Search Unity

With a new release candidate version of 2019.3, we’re entering the final stages of testing for release readiness.

When we label a Unity version a “release candidate,” it means that we’re finalizing the last step of the beta testing phase and are in the final stage just before the full official release. Each release candidate goes through a further set of Release Acceptance tests before we make a decision whether it’s ready for publishing. 

Unity 2019.3 is our largest and most complex release to date, with new foundational tech across the board and more than 170 features and improvements on top of a ton of fixes. We want to ensure all these tools work well together before the release leaves the beta testing phase.

Based on the feedback we have received and expect to receive during the end of the year, we aim to hit the quality bar needed to enter the official full release status in January 2020.

If you can’t wait to get your hands on it, the latest release candidate is available to download today. Unity 2019.3 has a ton of quality of life improvements, and lots of new big features (like HDRP and our new 2D tools) are becoming production-ready with this release. Feel free to take the latest version for a spin over the holidays.

The Unity 2019.3 public beta has been our most active yet, with around three thousand bug reports from the community. We take this feedback very seriously and really appreciate the contributions you have made throughout the 2019 TECH stream. Please reach out via bug reports or in the 2019.3 beta forum if you are encountering any unknown issues with this latest release so we can address them.

For those of you who have projects in production or want to update live projects, we highly recommend waiting for the 2019 Long-Term Support (LTS) release. While the TECH stream offers you the latest features and improvements, in the LTS releases we focus entirely on stability and quality. Therefore, the LTS version of Unity 2019 will only add fixes that address crashes, regressions, and issues that affect the wider community. It is our recommended version for those projects that are past the pre-production phase. We expect to ship the 2019 LTS in the spring of 2020 along with the first 2020.1 TECH Stream.

74 replies on “2019.3 is now in the final stages of beta testing”

After reading most of the comments here, I can honestly say I understand everyone’s frustration. But the reality is, this is how software development goes, especially for game dev. If you want the latest tech, prepare for prototypes. Unity states, if you want stable, go with LTS. They don’t recommend going with current year releases if you want production ready soon. However, that being said, I never used a ‘stable’ release of Unity. I’ve been with it since 2010. There has always been one issue or another, mainly with 3rd party assets not playing nice with the engine itself. The more assets you depend on, the less control you have. I always stuck with Unity though, because the community was strong, and it was easy to share solutions.

My take with 2019.3 and going forward, it is very promising. It is very early. It depends on your platform, but most issues can be resolved. I am dev that has done a variety of software, Unity, WPF, Winforms, and others. Believe me, it’s not just Unity that has become more ‘complex’ with packages. The entire industry is going in this direction. The key is to plan ahead, lock in what you need, keep a backup plan or two, and adapt. Never settle. I am currently using 2019.1 for production, and we haven’t had many issue that could not be solved from a bit of digging in forums, blogs, and Gits. Honestly, the propriety ‘solutions’ are the ones that are the biggest headaches I always find, not the engine itself. I’m sorry for the community for diving head first into all the new features. I did a few of them, and I get it, it changes with every minor release, but keep in mind, Unity made no secret that was how it was going to be.

I don’t care about the CEO so much as I care about Unity. There are always a few bad apples in any company, and hopefully Unity is looking for a replacement. But there are also hard working devs at Unity, really pushing the engine forward so give them a break. Considering how many things they have to build, and integrate with, everything moving in different directions at the same time, I don’t blame them for having so many dependency issues. I can see alight at the end of this tunnel though. Things are finally starting to come together. Major features are starting to stabilize. Major features are starting to integrate really well. I don’t see a need for any more new features other the RTX support, everything else they have is beyond what I expected from Unity. I just hope they focus on usability now. Narrow down the versions, and keep improving the documents.

Unity is not listening to the user base as the company nowadays is focussed on maximising profit, not democratising game development. The past two or three years have been an absolute nightmare in terms of stability as Unity prioritises doing an IPO as soon as possible – the very reason why the new CEO was hired. The mad scramble to become “relevant” and a real competitor to Unreal has eroded confidence in Unity as a company considerably. There has simply been to many mistakes, too much chaos, and too much calculation that the user base somehow will put up with the troubles consciously inflicted upon it by Unity management.

I was thinking that Unity somehow had fought through all the troubles, learnt its lessons and had come out on the other side. Unfortunately, that’s absolutely not the case.

I’ve used Unity since 2010 as a Pro user. I’ve switched to Plus, which is considerably more expensive. I’m now seriously thinking of ditching Unity altogether in favour of something more stable. Sure, it will involve learning another engine. But as developers we ditch technologies all the time, for all possible kinds of reasons.

To my mind, Unity has missed the train. And I have no confidence in Joachim Ante.

The two main things Unity needs asap (again) are stability and reliability.
The comments show these two are the main things lacking heavily right now (and for over a year if not two, but getting worse and worse).

I could list many examples, but just as simple one: in Unity 2019.2.x “low overhead mode” for mobile VR/ Quest is mentioned. So despite all the deserved upgrade worries, i upgrade my project to it (from 2019.1.2F1), of course then also have to upgrade Oculus integration and a bunch of other plugins and do various changes in own code.
Then after 1 full day of work finally the project runs again…and… it runs noticeably worse on device.
The performance got worse overall so much, that even with low overhead mode it does not achieve the same fps i was getting out of the box when deployed from 2019.1.2F1.

Like come on, W T F.

I expect ideally better, if not at least the same performance for a new version, but no, worse. again.

I am really fed up of upgrading to newer Unity versions and things running worse.

And no, huge performance gains advertised for when doing all in dots does not count, at all. If you expect your whole userbase to dump their entire code base and just simply code everything in different way, and btw that different way also does not work with half of the systems in Unity like physics etc, then no, this is completely misguided.
People expect better performance for when the same project runs in newer version. out of the box.
That’s why we pay the subscriptions, so things become better. Else i can stick with my several years older thing.
If you have to internally rewrite graphics in dots or similar under the hood to make that happen, ok, but it should be reconsidered at least twice whether a good idea to try to make your whole user base have to switch to a different coding paradigm in their code because unity performance itself when doing things the “old” way is not improving and even regularly regressing with dot versions.

And that the performance/fps got noticeably worse for me with my project in newer Unity dot version, that’s just a general example for too often updates bring regressions on many ends.
I can’t even switch to even newer dot version because that has even way more known big bugs.
And now i see 2019.3 release candidate builds with known issues like crash bugs and serious physics and other performance regressions.
This is not acceptable to even put a release candidate label on something with that type of issues and i am sure all developers agree this is ridiculous and puts huge harm to Unity’s reputation as a trustable partner that this partner would even consider to put labels like release candidate and production ready onto things in such a state.

If someone from marketing or higher up forces you to this, they should know: it is achieveing the exact opposite of what you want, you are extremely risking loosing customers and tarnishing Unity’s reputation in irreperable way, not increasing value short or long term.

Then there is the even way worse side that working with Unity meanwhile really feels way too much like having to search together many patchwork pieces where different versions of different pieces work to different reliability and functionality level together with other pieces.

This is not my job, don’t attempt to make users have to have this job on top of their actual job. It will drive everyone away over time to have to constantly deal with this unreliable crazy messy patchwork puzzle.

The theoretical benefits of turning all parts into separate packages Thomas mentioned are nice in theory, but the actual reality is that it has been a total gigantic mess and will of course stay a total mess, because if you think about it, it has to.
Of course when all the stuff is made up of many different parts which are independently developed further, updating and breaking on different ends at different speeds, it becomes a growingly impossible puzzle to have to check for every piece with which version of that other piece it will actually properly work together.
If you think it is acceptable to let users have to juggle and experiment with combinations of versions of all those packages each dot version, please think again.

It is bad. It drives away more and more people over time as people get fed up of it more and more. Again, reread the comments and forums. This is the voice of the majority meanwhile and only growing louder.

Attempts to counteract that loud public voice with ever more hubris and unfitting (untrue) marketing messages just then on top makes you come off even less trustworthy.

Please list me for the current Editor version with which packages in which version all will work ideally for platform x,y and z deploy.
It is impossible. (Both because a) such a combination does not exist, all are buggy in different version combinations with each other and b) even if it would exist, it would be impossible to figure out the ideal combination halfway quickly)

Sorry, i have to say it bluntly at this point: this is a wrong direction and it has been proven to have failed, for over a year now, and it has gotten worse and worse instead of better.
And it makes me seriously worried that unity is still hell bent on this being a good way to do things and hoping for this and that dependency control attempt to solve this.

No, you can’t, because it is not solvable, when different things release at different times with different version dependencies, there will always be dependency and bug issues and the more separate packages you have, the more unreliable any combination of them will become.
Please reconsider this over push for packages because it completely destroys any stability and reliability one could and should be able to expect from a game engine.
Anything that is a separate package should be made part of the engine base again asap so that one finally hopefully would get a chance again to have a unified engine one can download and knows all parts work together with each other properly, as, you know, one should actually be able to expect and completely absolutely can not right now.

Sorry for my for me unusually blunt tone, but i am really annoyed right now for having yet again, for the xth time trusted in some fancy marketing sentence of version x upgrade bringing this and that improvement and then i waste a whole day on it and the end result is things run worse now in newer version and i have to either revert all the work or try to see what kind of bugs and regressions in this new version cause this performance regression so i can try to work around those again if possible.

This should not happen and it happens constantly, with every dot version and every individually upgraded package now.

Sorry to have to send you with bad vibes to the weekend, but some things really have to be changed, massively, else no good next year.

Also one note on the scriptable render pipelines: They are at best in what i would call alpha preview state and it is not ok you are willing to at most assure us the builtin RP will not be deprecated for a few years, while really you should assure us it will be developed further in best way with ideal support for all current and future platforms for years while the other stuff is no matter what marketing decides really in alpha state and not even halfway close to being a good generally usable replacement for the builtin RP and trying to force everyone onto the new stuff asap just makes more people consider switching away from Unity, not switching to the new stuff when once one tries it it is clear it is not production ready for most use cases yet.

In general it is also a huge mistake to try to do two new RPs, one for lower end devices and one for highest end ones, because one of the main reasons lots of people use an engine like Unity is so they can deploy to all platforms they want without having to do decisions like hey, what platforms and devices i want to deploy onto when, based on that which RP should i use and then based on that have to make custom shaders and post processing stuff all over from scratch again, replace all asset store packages i bought and also replace half of the unity systems like physics, ui and other things again with something else.
It should be one RP which scales properly by itself, to all platforms from lowest to highest requiring only one way to make things and also allow using existing skills, assets and code bases.

At that point when you force people over and over on many ends to do things from scratch in yet another 12th new way of doing things, yeah, again, you’re each time inviting people to reconsider whether to stick with Unity at all.
How many UI, input, physics system, render pipelines (and with that all shaders, post processing etc) versions one has to re do and yet it does not become more generally usable, more stable and better performing (unless one on top also drops out all code and does everything in different coding way), you’re spreading yourself and the whole user base way too thin with two many options which all don’t work properly with the existing stuff, among each other and on as many platforms.

Unity used to do what a good engine does, which is support their users and allow them to work as the users wanted and hence unity would bend and stretch to meet the users’ expectations.

Now it feels way too much like the users have to constantly bend and reconsider to adapt to Unity’s ever changing constant direction twists, which are gowingly becoming less reliable and more confusing.
Like hey, i want to use the latest best physics setup, what should i use?
Hey, here’s 3 physics engines, enjoy sitting through 1-2 days of lectures to be able to decide which one you maybe should use for your next project.

This all is halfway ok for someone making for fun stuff in his spare time or a fancy showcase example project where it seems to not matter it breaks with the next dot version of the editor or the next version upgrade of package x, a complete no go for a serious production workflow.
Does it not make you wonder at least when already for your own sample and showcase projects most of them don’t work out of the box with the next dot version of the editor or some package update?

So yes, the users have spoken, i hope Unity listens and really reconsiders things, because it is really getting close to being completely unacceptable for more and more users.

I should also add that the load times when switching levels are around 4 times as long when running deployed from 2019.2.15F1 and also 2019.2.17 compared to when i was deploying from 2019.1.2F1.
4 times!
Please debug your stuff before you release it as “production ready”!

Pete, THANK YOU! As miserable as using Unity has made me, it hasn’t broken my resolve to finish. In those most fickle moments, when it’s supremely difficult, lonesome, painful and agonising to push on, your writings will help endure, push forth the stoic spirit and a little relief, cold comfort for all the time lost. Thank you!

This is true and even unfair to expect it but one person can make a huge difference in a companies direction, we’ve seen it many time. End of the day progress is never free and as a long time Unity user will be riding this one out

I think packages is the right direction, it’s really nice to get the source code and quick updates, the real problem is finding stability between packages and QC with the rapid development they provide. This is a serious issues that needs a solution ASAP before Unity loses more users. In production it so so tight, every bloody hour is so crucial to stay on budget and make deadlines. We simply can’t afford to debug engine issues and to make matters worse legally we can not use assets from the asset store. I read comments of people relying so heavily on the assets store during this transition but we don’t have that luxury and we don’t have time to develop feature that should be working. Even though we had success with unity I was out voted to make a switch for the next product simply because of the extra overhead instability would cost us. It crazy that it safer to retrain and refactor to a different engine !! I love working in Unity and when it works nothing is more rewarding, so Im holding on to that hope. Just speak to us Unity we need to navigate this one together

“I don’t like to give good people bad news” – a few facts:
As a co founder Joachim owns a hefty share of the company.
When Unity goes public said share will turn into an astronomically large $ amount.
Nicholas or Helgasson too.
Large financial entities are the ones deciding of the succes of an IPO, not the customers.
The current CEO has many friends in the financial industry.

JD said – Unity hired a rat who almost sunk EA as their new CEO. And I was like wtf?! So I googled “Unity CEO” to find out more. Wow, I can’t believe what I just read in that Forbes article :o I’ll refrain from commenting since the trial isn’t done yet.

But, I think I found the real reason why “production ready” tag is being slapped all around! And why the painted public image in blogs and conferences has nothing to do with reality. It’s in this sentence from Forbes: Riccitiello, who took over from the company’s cofounder David Helgason, was expected to take Unity public as early as 2020.

There you have it, gotta go public guys, sorry about your projects and businesses.

Can’t wait to be done with my project and make a switch. I just feel sorry for Unity employees who must be working like crazy to meet unrealistic goals.

We’ve all seen this scenario before. Devoted engineers with a vision make something wonderful and successful. But then they pass it on to corporate guys and … you know the rest.

The Forbes article is disturbing indeed… It also mentions Unity going public:
“Riccitiello, who took over from the company’s cofounder David Helgason, was expected to take Unity public as early as 2020.”

Its been a blast getting to know members of this community and I had lots of fun over the years. As of this morning my company is retraining staff into unreal using hired in experts and converting propietary code base to unreal.

I wish everyone the best of luck, and hope you guys also find a way to get out of the hole you are forced in. I am really sad to go but I am also too tired to stay anymore.

Best of luck guys and keep making awesome things!

Totally agree with those talking about being with Unity for so long and are having doubts.

I have been with Unity since version 2.64f1… and a Pro user since that time, 3.x, 4.x and the horrendous 5.x cycle.

I took a lot of cheap shots by my colleagues when I worked for a top-3 AAA developer and publisher when I told them, look, Unity is getting serious and you’d better watch out – way back when Unity 3.x was released.

Today I would not defend Unity against their volleys about it not being a legitimate game engine.

Unity hired a rat who almost sunk EA as their new CEO, who clearly was hired to raise the valuation of the company. From that, Unity threw out their principle of democratizing game development. While they moved to a subscription-based model and are now increasing the cost of the subscription, other game engine developers as well as open source projects are offering more for free to the community – including online services.

Through all of this, the stability of the game engine has gotten worse. So Unity’s direction is higher cost, lower reliability, no more democratization all the while other game engines are either catching up or are on the cusp of making people switch over to something else, including myself.

When someone’s been with you since 2.64f1 and they’re seriously considering saying it was nice but I don’t know who you are anymore so goodbye, it is time for some soul searching.

im used to all the whining and shouting out people who must show their disrespect and anger over unity’s work.
I relly think their transition was good thing, we had almost never problems with unity that we could not solve.
About your disrespecting words, you really should carefully hold on and think what you are saying because saying other they are rats qualifies you as a very unpleasant person and i really hope i never have to work with you in any moment in my life! Look forward, deal with the problems you face and keep your shit together!

pachermann –

If the shoe fits, the shoe fits. Some people take issue with the truth when it is said bluntly, that’s fine, there are a number of people like that in the world.

I wouldn’t worry about us every working together; I work with people who can handle the blunt truth instead of wasting time trying to be diplomatic or keep things that need to be said to themselves. They usually end up dragging down the productivity of teams rather than helping them progress.

This is a painful transition but I believe in the change. The biggest pain point in not knowing how to navigate the dependency issues. Simple things like no tessellation in shadergraph cost us dearly, switching a project to URP and realizing it doesn’t meet our current needs then switching back was very upsetting especial because you just dont know what will be a show stopper and how much work will be lost until you hit it.. I just have to keep believing it will all be worth it in the end

Sadly we have to agree with many of the comments here and in the forums over the past 18 months. Unfortunately for us and many other devs we know, Unity has become a constant frustration of bugs, regressions, unready and unstable new features, and an increasingly complicated and slow development tool to work with.

We are a long-time Unity Pro licensee and now subscriber, but it has become pretty hard to feel confident about the state of the engine, new releases, and timeframes and plans for future features. Unity has lost a huge amount of our trust with 2018.x and 2019.x being so unreliable, particularly this year as it felt like we spent as much time working around all of the Unity problems we encountered as we did working on our project.

Unity Editor performance on Mac seems to be regressing every release. Engine and package stability is patchy for us at best. Long-standing bugs we repeatedly encounter never seem to receive any attention but surely we are not the only ones hitting and reporting them. Documentation is often outdated and much of the Editor api remains pretty awful.

Fundamental issues with serialisation have caused us a lot of headaches this year, with Unity losing references in scenes and scriptable objects all of the time, renderers losing assigned materials, and taking multiple reimports or Editor restarts to pick up changes made to any type of asset on scm branches. Nested prefabs sounded great in theory but serialisation of scenes and overrides and nested prefabs has turned it into a bit of a nightmare. We talked to devs of a popular editor inspector plugin and they have abandoned support for nested prefabs and suggested that their support channels are full of the same sentiment, that serialisation issues in the editor have caused devs a lot of trouble this year, and nested prefabs are extremely unreliable.

This is both a feature of a flexible engine and perhaps a telling state of affairs, but we seem to be increasingly disabling built-in Unity features and replacing them with third-party plugins just to get basic stuff working with far less issues. Asset Store and plugin developers often offer more reliable features, better apis, better documentation and far better support than Unity itself, and they seem to appreciate the trouble devs have with the native features more than Unity itself. It is indeed a bit cynical to suggest that Unity is clipping the ticket twice here too, once for our Pro licenses, and again as they take a cut of the Asset Store revenue for plugins that replace core features.

Our current project is using plugins to replace input, the sound engine, and the editor inspector and api, and we are using shader builders, async libraries, leaderboard, achievement and localisation plugins. It is great that these plugins exist on a marketplace as a choice for devs to use, but we are past the point where we consider them optional because first and foremost they are making core features usable and reliable.

Unity has a handful of staff and community moderators on the forums and they are tirelessly answering questions and helping devs out while maintaining a respectful and professional manner. However, the beta forums are full of devs posting information and sharing their experience of an engine that is not delivering the features, performance, and stability they need. Unfortunately, Unity as a company appears to be a bit tone-deaf on some of the issues and does not seem to be responding well to this feedback. Often the threads are left unanswered which feels like Unity is simply sticking their heads in the sand about various issues.

Surely this is not the case, Unity has plenty of smart people building smart things… but read the posts below, look at the forums, talk to devs at conferences etc – it seems to indicate that more than just a few noisy devs are experiencing issues and pain points.

Good luck with the upcoming releases. We hope they are great but we are holding off longer and longer before upgrading these days because it comes with so much baggage. We are not expecting to use many, if any, of the new features over the next year or more.

Somewhat selfishly, we would really like to see some releases that focus on bugfixes, regressions, upgrading of core features, and some big editor performance gains.

Adding to my other comment: I hope these Unity problems doesn’t lead them to Microsoft arms. If that happens things are going to be much worse in the future…

I have been with Unity 10 years now. I also bought perpetual licenses worth of 5k and now I am with Plus. All this engine journey tells me that things are going to become much worse in the future. Why I say that? Because Unity management team doesn’t seem to listen to its core clients (which are the indie studios not AAA) and instead of making the engine easier to work with its making it much more complected and malfunctioning. The problem here is that I have many projects with Unity and I cant simply quit and move to another engine because I need to support my existing projects and not lose valuable time learning a new software from the beginning. So I really hope they somehow magically save the day because I cant afford to start all over from the beginning again. Maybe its not a bad idea for Unity to buy a small dev studio so that they will experience the core bugs first hand and maybe address them faster.

So as it can be seen, the current situation with Unity game engine looks like Unity team wants to bring all the new technologies but without real success. Old Unity stuff is really old and useless for next-gen games, but new Unity stuff is far not ready and has bugs. It seems like Unity team wants to make too much but they have not enough skilled professional people or they just produce too much chaos.

A good article –

Thank you Tomas for your detailed info on the situation. I understand that the complexity of what you’re dealing with is unimaginable.

Please also try to understand the position of us small devs. For example, here is my situation. I’m full into production, using the builtin pipeline and my game simply can’t work without some stuff that isn’t there in URP or HDRP yet (like RenderWithShader function). I’m in talks with a publisher and we want to get the game to PS5 eventually. But I have no idea when builtin will be deprecated and if it will be supported on PS5 at all. We’re also considering publishing in episodes, to try to start earning money early from the first episode while working on others etc. So, how many episodes can we plan? That entirely depends on how much more time with builtin we have. And we don’t know. You said builtin will be there for a “few” years. But I can’t tell my future publisher that we will have a “few” episodes and will need a “few” amount of $.

Deprecation of something so important as builtin rp should be outlined and communicated extremely clearly because many devs are so deep into production that they simply can’t switch without extra budget for that extra work. It’s really an important matter that I think you didn’t address seriously enough. And all this hurry with new features and labelling other render pipelines as ready smells like you just want to get rid of builtin as fast as possible.

Only when you start seeing actual games being made and finished on URP and HDRP by all sorts of devs (not just the vip ones who get extra support from you), might be the time to start thinking about builtin deprecation.

That’s just my situation but I see (as monoflow said) it’s equally frustrating for some who started development on URP and HDRP too soon. I’m sure many people started developing on HDRP/URP because it felt like builtin is soon to become past and losing support, and not becuse they really needed new features. Feels like it was all announced way too early.

In the end, it’s great that Unity is improving and all the planned features are great, but I can’t bet my hard earned budget and livelihood (or someone else’s funding) on something that’s not ready and not battle-tested. So, for us small devs who make business with your product, certainty and stability is paramount.

I’m very relieved that you said you’ll give us a fair warning when it comes to builtin deprecation, thanks. I hope that means an announcement at least three years in advance since game development normally takes years and people often support their published games with DLCs and updates for 2-3 years after release and I guess you know we’re forced to update Unity versions on consoles.

Sorry for such a long post, I thought it’s important to communicate how it feels from our side and hope this feedback is useful.

Here’s an idea! A clear blog post about the future of builtin! All I could find were a few words from Arisa Scott at Unite and it wasn’t really clear enough.

Let me decribe my feelings working with Unity over the last year:
I did a great mistake last year to start using HDRP. If i had known what a freaky journey i had to expect i never had made that decision. It is a daliy struggle to work with Unity at the moment.
You totally lost the focus on the developer community with all your new features. Your public image you try to paint with your blogs or Unite conferences has nothing to do with the harsh reality working with Unity at the moment.
After a year of pain i came to the conclusion: Use the old render pipeline with all the nice third party assets from the assetstore and publish for Standalone, VR and iOS from one project without getting insane.
I’m using Unity since 2010 (and make a living from) and i catch myself thinking about if the Unreal Engine would be the better choice to work with (crazy isn’t it?).

Hi Thomas, thanks for the reply. I’m pretty excited where Unity is heading with packages and the freedom it brings to developed. Loving the features and appreciate the hard work in this transition phase. One think that could potentially ease our pain is knowing what packages have been verified against other packages. Currently it a bit of a minefield and some sort of map to help us navigate dependency issues would save us so so much pain. I’m not exactly sure what I’m asking for maybe some sort of table, I’m sure someone has a better idea how get us this information.

The problem here isn’t that Unity accepts accountability – let me tell you – you don’t have a choice. There is no scenario where you accept whatever.

What happens is you don’t knowingly let this happen again. Every problem that came to this point, was knowingly embarked on. At every point in time, Unity knew where it was.

The fact is if it happens again, this kind of mess, you risk losing long term members like me. Because if a company is willing to take these kinds of risks with my business then I know I’m a fool for supporting them.

Simply: it won’t happen again.

The problems come with the “Production-ready” tag. I understand why you did it but that is not an excuse.

I started my next project with the “Latest Official Releases” according to Unity HUB (2019.2) which I expected to be “Production Ready” since it is official and a release, but the amounts of bugs and workarounds that I found has been crazy.

If I want the latest most stable version of Unity and I can’t trust your latest official production-ready version, How I am supposed to choose the right version?

Hey everyone,

Thanks for lots of good comments. I want to try to address some of the issues mentioned to the best of my ability. Please don’t take it as an excuse, we accept accountability of the current state.

First up packages. The promises of packages are many and good. Packages enable us to split out features from the core code, which should reduce the code churn on core and thus stabilize it. It also enables us to open source more things, which we can’t do for legal reasons on the core. With the out of band shipping, it will also be quicker to fix bugs in package versions and ship them without having to ship the core.

Those are all the good parts. The trouble we have been facing moving towards this new world is a combination of codebases being separated into different repos, dependencies between those repos, and the infrastructure for testing them all coming together to form an extremely complicated problem. For context, we run almost 1 billion automated tests every month, across many, many platforms, and they need to be ultra stable to function as a smoke detector. Because our “old” infrastructure was set up to do a mono-repo and single install, we have had to build a completely new tech to handle both execution and scheduling for the packages on top. This work hasn’t been as fast as developing the code in the packages, so we have not been on the same safe ground when it comes to testing all these packages together as we have had on the core repo.

Which leads me to “Production Ready”. The packages promise of shipping directly and out of band from the main engine has the benefit of putting out a “Preview” package, in order to get feedback on API’s and functionality, before locking it down to a backward compatibility workflow. Conversely, we decided to implement the “Verified” tag for packages which were tested and shipped together with core in order for users to make an informed decision on what to base a production on. Hence, “production ready”.

Did we nail it? No, we did not. The dependency testing is by far the most complicated problem I have seen in my time with Unity. It creeps into the production of the packages, because testing is obviously not the best place to discover problems in a dependency chain. We are still working on solving this with our infrastructure and it is a very high priority for us.

Finally, built-in renderer. It will still be around for a few years, while we improve the renderpipes. They are in many regards superior, but we recognize that many, many current productions are committed to built-in. When we deprecate it completely, we will give fair warning.

Now, all this said, I think we’re building a future where you will have access to a fast, small, stable core upon which you can add the features you need for your game without carrying a massive overhead of everything-in-one. What we need to do is to make the journey towards that smooth and delightful.

@ Joe
“WhoTF said you had to upgrade?” When you start developing and publishing games on consoles, you’ll find out who says you have to upgrade.

Its not about upgrading, its about the current “stable” version. What counters at the end of day is to have a stable version to deliver stable builds and upgrades to the user. Any if upgrading between minor version like Unity 2019.2.10f to 2019.2.15f is not possible without fear because of new regressions and more than two months no fixes it doesn’t matter if a new feature will be available in a few months. Will love new features too, but in our point of view there must be either a bigger team working faster on regressions or regression fixing must be of higher priority.

Please note that I don’t talk about bugs in new features, but regressions. There are times you have to upgrade (e.g. new Xcode requiring a new Unity version). And this is where things get complicated with regressions (sure they can happen, but they should be fixed way faster after being reported, or at least, better communicated if and when they will be fixed)

We are looking forward to DOTS too, but what counters at the end of day is to have a stable version to deliver stable build and upgrades to the user. Any if want get upgrade between minor version like Unity 2019.2.10f to 2019.2.15f because of regressions and more than two months no fixes it doesn’t matter if a new feature will be available in a few months

user said “Pause stuff like DOTS”
erm no thanks, keep going with DOTS… The results are shocking and no other engine comes close to that sort of performance gains.
DOTS is amazing and the team behind it are geniuses that deserve a huge applause for what they’ve achieved already.

wow so much fear people…
anyone care for a snowflake?
Look this “im so scared to upgrade” WhoTF said you had to upgrade?
If i want more stability i use unity 2019.2.16
If im willing to have more bugs but take advantage of the new features then i use unity 2019.3 . and thats put back to beta.. anyway i use it and it has little bugs here and there.
Maybe we can understand that yes LWRP was the original plan but as time goes on they see that it wasnt the ideal strategy and changing later would do more harm. its better to get the fundamental foundation right and build on top of it. does URP perform better than LWRP? YES IT DOES.. so its better to build on top of that.
its not like they will change it later its just they have a better understanding now than they did before.
work with them not against them and be more forgiving.
Report the bugs you find because they are waiting to repair them.

That would mean no new features ever.
Seriously.. just look at your game you’re making, do you think you will have all the bugs fixed before releasing? no of course not, because sometimes it takes other users to bring those bugs to light.
I think overall unity is quite stable compared to other engines. i would even go as far to say that it it the most stable game engine available to us.

I agree with everyone here too. What’s the rush? You shouldn’t have even mentioned lwrp and hdrp until they are really near production ready. What you did is promised stuff too soon and neglected builtin rp, thus leaving us in a void between something that’s working but being abandoned and something that isn’t going to be ready for a long time.

I’ve been using Unity since version 2 and this is the first time I’m wishing to have learned another engine. Working in Unity has become frightening, the fear is always present. Fear of updates, fear of uncertain future, fear of serious bugs (Last time I updated my project, I had to rework my game’s physics from simply using colliders into working around bugs with using raycasts instead (issue 1181054)), and the fear of the builtin pipeline being deprecated before our game is done. So many questions and whenever I asked, I got no reply. So there’s fear and silence.

We’d much more prefer stability, predictability and concrete answers about the future instead of getting plenty of new features fast. In the end, that’s what’s most important for doing business.

Don’t rush a major version out there. We can always use a previous version. It’s brave and wise to postpone this release because it has so many important bugs. Please pay attention to VR issues, the API upgrader, the documentation and try to avoid any undocumented breaking change please.

Yeeeaaaaaahhhhhhhhh… I really have to agree with everyone here. I’ve been using the 2019.3 beta and it’s… not great. There’s still editor rendering issues, stuff like the font hinting is a mess, there’s loads of stuff that just straight up does not work and I can not possibly see working within the release window. I know that there’s no way to have this release pushed back but I genuinely hope you slow things down a lot and at least give us something that isn’t a total mess when 2020.1 rolls around.

I agree with the devs here as well. It’s becoming too big of a risk to update to the latest version right away.
Pause stuff like DOTS and all the other half finished features for a while and fix what you have.

Unfortunately, I have to agree with a lot of devs here, Unity should just slow down and work on stability and bug fixing.
Currently Unity release a buggy “production ready” version every 3-5 months, and fix bugs on the next beta. I don’t know how they can think it’s a good idea, it’s just nearly impossible to get a stable release this days. For the love of the god Unity, please just fix your damn bugs and issues before going next.

Really sad not getting the latest release in the year it was promised. And to stretch it even more a 2019.3.0f# was released recently and somehow considered a Beta. Never heard of this before or presented in previous roadmaps, posts, conferences, etc.
As more than 50% of Unity is either in “preview” , “beta”, or having a new or regressing bug, this just drops the expectations even lower…

I’m starting to think that Unity starting a studio might not be such a bad idea. The only real way to call it production ready is to put it in production.

Please focus on stability rather than features we are end customer of your product not alpha/beta testers.Since unity5 i have made a thumb rule don’t upgrade to newer versions for at-least a year however shinny they may appear.

Please don’t make a habit of doing final releases the following year. So next year I would expect 2020.3 to be released in 2020, but 2020.4 LTS to be in 2021, which is fair for long term support. Slipping like this, points to unexpected management inefficiencies.

I am using 2019.3 for some weeks now and its not more buggy than 2019.2. I agree with all other developers here saying that Unity should try harder to eliminate bugs and regressions. All Unity versions are more or less unstable but after some workarounds they do the job needed to be done. We all like new features but we also want stability. I personalliy like 2019.3 and I hope it didnt carry bugs I see (and also reported) from 2018.3 and new regressions. Also I hope it didnt crash on me at least 3 times a day. (2019.2 crashed too)

Congrats guys.
Unity is becoming the beast i always wanted.
I know you’ve taken a huge risk by introducing very powerful features.
You’ve worked hard and we are seeing results.
sure developers took a hit on asset compatibly etc but you’ve done the right thing.
its important to evolve and push this engine far past its competitors.
the question many people will be asking others in the future is “why wouldn’t you use Unity? it doesnt make sense not to”

Unity should slow down a bit and focus more on quality instead of quantity releases.
Not much of the introduced tech is production ready while at the same time the old, proven systems are getting little to no love.
It looks like when their replacement tech ideas are promoted to development level, old tech is getting deprecated, or at least getting sidetracked, asap.
IMHO things went downhill after Unity 5.

I wouldn’t mind 15 hotfixes if they will finally make “Update engine” feature for the minor releases. Deleting and installing the engine for a proper minor hotfix seems pretty dumb for me in 2019.

Could you make only a few hotfixes per TECH cycle? One TECH release could have 3 hotfixes instead 15 or so. It’s too much.

Will LTS come in late February or March ?

I feel the editor has become too bloated and clunky. So much so that I’m starting to wish Unity Tech would just drop it and let us use the underlying engine instead. The engine is also becoming so complicated that I’m beginning to wonder why anybody doesn’t just use OpenGL or Direct3D and C++ instead.

Well, that is the whole purpose of democratizing game development. The previous render pipeline wasn’t ‘scalable’ enough for large, high quality games (it was very much possible, but you had to do a lot of time consuming work to extend functionality, or you had to download assets from the asset store). URP is pretty much reliable enough in its current state, cannot comment much on HDRP since I haven’t used it.

As long as we have all of these new features and it gets constant updates then I am happy. What makes me sad is seeing a feature get introduced and not get much love (I am looking at you New Input System, but I am glad they have started making progress on it once again).

I hate to agree with most other comments: by adding too many features too fast, many of which are still far from being production ready but jeopardize the future of stablished, robust alternatives, Unity is becoming a bloated and disorganized mess.

We used to have one render pipeline, one physics engine, one UI system, one package system… now we have 3 render pipelines (none of them reliable enough imho), 3 physics engines (2 of them that solve no real problems, and introduce new ones), wait, 3 UI systems, 2 mutually incompatible package systems….sigh.

Dear Unity Team.
We are really appreciated that you can’t wait to share new very cool stuff you are working on with us. But please hear us; You are announcing a new tech or feature almost every other week while the old ones getting lost and never get into production-ready phase. And stop calling those very buggy releases as prıduction ready.

Maybe 3 tech releases and 1 LTS release for every year is a bit much? maybe you should slow down, maybe you should hold your horses and take your sweet time with improving existing features. I have been using Unity since 2011 and I must say that last while 2 years bring a lot of new cool stuff, we kinda tired of dealing with bugs, finding workarounds and never being able to sure if our projects will be damaged. I have been working paranoidly with Unity in 2019. Constantly pushing to git at almost every click.
In 2019 Unity lost my trust.

Can you please calm down and take more sturdy steps.

By the way. Still, I am very thankful for democratizing game development for us. Have a good new year.

This time we will skip the next major release 2019.3 until at least a year has passed with bug fixing. Too many bugs (build pipeline, android/iOS runtime, shader compiler) we have to wait sometimes for months until they get fixed in 2019.2 which always takes quite an effort to workaround. Can already imagine how this will quadruplicate in 2019.3 with that much changes

Gonna throw my own wrench into this, but I too feel that Unity is not doing things well… All the features are nice and needed, but they are *far* from being “production-ready” and need a lot more development. I honestly think you people are trying to bring lots of features all at the same time, but that’s arguably wrong, leading to all this nonsence when various features are clearly not finished and are just left there to die.

Unity, please, pull yourself together. Sorry.

Please give us some insights what you mean with Production Ready. Do you have a quality state above this, and what are the characteristics?

Like the other users below, We are also not using features which are called Production Ready until ~6 months after announcing that. Maybe you can communicate it a bit different. We know that you must push out new features etc. so that stake holders and others feeling quite confident, but at the end of the day, your target audience has a lot of technical background and can see the hurdles and bugs in it really fast.

We do not think that everybody is just marketing driven (even in the current time).

On our wish list for 2020: Please communicate more honestly with features which are in the works (or nearly Production Ready) and do not rush out new features or mark them as Production Ready (or communicate more openly what Production Ready means for you).

At the end of the day, this will give you more reputation then just the marketing bla bla…

I know this isn’t contributing much to what has already been written, but I have to give an amen to Isaac, and his very real plight. There seems to be a silent mutual agreement between seasoned Unity users in regards to the state of its’ features. If you’re new to Unity and reading Isaacs’ comment, you should know you have just caught a brief glimpse of the scary thing that lies below.

@Isaac Thanks for putting it into words.
I’m a programmer that works with an artists, and while I was able to update without not much pain before, now I’m frightened to even update between LTS versions, because when I do you never know when a new bug that wasn’t there will appear, something won’t work anymore or I will simply break my artist’s workflow.
Right now LTS means “less bugs that we may not fix” while non LTS means “cutting edge with many NEW bugs, but LTS bugs may not be here”. It is frightening to work on unstable software.
So many problems that I just don’t want to report them anymore, I prefer to work around them.
Exciting new stuff, but not so much when you know it will be abandoned mid-development.

I wouldn’t say this if I didn’t care about Unity, I want this tool to be good like we all, but I keep on remaking stuff Unity abandoned because I can’t use it in that unfinished state and it doesn’t support building on top of it.

Please add support for layered terrain shaders and foliage in HDRP 7.1.6. Please optimze draw distance for foliage.

If Unity is actually producing a game and call it production-ready, then I’ll bite it. Unity has no idea what it feels like to producing a game. I mean a real AAA game, not a Flappybird.

2019.3 will be the THIRD release where URP (former LWRP) is in Production Ready state. With what I saw in RC1 it still has a lot of bugs out of the box and I can’t, from my developer perspective, call it production ready.

So I will take all production ready stuff (especially HDRP) with a huuuuuge grain of salt. And I won’t recommend anyone to use anything “production ready” from Unity at least for a year.

Re: Isaac – I can confirm “production ready” has become a euphemism for “janky as hell” at our studio, thanks to Unity. I appreciate Unity is expanding the feature set aggressively and there is some really cool stuff in there, but the marketing team desperately needs some restraint in branding things as “production ready”. You are dealing with some really seasoned devs here after all, who can see under the hood and right through the bs very quickly. There is some work to be done in rebuilding credibility.

This needed to be said. Thank you @isaac! The pain these processes, misrepresentations and claims cause are very real, even if the docs are not.

I had hoped you guys had learnt about calling stuff “production ready” after the absolute cluster**** that has been the last 2 years.

Maybe take a step back and stop trying to tout everything as being ready when -ANYONE- using the forums or git repos or packages knows full well they very much are not.

So if HDRP is “production ready” what is URP when that was meant to be out of preview already? Your conventions on what is ready and what is not is currently a complete joke and some clarity is needed for people so they dont make the mistake of jumping on these bandwagons and then pulling hair out literally a year down the line from when they were meant to be usable.

I am looking at you timeline.
And you addressables.
And you HDRP/URP.
And you shadergraph.
And you VFX graph.
And you navmesh surface components.
And you lightmapper.
And you AutoLOD
Should I go on? Because the list of unfinished semi functioning systems that are well below the point they were advertised to be is extremely lengthy at this stage.

Like, I am a unity fanboy but even the hardcore fans at this stage know that something is not right.

Dont get me started about docs surrounding all these things.

What docs you say? I know brother, I feel your pain.

Agreed, Mechanim issues, UNET deprecated, Terrains, UI, the list covers all areas. Not to mention other design flaws that creates performance issues in large games (runtime) and Editor have been in a spiral downhill since 2017. This is what happen when Unity focus on services that no one really needs instead of focus on the core product, the game engine itself.

I’d be happy with a working Unity Hub and a working Android build system out of the box if the Android options were selected during the installation process of Unity 2019.2…

Comments are closed.