You know, whether there is a bug here or no (can’t even view the video, it just buffers on and on), why did you wait for release? Why not pitch in during the beta to see if there are any issues in the areas you’re working with, and report them? Problems aren’t going to find themselves.
1 Like
I was too busy with work till yesterday, as simple as that…
1 Like
Beta didn’t start yesterday though. Ultimately, if there are indeed bugs in release, it’s as much our fault as the devs.
Maybe if you move your sphere for example to z -0.1 then it works ,can you try this
I think that I got it figured out: move the shell a bit up in object mode, and it will start working.
There seems to be an issue with both objects being in the exact same original position.
When I reproduce your steps, it will not work until I move the shell a bit in the z-axis.
This seems indeed to be a bug.
PS I just noticed it does not matter whether the original object or the new shell is moved in object mode. Once either one is re-positioned, it works.
1 Like
I can reproduce it on my end. It does appear to be a bug.
If your shell exactly overlaps the object its supposed to shrinkwrap it seems the offset value does nothing. Almost like it can’t decide what side to offset to.
the moment you remove this exact overlap by say, using an object with different geometry, or moving the shell or the target object a little bit, the offset value will start working again.
I see Herbert123 beat me to it so lets just say I can reproduce it too. 
PS
Just checked in 2.80 The bug (I think we can call it that) is present there too.
3 Likes
Yup, if I understand the problem from comments (video is still buffering oO), On Surface + Offset doesn’t work in 2.80 either, unless objects are not exactly overlapping.
I don’t agree with this. While some people are more than happy to be beta testers for software others are not. They won’t use something until it’s final release. Beta testing takes significant time and if you are working on a production using beta software is not a good idea in the first place. The situation in Blender is a bit different as there is no actual support to deal with. Beta or otherwise you are pretty much in the same boat but there is an expectation (due to past history) that major issues will be stamped out by final release and that’s when users jump on board. Not everyone follows Blender’s development they see a new official version download it and just start using it.
2 Likes
Right. “Let someone else do it, not my problem, I’m too busy yadda yadda”. Awesome.
1 Like
This attitude has always baffled me.
One one hand everyone is cheering for blender to be more widely adopted and used on commercial projects, preferably big and high profile like Next Gen.
On the other hand, people who use it for exactly that are spit upon when they expect Blender to be fit for that purpose.
When they say they are too busy with actually earning a living with Blender, to spend their time beta testing they are ridiculed for it.
What the actual bleep. It is perfectly fine for people to JUST use Blender and have certain expectations for point releases.
Now you can argue that his initial post was a little over the top. Sure. We are all people. We all get disappointed sometimes and react a little fierce. But to ridicule him for saying he does not beta test?
Come on now. That is just ridiculous.
2 Likes
Where did I ridicule @jacobo, exactly? What the actual bleep indeed. “Spit upon”??? Are you reading the same thread or what?
Bugs don’t find themselves, it’s as simple as that. Weird this has to be even said.
Stan is way off course. Although what he claims seems to make sense from a distance, it just goes to show that he has no regard for those who would be willing beta testers if they had time on their hands. For 3 months I’ve been wrestling with a difficult job (and taking on side jobs) and had absolutely no time at all to devote to setting up Blender 2.81 alpha or beta to check for bugs. Granted, if I did, I would have detected these very bugs that came my way today beforehand, regardless of the fact that they may have been addressed before the actual release by the devs or not… But, I simply couldn’t find the time till I completed the project yesterday, and that’s it…
2 Likes
Ok, then there is a bug, it was not obvious, but bug it is, now all is required is to report it and wait to be solved, download the next day buildbot build and problem solved.
That was a reaction to apocalypse who comment on your reply to jacobo where you inferred this bug was his very own fault for not partaking in the beta.
Now I’m not going to turn this into a quote war or a he said she said. You know perfectly well what you said and to whom. As you can see, I disagree.
I think this thread has been very useful in finding a bug that appearantly has been present since at least 2.80. Now it can be reproduced and it will be resolved in time. Win for everyone in the end. 
I agree with Stan. Devs can’t fix bugs if they aren’t aware of them.
The nightly builds run right out of a folder without installation. For some reason a lot of people are too busy to test betas but conveniently are available on the day of the official release. 2.81 was even delayed a week.
If people want less buggy official releases, they need to step up and test their files/workflows in the nightly builds before the official release.
4 Likes
The real question should be how this (relatively small) issue flew below the radar since 2.8 and wasn’t picked up earlier by anyone: developers, beta testers, or regular users.
I myself never picked up on this, and I use the project option quite often. But I never left both objects in the same position, and therefore never noticed it before.
Even when trying to replicate this problem, I initially moved the shell away from the parent object in the direction of the wrap. Which meant it just works as expected.
In effect, it is a relatively small issue, with a simple work-around (just move either object away by anything but 0 (even 0.0001 positional offset will fix this).
Anyway, report the bug, and I am sure it will be fixed in the next betas.
As for who is responsible, or whatever: I have worked with software since c64 times, and there are always obscure bugs. This is one such relative outlier as well, it seems. Otherwise it would have been noticed a long time ago.
Software will have outlier bugs. Until someone (often accidentally) identifies these and reports them. No-one is to blame, it just happens and is part of development (as a web developer obscure bugs happen all the time). I myself identified tenths of outlier obscure bugs in PhotoLine doing stuff no-one else did before in PhotoLine.
As for expectations regarding so-called “point releases”: it is not as if awarding that predicate is a magical guarantee for a completely bug-free release. Gods, Photoshop has had some pretty bad workflow stopping bugs, some of which still up till day haven’t been addressed. But these remain outlier cases.
Bugs identified by users present in C4D, Modo, and Lightwave took a complete new release to be fixed, and often even years, or are still left unresolved (Modo I am looking at you!).
Pointing fingers at who’s to blame for outlier bugs is just a pointless exercise (pun! See what I did there 
)
4 Likes
There’s no need to put words in my mouth, thank you very much. Disagreeing is one thing, building strawmen is another.
And to be clear; what gets to me is the fact that there are hardly any bug related releases between point releases. Now, I must admit that it was astounding to read on this thread that this bug persisted till 2.8 and it wasn’t addressed till then. Instead they diverted their attention to file browser and etc. There are admittedly many cool features in 2.81, but for the sake of what? In this instance, two objects can very well share the same transform space and the shrink wrap mod is supposed to work in that case in its default state. And since this was reported, why wasn’t anything solid done about it, and this was allowed to seep into this point release?
Careful now, lest they start saying you “inferred” something inappropriate.
In this thread we’ve found out that at least with your test case the behavior in 2.80 is the same as in 2.81, not that the issue was already known. Unless I missed a post. Was this reported before?..