better name pending)

There are lots of great package managers in CC. The problem is that there are either not enough packages for that package manager, or normal people don't know how to use it.

This idea is here to solve the 'not enough packages' problem for packman.

All you have to do, as a script developer, is ask us to add your program/api/script/thing. We'll do all the rest.

Would these ideas be useful and should I start working on them (WPIFY at least, namespaces seems quite complicated).

comp500, on 28 February 2015 - 07:24 AM said:

also, another idea:

type = none

would be useful for OSes or easy install scripts with dependencies
it's used in apt somewhere - I've forgotten what it's called

local trueW, trueH=process.compositor.target.getSize()

    local trueW, trueH=process.compositor.target.getSize()

lsh:531: attempt to get length of nil
window:84: Too long without yielding

window:77: Too long without yielding

Lyqyd, on 20 August 2016 - 04:52 AM said:

The problem with luaide should be resolved, just run packman fetch update to grab the new version of lsh. I'll look into what's going on with the pocket computers tomorrow.

compositor:26: Too long without yielding

packman fetch update
rm /LyqydOS/.lyqwin
/LyqydOS/installer setup

Lyqyd, on 20 August 2016 - 08:12 PM said:

I looked into the pocket computer issue and it turned out to be a simple case of the default window size set up by the installer being outside the bounds of the pocket computer's screen. If you run these commands on your pocket computer, you should be good to go again:

packman fetch update
rm /LyqydOS/.lyqwin
/LyqydOS/installer setup

The updated installer program will detect that it's being run on a pocket computer and set up default window sizes that make sense for the pocket computer.

You should be able to edit the items in the menu by editing /LyqydOS/.lyqrunhistory and adding/removing items. The number of times an item appears affects its position in the menu.

window:199: vm error: java.lang.NegativeArraySizeExpection
window:84: Too long without yielding

compositor:26: Too long without yielding

Too long without yielding

Lyqyd, on 22 August 2016 - 12:21 AM said:

The settings program requires a certain size, and I designed it without thinking about pocket computers. It's not going to work until I can get around to re-designing it or coming up with another solution for the pocket computers. Whoops.

I'll have to see what the exit program is doing, can't say as I've ever tried running that one!

YogCrafteR, on 25 August 2016 - 11:20 PM said:

packman:265: invalid key next to 'next'

MarcoPolo0306, on 28 August 2016 - 01:40 AM said:

I am getting this following error:

packman:265: invalid key next to 'next'

How do I fix this?

 print("Hello") 

Lyqyd, on 19 November 2016 - 08:47 PM said:

Whoops! I've fixed the installer to bootstrap packman correctly. It should work fine now.