Opened 4 years ago
Last modified 3 years ago
#6175 assigned defect
stateSelect regression from v1.14.1 to v1.16.0
| Reported by: | Owned by: | Karim Adbdelhak | |
|---|---|---|---|
| Priority: | blocker | Milestone: | 1.19.0 |
| Component: | Backend | Version: | |
| Keywords: | Cc: |
Description
The attached model of the 6DOF Rotary Stewart Platform (a parallel mechanism) works with v1.14.1 (64-bit) on Win 10 but doesn't work with the current stable (v1.16.0).
Attachments (1)
Change History (9)
comment:1 by , 4 years ago
| Milestone: | Future → 1.16.1 |
|---|---|
| Priority: | high → blocker |
by , 4 years ago
| Attachment: | modelStatikaParametarski.mo added |
|---|
comment:2 by , 4 years ago
I added the missing 'each' prefixes to the attachement so you can use the new front end. BTW, the new front end was already the default in 1.14.1, I guess you switched to the old one to run your model.
I understand the handling of 'each' with the GUI is still incomplete, this is know issue, see #5405, I hope we can fix it in the next release.
comment:3 by , 4 years ago
| Component: | *unknown* → Backend |
|---|---|
| Owner: | changed from to |
| Status: | new → assigned |
I see the following state choice:
Number of states: 24 ($STATESET2.x[3],$STATESET2.x[2],$STATESET2.x[1],$STATESET1.x[6],$STATESET1.x[5], $STATESET1.x[4],$STATESET1.x[3],$STATESET1.x[2],$STATESET1.x[1], B1.phi,B1.w,B2.phi,B2.w,B3.phi,B3.w,B4.phi,B4.w,B5.phi,B5.w,B6.phi,B6.w, A1P1.body.phi[1],A1P1.body.phi[2],A1P1.body.phi[3])
I'm not too familiar with the Stewart platform modelling, but if this system has 6 d.o.f. it should have 12 states, not 24.
Can you please run this in 1.14.1 with additional translation flag -d=stateselection and report what state choice you see?
comment:4 by , 4 years ago
Thank you for all responses, I already inserted each but this was not what I have reported as a bug. This model stopped working between releases, which means that I get results after running this model (whole animation) in v1.14.1 but in v1.16.0 simulation is aborted without results.
comment:5 by , 4 years ago
| Milestone: | 1.16.1 → 1.17.0 |
|---|
@Milos, we need to proceed with the release of 1.16.1, because it has some critical bugfixes. We understand this is a regression, but I'm afraid we'll have to fix it in 1.17.0. Once that is done, you may use the nightly build to access the fix. If you really need this thing to work in an official release, we may do a 1.16.2 bugfix release.
If you could please run this in 1.14.1 and report the state choice, that would help figuring out what went wrong.
Is the error related to unit checking? The unit checking in the old front end experiences some problems at the moment, so you may want to try the newInst.
If you do that you need to insert about a dozen missing
eachkeywords in the model. In the old front end those were warnings but in the new front end they are errors.