I should also mention that I want to redo the website at some point. It's built on a platform that was fast and easy to get up in time for launch, but is now charging me money that doesn't need to be spent. I'll rebuild the website into a cleaner/ more professional looking 'scroller' website on wordpress when I get the chance. I want to wrap up the proposed core upgrade before I work on that though so you guys have a chance to go over it while I do the site.

Basically when I get happy with the core upgrade proposal I'll post it. You guys can look it over, see the exact changes it would make, and decide if you want to go through with an upgrade or make further changes to the proposal. We don't HAVE to upgrade if people are largely content, but I want the option to be there for when / if we come to the agreement to rework the reward code. We could just put the upgrade on standby and focus on things like marketing and increasing our userbase. If / when we do decide the proposal is good and largely agreed upon, we will then go over what the upgrade process would look like and proceed from there.

I think explaining the room for improvement and putting out proposals like this is a good way to get people thinking about the core code and potentially stepping up as a volunteer dev willing to keep improving Swing as a community.

Just for clarity sake I haven't had any issues myself either. I went out of sync once weeks ago when i was offline for a while but aside from that no problems.

I'm just going off what people have reported as opportunities to strengthen the core code.

Let me give some insight.

I've recently been digging into what happens with the rewards when they are calculated, and exactly what's going on when getting calculated.

With the current setup, that process looks something like this:


HERE is the section of code that spits this out. I just used block 1 - 999 as sample output. For the sake of clarity i re-coded the process in java for faster debugging. (im just better at java is all, but java may produce slightly different results is the point)

So what I discovered is that somewhere else in the code the decimal is inserted into a full number with 9+ digits.
The end result in this section of code that gets passed out looks something like '9.87 * COIN' (=987000000).

Me attempting to put in the decimal during the reward calculation instead of letting it do it after results in some unnecessary math. Math that doesn't always turn out clean, and leaves room to lose accuracy when clipping/rounding the final result to a full number.

I'm working on re-coding this section so a result like this:

Count: 987 Three: 987 Decimal: 9.870000000000001 Final:987000000.00000011920928955078125

Acts like this:

Count: 987 Last Three: 987 Final:987000000

Now, this is an example of POS. POW gets a little more challenging since rewards need to be multiplied by 10%.. 20%.. 30%.. etc.. But I'm working on doing that cleanly too.

But that's essentially where we have room for core upgrades and cleaning up the internal math for increased accuracy.

So that should give you guys a little better idea at what I've been looking at.

Will post more later as I make progress.

Me contacting them doesn't resolve this. We just need to make trades or encourage new traders to automatically resolve this. More volume = warning vanishes.

(I'll give it a shot though.)

Hey guys, just wanted to touch base.

I'm at the point where I have a good feel for exactly where and how the Swing core can be improved. I'm on a buisness trip this weekend but when I get back I'm going to work on pulling together/finalizing and posting the proposed upgrade.

I plan to explain the issues currently occurring in the code with example debug logs, along with the difference the upgrade would make. That way everyone can easily understand the issues and see what it is were talking about changing.

Thanks guys. More info to come later.

x64 or x32?

Alright so we've dug into the code to take a look for potential connectivity upgrades and have made progress.

I'd like to try to force some connectivity problems in a test environment, that way I have something I can use to test potential upgrades against.

In order to do that I need to get a better feel for the environments any forking occurs in.

Put simply, if you think you have experienced forking, i NEED to know what operating system it happened in.

I'll check back to see if anyone responds. If not I'll start prodding people from earlier in the thread.

Thanks.

I'm quoting this just so it's visible on the new page.

Thanks for voting guys, keep it up!

Quick status report.

I've been taking some time to research into the best ways to go about this upcoming core upgrade. I've spoken with a few other devs about the issues we are looking to upgrade and gotten some valuable input and suggestions on alternate solutions. I'm meeting in person this week with an intelligent coder friend of mine who is familiar with the issues we're facing. We plan to take a deeper look at the code itself and potentially come up with the proposed patch. Along with this I'm factoring in some minor core upgrades to be packaged into the update (checkpoints, seed node update, protocol versioning, etc.. etc..).

In addition I've been looking into getting more devs on the swing team. The other coin devs I've spoken with have expressed an interest in helping but had to keep focus on their own projects. One of them has expressed an interest in some cross marketing down the road. The dev I'm meeting with is busy chasing his phd, but has expressed a willingness to work on the code with me and get involved to an extent.

Trying to make sure this is done right rather than rushing a quick patch.

Will keep you all updated.

We will keep progressing as a community however we can. I'm focused more on upgrading the core right now, but I'm open to any ideas.

Glad to hear you are "well in profit" though. Shouldn't be too bad for you if it takes us some time to get our footing.
There are no plans to abandon ship (Even with how much core updates stress me the hell out).

I'd make some volume for us if I could but I'm down to .005 BTC after buying that chart on TrySwing.com. Cost me something like 0.95, I didn't get as much support paying for it as i had hoped. So I can't do much volume generation for you guys. That's up to all of you.

On a side note, I'm thinking about adding more seed nodes. If you want your nodes IP to be hardcoded (or something similar) tell me. Would be good to get a couple pools doing this.

Good to hear. Sorry for the lack of posts been on a thanksgiving trip. It's probably worth mentioning the seed node runs on Ubuntu 14.04 x64. I would expect wallets running on the same operating system likely won't have forking troubles.

While I've been gone I've been touching base with other devs to get their opinions on the state of swing and update options. Will try to keep you guys in the loop. Trekking back to the home office now.

Hope everyone who does thanksgiving had a nice one.

Thank you. This should help reassure users concerned about their personal wallet status while I investigate.

For GUI users:

(At least in windows)

You will not have the "Target" field in the Swing.exe properties.

To rescan:

Right click and drag/drop Swing.exe -> Make shortcut here
Right click shortcut -> properties
In the "Target" field you will see the path to Swing.exe.
make it say: C:\Directory\Swing.exe -rescan
Apply -> OK
Launch VIA shortcut.

Now when you startup Swing you have a regular launcher and a launcher that rescans. (It will happen during the splash screen and will take a while. Mine hasn't finished yet so I'm not sure what to expect when it finishes.)

Also I should mention that I'm not firmly against it, I just think it's a bad idea. If there was big community support for this we could investigate.

My thinking is just:

A: SHA256 makes it more likely for miners to discover and sample Swing.
B: SHA256 is easiest to rent and mine with for most people.
C: I don't have any experience with an algo swap and the more we change the more we introduce room for error. I don't want to break things.

I like the thought of this. Can I get some links?

My initial thought is about how Swing already seems to split and combine inputs on its own. I'm unaware if this is compatible with coin control. (I also don't know why or where in the code this occurs.. other than guesses that is. I haven't explored that yet.)

I also don't know how they accomplished coin control or how to implement it.

Being the only dev working on this (we could really use more) it may be tough for me to implement this without being able to bounce code off people.

That being said, I'm absolutely interested in exploring it.

While I'm digging around trying to sort out the connectivity upgrade I'll see if I can figure out how to make the new version's wallet go fullscreen too.

I should also mention:

@Berzeck in my slack group (Part of the WorldCoin team) has offered to potentially help us implement his "Dancing Hippo / Worldcoin Business Center" wallet as an alternative wallet for Swing.

I gave it a look and it seems pretty neat. He's also the only dev to suggest or seem interested in working with Swing. If people are interested I can poke him about it.

(Oh and it goes fullscreen)

Preview of what that looks like:

Let me be a little clearer if you don't mind.


How did you notice this? A Swingd command? Comparing your wallet to the block explorer? Comparing two wallets?


"I did a rescan"

How? Is this a Swingd command? How did you get the readout you posted above?

Excellent to hear!

I'm unable to replicate any issues on my end either, but that doesn't mean they don't exist. I will be investigating.

This is the last we spoke of it. Do you get this issue very frequently? It's not frequent for me and nobody has told me it's severe so it's been categorized as minor up to this point.


That being said, I'm upgrading this issue from minor because I think it's related to:


So that being said, I'm now upgrading the issue from minor and investigating a core update.

I believe I know what needs to be done to fix both of these problems. I'm currently looking into verifying and figuring out what the best way to push an update is.

If anyone wants to discuss other core upgrades, now is the time! (Keeping in mind we are currently looking at core upgrades, not wallet upgrades.)

Crackfoo, if you can PM me or post here on information regarding how you went about investigating this I'd appreciate it. Others might find it useful as well.

I'll need some time to confirm things and prep a suggested update. Please bear with me through this process. I'll keep you all updated best I can. We'll get it all sorted out.

I did a test withdraw and it went through. It didn't hit quickly so I re-downloaded the chain and got it. I may have just been impatient I'm not sure.

To be clear for debugging/clarity, did your wallet fall out of sync, or continue to progress on a side chain out of sync from the seed node/bittrex?

Also, are your coins where you want them now?

I submitted a ticket as well to ask about it.

Could be their wallet just fell out of sync and payments are queued but not being processed. Whatever it is we will get it sorted when we hear from Bittrex.

I would caution people to do test deposits/withdraws and wait for them to clear before making major transactions.

Thanks for the heads up crackfoo, sorry for the inconvenience.

I've had deposits take much longer than they should in the past in a few coins including Swing. I think it's something to do with their deposit platform. Bittrex isn't showing any warnings about the coin being offline so I'd just give it a little more time. If someone wants to test and report back it can only help.

Excellent, thanks!