Page MenuHome

"Batch-Generate Previews" crashes
Closed, ResolvedPublic


Operating system: Windows 10 Home x64
Graphics card: GTX 970M

Blender Version
Broken: 2.80.69, 06c4139a6833 (2019-5-19)

Short description of error
The operator to batch generate previews is broken. It ends up with a console window with the error, but it can't be closed and the ui window is completely normal. It also says "Blender quit."

Exact steps for others to reproduce the error

  1. Save the default file.
  2. File -> Data Previews -> Batch-Generate Previews -> point to the newly-saved file.
  3. The console comes up.

Event Timeline

Philipp Oeser (lichtwerk) claimed this task.
Philipp Oeser (lichtwerk) triaged this task as Confirmed, Medium priority.

Yep, this was left out from rB77f92b8cb73c. (still wanted to check something).

Brecht Van Lommel (brecht) raised the priority of this task from Confirmed, Medium to Confirmed, High.May 21 2019, 4:21 PM

(Version: 2.8 Release Candidate 1)
This issue is still present in the form of either another command prompt message, or it doesn't say anything and no previews are generated when viewing them when appending.

Some files, like the default one, shows this message and writes the previews, but most don't write when that comes up.

Philipp Oeser (lichtwerk) lowered the priority of this task from Confirmed, High to Needs Information from User.Jul 14 2019, 9:54 PM

I noticed a memleak in D4917 as well, this I will track down [this should not be the reason for not writing the previews though...and is probably not that high priority atm...].

However, I cannot reproduce the issue of "it doesn't say anything and no previews are generated when viewing them when appending":
I ran this on multiple files and the previews seem to show:

  • Could you provide a file where generating the previews fails?
  • Do you possibly get a message like *NOT* Saving someblendfile.blend, because some error(s) happened while deleting temp render data... (or any other ERROR message)?

That message is what I was talking about at the end of the second comment. Here is that message with wasp_bot.blend from 2.8's page on

I also can't quite remember what I did to get it to fail and not say anything.

Philipp Oeser (lichtwerk) raised the priority of this task from Needs Information from User to Confirmed, High.Jul 14 2019, 10:51 PM

OK, get it, too, (on the race_spaceship.blend) if I remove try/except, it fails here:

Traceback (most recent call last):
  File "/build_linux/bin/2.80/scripts/modules/bl_previews_utils/", line 511, in <module>
  File "/build_linux/bin/2.80/scripts/modules/bl_previews_utils/", line 502, in main
  File "/build_linux/bin/2.80/scripts/modules/bl_previews_utils/", line 425, in do_previews
    if not render_context_delete(render_context):
  File "/build_linux/bin/2.80/scripts/modules/bl_previews_utils/", line 167, in render_context_delete
    rna_backup_restore(scene, render_context.backup_scene)
  File "/build_linux/bin/2.80/scripts/modules/bl_previews_utils/", line 68, in rna_backup_restore
    setattr(dt, path[-1], val)
TypeError: bpy_struct: item.attr = val: enum "" not found in ('ADAPTIVE')
Error: Not freed memory blocks: 2413, total unfreed memory 1.000534

Here it seems to be about UnitSettings.length_unit [this might be more general though and affect other properties which have changed?]
Checking further...

I think this is due to settings with bad/missing do_versions code?
Those values are empty in the failing files and then this empty value refuses to be set (e.g. on an enum)

seems like this fails for (again, I've disabled try/exept in rna_backup_restore):

  • UnitSettings (length_unit, ...), introduced in rB2d21eb79ad48 ERROR: bpy_struct: item.attr = val: enum "" not found in ('ADAPTIVE')
  • FFmpegSettings (ffmpeg_preset, ...), introduced in rB7e39d151d850, fails with ERROR: bpy_struct: item.attr = val: enum "" not found in ('BEST', 'GOOD', 'REALTIME'), not sure about FFM_PRESET_NONE in code...

@Bastien Montagne (mont29): do you have a quick solution to this? Just skip them in exclude_props (D5255)?

@Grayson Riffe (MrGrapeDave) next time please open a new task, that new issue you are reporting here has absolutely nothing in common with the original one (except for the fact that they affect the same end-user tool), having different issues in a single task is really not ideal to manage the tracker.