You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Here are some steps to recreate a bug in smokeview:
(1) Run the attached FDS model in FDS6.7.9. (This is tunnel0003.fds, renamed as tunnel0003.txt, but I have seen this bug in several models.)
(2) Open the results in the corresponding SMV6.7.21. Rotate the view using the left mouse button. Zoom out by holding down SHIFT and dragging the left mouse button downwards -- I'm not sure, but I think it is critical that this zoom method is used (rather than, for example, dragging the middle mouse button). Try to reach a view with a memorable zoom level -- for example, all of the model is just in view. Press ALT-g to raise the Motion/View/Render dialog. Name the view (for example "view1" in the Edit: line), press "Add", press "Apply at startup", press "Save settings", press "Close". Right-click on the smokeview window and quit.
(3) Open the results again in smokeview. There are two problems: First, smokeview takes longer than usual to load (maybe 20 seconds -- possibly less if you nudge the system somehow, for example pressing RETURN in the bash window or changing something in the X display). This time is not reported in the "Startup time" message to stdout; the delay appears to happen before the "reading tunnel0003.smv" message. (Actually, the messages to screen sometimes seem to not update themselves without a nudge, but this is an intermittent fault.) Second, the view that is applied is an incorrect, zoomed-in version of the saved view.
It is the second problem that bothers me more. I think this bug applies not only to the "Apply at startup" view -- various views can be saved incorrectly. During step (2), in the first session of smokeview, the views are reconstructed correctly by selecting or by the "Cycle User" button (if you expand the steps to include several saved views). The views seem to be held correctly in memory but not reconstructed correctly via the saved .ini file. It (sometimes) applies to both perspective and orthographic views.
I do not know if this behaviour is new or old -- I only recently found the SHIFT+leftmouse method of zooming, and I find it more convenient than holding the middle mouse button.
By the way, there is another odd quirk of saving views. Often, when I press "Add" in the Viewpoints part of that dialog, the displayed view jumps to a zoomed-out version of the added view. This is slightly inconvenient, but I can just re-select the view that I have just added. The view has been added as intended. This quirk has been around for many versions, but it feels possibly relevant here. It is possible that a view reached in step (2) with SHIFT-leftmouse does not show the quirk -- the whole business is slightly difficult to replicate. I can live with the jumping-out quirk, but having a view saved incorrectly to .ini means that I produce useless images. As I mentioned, I think one workaround might be to use middle-button zooming.
The text was updated successfully, but these errors were encountered:
Here are some steps to recreate a bug in smokeview:
(1) Run the attached FDS model in FDS6.7.9. (This is tunnel0003.fds, renamed as tunnel0003.txt, but I have seen this bug in several models.)
(2) Open the results in the corresponding SMV6.7.21. Rotate the view using the left mouse button. Zoom out by holding down SHIFT and dragging the left mouse button downwards -- I'm not sure, but I think it is critical that this zoom method is used (rather than, for example, dragging the middle mouse button). Try to reach a view with a memorable zoom level -- for example, all of the model is just in view. Press ALT-g to raise the Motion/View/Render dialog. Name the view (for example "view1" in the Edit: line), press "Add", press "Apply at startup", press "Save settings", press "Close". Right-click on the smokeview window and quit.
(3) Open the results again in smokeview. There are two problems: First, smokeview takes longer than usual to load (maybe 20 seconds -- possibly less if you nudge the system somehow, for example pressing RETURN in the bash window or changing something in the X display). This time is not reported in the "Startup time" message to stdout; the delay appears to happen before the "reading tunnel0003.smv" message. (Actually, the messages to screen sometimes seem to not update themselves without a nudge, but this is an intermittent fault.) Second, the view that is applied is an incorrect, zoomed-in version of the saved view.
It is the second problem that bothers me more. I think this bug applies not only to the "Apply at startup" view -- various views can be saved incorrectly. During step (2), in the first session of smokeview, the views are reconstructed correctly by selecting or by the "Cycle User" button (if you expand the steps to include several saved views). The views seem to be held correctly in memory but not reconstructed correctly via the saved .ini file. It (sometimes) applies to both perspective and orthographic views.
I do not know if this behaviour is new or old -- I only recently found the SHIFT+leftmouse method of zooming, and I find it more convenient than holding the middle mouse button.
By the way, there is another odd quirk of saving views. Often, when I press "Add" in the Viewpoints part of that dialog, the displayed view jumps to a zoomed-out version of the added view. This is slightly inconvenient, but I can just re-select the view that I have just added. The view has been added as intended. This quirk has been around for many versions, but it feels possibly relevant here. It is possible that a view reached in step (2) with SHIFT-leftmouse does not show the quirk -- the whole business is slightly difficult to replicate. I can live with the jumping-out quirk, but having a view saved incorrectly to .ini means that I produce useless images. As I mentioned, I think one workaround might be to use middle-button zooming.
The text was updated successfully, but these errors were encountered: