Multitude is the correct term. I put in a lot of effort doing the same some months ago and gave up. I have to stick with Rhino 7 indefinitely.
We eat bugs for breakfast and appreciate everyone reporting. Some bugs are easier than others, I submitted this one yesterday and the fix is ready for testing. If we can repeat it, we do our best to get it resolved in a timely manner.
I added a test command for you for the feature you were requesting last month. Have you had a chance to try it?
I wonder if this is the object properties panel. It does an awful lot of work when things change in the document. I would be curious to know if you see this behavior when the object properties panel is not visible. If it is to blame, I can try to optimize this.
[edit] Iām going to investigate seeing if I can get the object properties to notice this specific situation and skip rebuilding itself.
Hi @stevebaer,
Sure, Iāll keep an eye out for this, and when I find it again, Iāll save you a file with just that part. If when I save with that part alone does not show me the problem, Iāll have to give you my full ~1GB file.
Re: Push-Pull
This is a much bigger ask. The tool is too limited for me to explain and document all the ways it fails. This definitely falls into the category of an internal expert testing project or an outsourced paid project. If you want to hire us to do this, weāll happily provide you with a quote. We are very thorough and efficient, so having us do this could be of excellent value to you guys.
IMO, this is the most worrisome and critical situation. Are you aware of these post-gumball delays?
G
When the gumball drag operation completes, transformed objects are replaced in the document and an event is issued that objects have been replaced. This event is listened to by several parts of the Rhino user interface as well as 3rd party plug-ins. We try to optimize so we can efficiently handle this event, but it is impossible to control how much work 3rd party plug-ins do when this event occurs. Right now it is hard to determine what the offending piece of code is for your specific situation.
Some of the form issues would literally take minutes of coding to solve. I would guess that whoeverās in charge of that is overworked. With all the tech lay-offs thereās gotta be some hungry programmers who would be more than glad to sweep through the forms and sort the loose ends out.
It would be helpful to know what these specific issues are. It is nearly impossible to track with a long thread like this so if you can point me to what you are referring to I would appreciate it.
The Linetypes dialog - the one accessed through the options menu is pooched. If you change the line thickness unit it doesnāt hold onto it. The dialog accessed through the ālinetypesā commands works better but reverts to āpixelsā when you duplicate a linetype. The later isnāt a big issue but itās a sign of sloppy/rushed code - I mean the person that coded it is probably a hard worker but just over-worked.
Thereās lots of little examples like these maybe I can just make a list. Although these donāt sound like big issues they can really gang up on us.
Edit ā I did bring this up a while back too it never got fix⦠lost in the queu I guess.
Thanks for the explanation. I understand, but it MUST be figured out and resolved. I see this current performance as unacceptable. Are we aligned that this is not a state of performance that we should accept? I donāt see how I could use V8 if your team is going to accept that this level of performance is the new normal.
G
Are you dragging a single object when you see this lag?
Thatās exactly what Iām trying to do.
Iāll continue to work on an optimization for the object properties panel and we can see if this makes a difference in a couple days.
https://mcneel.myjetbrains.com/youtrack/issue/RH-86626/Linetype-width-units-not-sticking
thanks for the report, unfortunately I never saw your first mention if.
I just issued a fix. Let me know it if still acts up.
I have not. Been busy with other work.
To be clear, my complaint above was in regards to the myriad issues Iāve had with Rhino 8 in months past that have mostly gone up in smoke. The object properties menu was an issue for me before Rhino 8.
As there arenāt any specifics in this thread beyond this issue that you posted above, I will need to go through your post history to find more
@gustojunk we have a diagnostic that could help narrow down where the lag after drag is coming from. I donāt want to adjust code and make attempts at optimizations when I donāt know what the cause of the problem is. Please try the following
- Start Rhino and open your model
- Run TestTrace and change Enabled to True
- Drag your objects around a few times. If weāre lucky in this case there will be a case where a large delay occurs.
- Run TestTrace again and this time choose the Save option. This will save a json file that will contain information about different events that occured and how long they took. Please send me this json file.
This json file is human readable and only contains a list of events in Rhino and how long they took.
hi @stevebaer,
I had the problem twice today, but every time I try to repeat it to capture it on video, it does not happen. But Iām pretty sure the second time around, I still had the TestTrace running. Itās the last command I run before saving the test trace.
https://www.dropbox.com/t/KdhKOOHV1G6wvut0
testtrace .sjon file here:
test_trace_v1_gf_250322.zip (52.8 KB)
Please let me know if you see the day in that file.
G
The video doesnāt seem to work for me.
Thanks for doing this; we havenāt had many users try this tool out but I think it can become something very useful for figuring out what is causing lags in Rhino.
This didnāt point at a 20 second delay, but I do see a 1.5 second delay happening at some points in time which can be jarring and is not acceptable. Iāll see if I can tweak the tracing code to help point at exactly what is causing the delay. It does not look like it is the properties panel which I had assumed was at fault.
Well maybe this tool is not caching the very long delay. Or by mistake I had stopped the trace process before the long delay happened. There were a few 1.5 sec delays but that wasnāt as back as the much longer one. Maybe it wasnāt 20 seconds but it was at least 10.
Iāll place closer attention next time. Iāll also check on the video soon and resend.
G
You are absolutely correct that this may be coming from something that I havenāt set up tracing for. Iām going through the code today and seeing if there are other suspect areas that we should be including tracing data.
hi @stevebaer,
Re: video link. For some reason discourse now disables dropbox links? ( changed the link above with a dropbox file share link). In the video I show the type of situations when this happens, but of course it didnāt happen while recording.
This is the link i had shared: removethishttps://www.dropbox.com/scl/fi/2u2atbbk8ic7nxohg0c3q/test_trace01_gf_250322.mp4?rlkey=7eae3b4qs999txgegqb6lra8c&e=1&dl=0
if you remove the characters āremovethisā above and you paste that URL in another window you will see that it works, correct? Is this a bug or a regression on Discourse?
Re: The gumball transform bug: itās really random, but itās also really bad. Is this something unique to me? or have you had reports of this happening before?
G