Help with testing - part 5

[PrzemoF]

Small changes, but mostly visible to the user - please test carefully. There is a new button (Purge FEM results), multiple results problem is now solved and results icon is active only if results are present.

FEM: Clean ccxFrdReader
FEM: Disable Show result button if there are no results in analysis
FEM: Add Purge FEM results button
FEM: Purge exisitng results before loading new
FEM: Always import FemGui

https://github.com/PrzemoF/FreeCAD_sf_m ... /fem_devel

This is my devel branch, so resetting/force upgrading _will_happen.

[bernd]

works great on my debian jessie. Mhh sometimes I really like to have the results of more runs availeble. Would it be possible to make an preference for this. The user could decide if he would like to have the exisitng results purged before loading the new ones than. Especially with the new button all existing results could be deleted just with one click.

Checkbox in FEM prefs: Purge exisitng results before loading new: on/off.

What do other say on this?

[PrzemoF]

Thanks for testing! That option would be easy to add. To be honest I don't like the way we're presenting the results al all: 2 separate objects in the tree (Displacement & MisesStress), then all together in the results dialog with names (U1, U2 ...) not linked in any way with the results names. After a re-run there is just mess in the dialog window with the results.

What about:
1. Switching off purging after re-run
2. Leaving Purge button as it is now
3. Changing names in the results dialog to show "result name - type" i.e. : "Displacement001 - Uabs (Disp. abs)" or "MisesStress002 - Von Mises Stress" something similar? It will require some renaming to avoid duplicating the same info in a name.
4. Renaming MisesStress to Stress, so we'd have Displacement & Stress in the tree.

I also want to add a tick box below the 3 buttons in with default value controlled from Preferences to allow "Write & run" - most of the time I don't do any .inp file editing [1] , but I have to click twice to run analysis. Another thing is that there is a disabled path selector [2] above the 3 buttons showing "/tmp" - we should remove it or enable and use it.

[1][2] - see screenshot

P.S. Do you know why there is a warning when trying to delete result by hitting "Del" key:
"This object is referenced by other objects and thus these objects might get broken. Are you sure to continue?"
I want to be sure that purging results is safe.

[bernd]

All the proposals you made sound good to me.

...
P.S. Do you know why there is a warning when trying to delete result by hitting "Del" key:
"This object is referenced by other objects and thus these objects might get broken. Are you sure to continue?"
I want to be sure that purging results is safe.

no. May be it is because they are inside the MechanicalAnalysisContainer.

[sgrogan]

The warning makes sense if you look at the dependency graph. But the dependency graph looks wrong.

dg.png (56.83 KiB) Viewed 2789 times

[PrzemoF]

The warning makes sense if you look at the dependency graph. But the dependency graph looks wrong.

dg.png

OK, thanks! I'm going to ignore the problem for now. I have to learn more about dependencies in FreeCAD to be able to fix it.

[PrzemoF]

The warning makes sense if you look at the dependency graph. But the dependency graph looks wrong.

dg.png

OK, thanks! I'm going to ignore the problem for now. I have to learn more about dependencies in FreeCAD to be able to fix it.

Bug report: http://www.freecadweb.org/tracker/view.php?id=2080

[PrzemoF]

works great on my debian jessie. Mhh sometimes I really like to have the results of more runs availeble. [..]

Can you double check that you actually have 2 different results? It doesn't look like that from the code and I don't see it in my results. There are to positions, but one is just a place holder - all values are from the latest run.

[sgrogan]

@PrzemoF

I can confirm the results are all duplicates of the last run.
I think this is how it should be. There only one set of results for a given analysis. If you change some parameters it should be a new analysis, eg Mechanical Analysis001, with its results.

[PrzemoF]

@PrzemoF

I can confirm the results are all duplicates of the last run.
I think this is how it should be. There only one set of results for a given analysis. If you change some parameters it should be a new analysis, eg Mechanical Analysis001, with its results.

Thanks! I agree with you. I just coded handling of multiple results anyway, because what we have right now is plain wrong.
We might remove it when proper handling of multiple analysis will be implemented.