Buglist Sep'17

We are getting too many topics and posts about bugs and issues. To help organize a checklist, I will be making a list of confirmed and unconfirmed bugs. Please feel free to test these yourself, and contribute your own bugs to the list.

Thank you.

A dash - means that grazer has also confirmed the bug, and has been checked.
(These bugs have been tested several times by at least a few members, or by myself, and do exist.)
[spoiler] I will be double checking all the bugs I know of, when I get the chance. They will be placed here, after I have tested them on both Chrome and Firefox. I just installed the newest version of adobe flash player. I will have no other programs or tabs open in the background, to reduce outside lag. I will also be testing these bugs in my debug mode concept, when applicable. If you wish to try it yourself, set every key and action as a number, so when you activate it, the number displays on a label. There is a claim of collision not working, so I can have each collision display a number on screen, so I can prove I’m touching one or more collisions, and if the action did activate, even if nothing actually happens, as a wire could be disconnected or plugged in wrong, somewhere. [/spoiler]
Animations are desynced in old games, but work almost perfectly fine in new games. Even off screen animations still almost always play without starting at the wrong time.
Cutting wires opens pop-up circle in editor.
- Timers are desynced, and do not always start at the same tick as other timers.
The mouse assumed location will be very far off to the left of where the mouse cursor actually is, when editing a game in the Firefox web browser.
Some games that previously worked flawlessly now have massive lag spikes and become entirely unresponsive to controls, eventually freezing your web browser.
- Rotating objects in older games rotate off-centered.
Attached objects do not receive messages.
Emitters sometimes don’t emit the direction correctly.

(These bugs are not confirmed nor unconfirmed, but have been tested by several people in the community, and likely do exist.)

Android apps crashing?

(These bugs have not yet been tested by either grazer nor I.)

All currently known bugs have been tested, proven, or debunked.

(These bugs have been shown to not exist for others, or could not be recreated, and may be a result of the game itself having bad wiring.)

Collision on both ends
Sound looping when game is not playing

If you have any other bugs, comment them below. I will add them to the list.

@grazer, I will pin this post to help others contribute to it easier. Once you have the list copied or checked off, you may unpin and delete the topic. Give it until the end of the month, since this is a September 2017 buglist.

[spoiler] Here’s a build of a simple Bug Tester engine. You can clone it and use it yourself.
Directional Keys
(Use to check if your buttons are being pressed right, if something they do is not working.)
Up = 1 ; Right = 2 ; Left = 4 ; Down = 8
(Shows a collision marker for 1 second. Use to determine if you are actually touching an object.)
Bottom = 1 ; Right = 100 ; Left = 10 ; Up = 1000
X/Y Velocity
(Use to check your movement, jump, or fall speed.)
-Left +Right ; -Up +Down

Bug Tester

Wow, thanks @“Mhx Ar” - this will be super helpful

No problem. I know you’re just as busy, if not much busier than I am, so I’m happy to take away some of the trouble of trying to recreate the bugs, so you can go straight to fixing them. If the main users on here would like to test out some of these bugs, I’ll add a COMMUNITY TESTED list, so they can further confirm the bugs for you. It’s been 4 long years since I joined, the program and community have been coming along nicely.

Tested several times in several browsers, signed in and as a guest, with updated adobe flash.
Flowlab does not work right in Firefox web browser. Whether that’s a Flowlab issue or a Firefox issue, I do not know, but this is the latest version of Firefox and adobe flash player, but Flowlab works perfectly fine in Chrome, and on some flash enabled mobile web browsers for Android.

I’ll add this to CONFIRMED, and it’s up to grazer to decide if this is a browser issue or Flowlab issue, and whether or not he wants to fix this.

I can confirm in multiple browsers that several, if not many games now lag to the point of freezing the browser, and become entirely unresponsive to key inputs. These two of my games are some examples. They worked flawlessly back when I made them, and now are completely unplayable.

Xediga Vade

jngthree’s Starblast 3 gets very close to this stage of lag, but as long as you aren’t shooting, it is playable.

I’m going to assume that I missed an update that limits the amount of on screen objects, or the old “off screen objects get disabled” feature was removed. Either way, the lag is unbearable in these space shooter genre games.

I have tested every bug on the list, and have updated it accordingly. Let me know if there are any other bugs that need testing, and I’ll see if I have any free time after this weekend to test them out.


The collision detector is having problems. Which ever side I put the collision on it has to be on the opposite side as well for it to work.

Using this bug tester, I have determined that the collisions always work, from any and all directions. I’m not sure what your particular bug is, but the collision detector is working fine. I’ll move this bug to the DEBUNKED section, unless you can provide a link to your game, and explain the bug in more detail.

I found a new bug, and have added it to the CONFIRMED list.
It turns out that animations are not the only thing that get desynced.
All of those background grid tiles are the same, all set to subtract or add to the alpha, every 3/10 ticks. If you look at the image, some are a tick faster, and some are a tick slower at fading, which should not even be possible, since they are all the exact same object, using the exact same logic. I have done nothing to trigger this bug. It happens from the start. You don’t have to press anything, and it is always different tiles. This could be why my older games break. This is a brand new fresh game with only 3 objects, the player, the ground, and the background. There should be no lag. I have no other tabs or programs running.


can you make the music load when the game starts, but will load newly added music, instead of loading them when the que is given. Otherwise they don't stop, even when the level is over, but only if the music doesn't start immediately

I have added a subsection (Press Spacebar to switch modes) with a music box and a sound box, both set to loop. I’ve hit them over and over again, trying to get them to keep playing when I open editor, or switch levels, but they always end as they normally should. Unless you can provide more detail, I’ll move this to the DEBUNKED section.

For some strange reason, older games have a bug where rotating objects are no longer centered, and I cannot seem to fix the problem. Watch the Shifter menu (click once on the title screen) vs the spinner in the Bug Tester (Press Spacebar to switch modes).

Shifter v2
Bug Tester

In Shifter v2, the white circle and black are lined up in the editor, but not when spinning, yet spin perfectly within the axis in the Bug Tester. This was a result of an update I missed breaking some of my games.


I think it may be the school wifi, they’re slow with stuff. Keep in in debunked, because it’s only where I do my work.

Open the editor or end the level before the song begins. It just did it somewhere else.

It’s on SB3 if it doesn’t work on the tester

Open the editor or end the level before the song begins? That’s an odd suggestion, but I’ll look into it in maybe half an hour. I’ll also check if it has something to do with your wiring.

The wiring should be fine. All it does is load what the game mode setting is set to, then plays the music for it

Wow, these are some interesting ones - I have been investigating them, and here’s what I found so far:

  1. Timer synchronization. This one is tricky. Every object gets its own unique timer, and it pretty much has to be this way, since every object can control its timer independently. For example, you can have logic that stops or restarts a timer for each object when you click on that object, so if they were synchronized, then stopping one timer would stop them all.

So, this creates a problem if you want them all to be exactly synchronized, since when we start them they get started in series (i.e. you have 160 timers here, so the last one will start running a few ms after the first one.). Over time, the individual timers will probably become even further desynchronized as tiny errors accumulate.

If we have a set of objects that we want to stay synced up, it’s probably a better idea to use a mechanism where they aren’t timed independently. For example, maybe have a single timer that sends an update message to each object. A more efficient example might be to just use a counter and increment each frame - something like this: https://flowlab.io/game/play/718658 - that one shouldn’t get out of sync since the update is based on a counter which always increments by one.

  1. Spinning alignment. This was another tricky one, but I actually think the alignment is working OK. The way the alignment works is that the sprite is first cropped down to the smallest rectangle that it can fit into, and the rotation is based on that shape.

I downloaded the spinner sprite from ShifterV2 here: http://flowlab.io/assets/users/16/user_16894/game_488603/img_asset_2104711.png?stamp=1505515556554 and opened it up in the Gimp image editor. Although the image looks fine to the naked eye, there is a thin row of almost completely transparent pixels on the right, left, and top edge of the image, but not on the bottom. This means that when the image is cropped, extra area is left all around the image except for the bottom, making the sprite kind of “lopsided”, so that it spins off-axis. If I adjust the alpha channel threshold for the image, the nearly invisible pixels become easier to see:


Another way to test this is to open the sprite in the sprite editor in Flowlab and just draw a single pixel on the bottom edge. This will prevent the bottom of the sprite from getting cropped, and it should spin correctly. Alternatively, you could enable collisions & set the collision shape to “rectangle”, then when you turn on debug view you can see that the rectangle isn’t where you would expect.

The reason that the rotation alignment works properly in your Bug Tester is that the sprite is cropped using the main sprite dimensions (a big “X”), and the animation frames just continue to use those. dimensions.

Next up, I’ll investigate Song Loading and Lag.

Lots of typing. Added spoiler to conserve scrolling time.

[spoiler] Interesting finds. I was aware of those lines on the edges, and was planning on erasing them in gimp, but I didn’t think of the crop offsetting the rotation. Even though they are both the same, only opposite colors, they were cropped differently and ended up in different squares. Using the X as a base shape and having the animation not be, would guarantee it rotates correctly, in theory. I’ve used a base shape in the past for a hitbox, and the animation automatically started, hiding the hitbox. I can’t believe I didn’t think about that.

As for the timers, wouldn’t sending a message to 160 objects also eventually lag? I’ll look at your example when I have a chance. I’ll be busy for a few days.

I know there are a few more bugs I personally have that I’ll look into. One of them is proximity lag. I first noticed this when making Vortex. Vortex doesn’t work anymore, but I ran into the problem again in Shifter v3, making a homing attack, but the x and y position numbers absolutely refused to update to the next enemy, after the old one was destroyed, for the longest amount of time. Even if the object gets destroyed, it still keeps sending messages with coordinates, so the alive enemy and the dead enemy mix up their numbers and you teleport to odd locations. In Vortex, the black hole would continue to suck you in towards it, long after it was destroyed. I’ll try to find a workaround, but it has been a pain in my side for years.

Anyway, I’m glad I could help you track down bugs. Before flowlab, I used to be a bug tester for beta games, and I’d report bad bugs, while keeping fun glitches a secret. I’ve glitched and hacked games most of my life, so I’ll be happy to keep finding and testing new bugs with each update, if I have time. [/spoiler]

@jngthree - I cannot seem to get any music to play in SB3. It looks like everything is correct, but for whatever reason the mp3 files don’t load into the game for me. It appears to be something to do with the server that’s hosting them, because if I upload the file somewhere else (like http://quadbuffer.com/DeepSpace.mp3 for example) then the music loads and plays fine. Am I the only one seeing this issue?

If I do change the initial title link to http://quadbuffer.com/DeepSpace.mp3 then I can hear music play, but I can’t reproduce the problem the way it’s described. Maybe I’m misunderstanding the issue, but here is what I am trying:

Start the game, then open the editor before the music starts playing. At this point, I’m expecting the music to start misbehaving, but I don;t hear anything. Maybe I am not testing correctly?

@grazer, I’m not sure either. @jngthree, I’ve tried dozens of times, on several games, and Starblast 3. The fastest possible way to back out of a game is with the esc key. I’ve spammed it over and over, and even tried to exit with the pause button. The music always ends. I am not getting any looping before or after the music plays. It is a problem with your cache? Sometimes clearing the cache on your browser helps remove old data. It also may be your connection or music host. I cannot seem to get the song to play outside the normal gameplay. Maybe it only happens on your computer, or your account.

It takes a bit to load. Eventually it does turn on. But once it does it doesn’t turn off

Maybe it’s the host. I’ll have to move the music to kiwi6 or something.

Ok, for the moment I’m going to skip over the sound loading, until we can figure out a solid way to reproduce it. Next I’ll check out the lag in Vortex, Xediga Vade, and SB3 and see what the bottlenecks are there.


I haven’t had any lag issues with SB3, because the main reason SB2 was lagging was because all the proximity behaviors I used, and the scrolling system.

This time I used a messaging system that got the job done and scrolling is no longer present.