Official Choosing between Collaborate’s 1.2.16 and 2.0-Preview-15 Packages

It's maybe a bit OT, still, may I ask, what is this "Unity Accelerator" magic?

BTW, on a 80GB project the background file scanning (I love not being stuck watching the previous loading bar) is currently taking 1/4 of the time it took in the past. Running in background.
And, up/down speeds seem pretty fast compared to 1.x

I still have some issues since I switched to the new preview version in mid-project, so it's asking me to push everything to the server once again. Failing while doing so (I will now check the whole forum, I have been unaware about 2.x until yesterday). But, considering I had issues during the pre-collab era, during the collab 1.x era, I'm not complaining to have a few in the 2.x preview era

I'm really optimist and have good faith in this 2.x - when it's going to be available - I guess in March or June 2020, or sooner in the alpha/beta

 

Personally I'm running out of good faith. I wanted to release this under 2018. Unity has forced me to migrate my game to new versions twice now and both times I had to fix a bunch of assets and replace some completely because they were no longer supported.

AND as if this wasn't enough work, 6 months later when everyone updated their assets you have to replace the ones you temporally patched with their real "updated" versions and make sure everything works... of course it never does.

I feel like trying Unreal or Lumberyard for our next title.

 

If you use Unreal you will have to use Git, Plastic or Perforce anyway. Just switch to Plastic now, its significantly better - just set it up with the correct .ignore settings and don't rush it.

 

https://blogs.unity3d.com/2019/09/11/speed-up-your-team-with-the-unity-accelerator/

 

Is it possible to switch a medium sized (20 GB) project to plastic from collab?

 

We switched a 40GB project, but we are having some problems now which I suspect are due to rushing into it without proper ignore files (we were desperate to get something working, since Collaborate had corrupted the cloud-saved files).

Just get your ignore files completely and comprehensively setup and you should be fine. Also, you will want fast connections for all users, since there is no compression. I'd suggest 100/100 minimum for everyone.

 

Have you used the git export feature from collab or have you just started a new repo from scratch? We've been having a lot of issues with exporting from collab lately and we eventually gave up (wasted more time on it than we could afford at this point)

 

We started a new repo based on the up to date project that one of the developers had locally. My team (including me) is only two people, developing the final set of content for Frontline Zed.

We're having another problem today though where something is working on another PC locally, but not any of the remote versions. It could just be because we are on 2017.4 and are stretching the project beyond its initial scope due to the positive reception and good sales it received on release.

Remote work is very problematic and I plan to de-remote the team next year if possible, then just run some kind of local version control system, maybe use Git etc. and save a little cash on Plastic version control fees if we can run a local office NAS with a manual automated cloud backup.

If you aren't using Collaborate now, what are you using? How many people in your team?

 

Still on Collaborate. We are 4 people out of which only 3 use the repo actively. We all work remotely as well so hopefully if we switch that won't be too big of a problem (network speed is not an issue - we all have 1gbit fiber connections)

We really have to switch away from Collab at some point though. As we are getting close to the release as early access we would really benefit from having different development branches, especially since we plan to upgrade the whole project to HDRP by the end of early access (if viable).

So I guess for big projects (ours is currently 45GB but we are still adding audio and props so we expect it to end somewhere at 60-70GB) with remote work plastic would be the best option...

 

Dear Unity, collab decided to waste my time again so you get another complaint.

What is preventing you from adding a cancel button to collab 1.x.x? I would be 400% happier with it. AT LEAST with all its problems it wouldn't be wasting my time so often.

FYI I know about V2 and I don't want to use it. I keep reading posts about missing files so let's forget it even exists.

Why not have one of the programmer's skip his 420 break and implement it already? I mean... look, took me 2 seconds:


But seriously, I program all day long. Adding a cancel button is NOT that hard. Telling everyone to just wait for 2020.1 without mitigating the current situation is NOT a̶c̶c̶e̶p̶t̶a̶b̶l̶e̶ cool. If you want to treat your ̶f̶r̶e̶e̶ future customers like they don't matter, fine but I would never dare to say this to any paying customers.

The reaction I would have expected from any company would have been to bring collab to a usable state and then work on it's replacement but NO instead you embark on a never ending adventure to redesign the whole thing from scratch.

I am NOT happy with you guys and the more I read the more I realize how broad the problem is; your new multiplayer code is stuck in limbo but you still tell every one to start using it, the constant rendering pipeline changes are making a mess of things for asset developers (which end up deprecating their assets which we have paid for, thank you!), you are releasing a visual coding system, there's also the shader development tool you published not long ago and many many other things in progress. Oh yes there is also the overhaul to the built-in animation tools and the Post processing stack V2's that has constant issues with every new version of unity... What a pain.

How many of these systems do you people think will be production ready when they are released?

This is what happens when your development team is spread out on too many projects at once.

It creates the kind of bad customer experience that makes you ditch a game engine the moment you have enough money to get serious and quite frankly, now that I have to think about how we will handle working with a larger team, I am starting to paint a bigger picture of the whole situation and I don't like what I am seeing.

Modernizing the engine across the board is an AWESOME goal but by paving our road to success with solutions like "please wait" to critical issues while constantly "pushing the release date" of said fixes you will lose the customer base you are trying to build.

...

Anyways... I am not asking for the moon, I just want ̶B̶A̶S̶S̶!̶

EDIT: I was just thinking, while collaborate was wasting my time AGAIN, with solutions like Bolt and Amplify shader which are both amazing, why did you even invest resources into developing your own visual coding solutions? There are so many places you could have invested the time instead.

These two assets are complex to develop and you are going to kill them with something that will most likely, at least initially, be a poor replacement. Your entire reputation has been built by a community of devs who weren't afraid to pitch in and code add-ons to supplement the parts of Unity that were lacking and you keep f***ing with them. Why would ANY developer invest time on producing quality assets when they see this now? You should do more of what you did with the dev of Text Mesh Pro... buy the add-on and hire the guy to properly modify it to your taste.

 

I agree, take the engineering resources off Collaborate and keep it as a basic backup plugin for solo developers.

Plastic SCM *already* does everything that a future Collaborate version will do, and its an industry standard piece of software.

Just partner with Plastic, and work with them to improve the documentation. Make it the official recommended version control system.

 

I agree with this but I must admit I really like the things they are adding to the UI for comparing diff in prefabs, the hierarchy and in scene. I always thought Plastic-SCM and Perforce could use a better integration.

 

Can't agree more. Its unbelievable how unity is ruining its own reputation releasing stuff as "production ready" while its not. Its occurring too much times. Cant understand how its possible unity management is not yet aware about this.

 

this /thread /thisforum

 

I guess the question is whether you're willing to trade a few nice features inside the Editor for losing man-weeks of productivity to general faults and limitations in the Collaborate system, usually at the most critical of moments (before a release or patch). That has been our experience with Collaborate.

 

SURE go ahead, try to logon for not god damn reason and if you fail try again until it works and then RESCAN EVERYTHING!

You have to understand that scans take a good 20mins to complete...

 

@tcz8 Im pretty sure that you'll need to be VERY patient. @ryanc-unity stated that this bug will be fixed AFTER the release of Unity 2020.1 (April 2020).
https://forum.unity.com/threads/col...ce-has-not-changed.468150/page-2#post-5215895

 

Wouldn't recomend using collab for anything bigger than lets say 5gb without an ssd to be honest (from our experience here - we had to buy nvme ssds for our workstations just for unity projects of 40gb+ and collab)

With big projects even regular ssds can be a bottleneck if you don't like waiting and looking at a progress bar. I see collab going over 2GB/s in read speeds on our drives.

 

I just lost 2 days of progress thanks to the nice custom merge things collab does and like 3hrs of "checking for changes" today so I'm here to take a dump on collab while I wait for it to finish, wishing I switched to Plastic already but I'm too busy doing other stuff...

 

My turn to give up Collab.
I don't understand how slow things are going, for users it is like testing a beta program far from production-ready, but paying for it...

After 8 months using Collab 1.2 then 2.0, my team just switched to AWS CodeCommit in a day:

• 5 users
• 50Go storage
• SourceTree compatible
• Instant changes awareness
  
• Instant pushes and pulls
  
• FREE! Lifetime.

No more problems.
Maybe Collab will be worth it for Unity 2021 according to the current plans and the fact that not a single deadline has been respected...
I would not recommand it until then.

 

How big is your project?

We're currently considering multiple options for version control hosting but our project is rather big and with some bigger binary files.

CodeCommit seems awesome, but would it be suitable for larger projects?

 

CodeCommit allows up to 2Go single binary file.
Our Assets folder weights 8.2Go

 

Interesting!

 

So... I am working on a Cinemachine Bug and had to upgrade to the new Unity version to test their patch... While waiting for Collab to do its "thing" I came up with this:

 

Attempt number 478 of trying to get collab to work. Collab 2 is in a much worse state than Collab 1 was, and I have no idea why they released it. I've encountered a non-stop title wave of different bugs. I fix one, only to have another pop-up. Right now it's trying to get me to get changes that I've just pushed to the server LMAO. The worst offense is that they actually didn't even get rid of the forced "searching for changes" on opening unity, which should have been priority #1.

For all reading this, stay away. I would not even recommend it for solo devs. Its an absolute mess.

 

I'm surprised people bother so much with collab. If you have any git experience I highly recommend https://desktop.github.com. I only use collab right now because my team doesn't have any real programmers on it, and they use macs so they don't have window collab issues but I've worked on a project with an experienced programmer who loved GitHub Desktop (he was able to just use git, and I was able to use desktop without any real problems). Git is way faster, clearer, and I can even compare images with their image diff support. No loading issues, proper file updates, proper branching, and easy enough I can use it without too much trouble. And yes, the unity desktop plug-in is trash, that's why I recommend just using a regular GitHub tool instead. Especially if you are looking for something free to use.

 

So does this mean that codecommit is affected by the same 2GB github issue? I mean, moving a big project from collab to github is a nightmare cause a single commit cant be larger than 2GB (dont know if this issue affects plastic too).
Is there a quick way to import a collab project to github/plastic?

 

FYI that problem doesn't exist on Plastic. Our initial commit (which was over 6gb) was really quick. I remember being surprised how fast it took to upload compared to the last time I tried to resubmit the whole project to collab (which ended up in the editor crashing). Collab is a ford focus, Plastic is a Tesla Roadster.

 

Recently I migrated from GitLab to GitHub with a project that is over 12GB and I was able to do a single push, not sure where that 2GB limit came from (didn't found anything in their documentation neither)

 

Are you sure? This is a git limitation, so any version system based on git has this limitation.

https://github.community/t/working-with-large-files-and-repositories/10203
you cant even push a filer larger than 100mb (you need to use an extension called Git LFS to avoid this limitation).

 

There is a 2gb limit per-file, not per push (info taken from the link you posted, I didn't know that before, but neither I see a reason to have a filer bigger than 2gb in a repository). Also, I am using Git LFS on that project, I was already aware of the 100mb limit per file.

EDIT: re-reading it again there seems to actually have a 2gb size limit per push, but probably it doesn't take into account the LFS files, which could be reason why I didn't have any issue with that.

 

Interesting how prescient your post was, from 2019.