Blender files are really good, in 8 years I only see one corrupt file (today). And the idea of rename last save file to blend1 also help a lot. But a lot of time blender crashes, I destroy a mesh, I forgot to save the file (yeah, sometimes happens)… specially when you work with experimental branches. And Always I go to autosave files an try to recover the file that I had, but not always I can recover my work, by some reasons.
The last autosave is after the problem, so I have a savefile with same problem or with loss data
The quit file is overwrited by a empty scene
The “recover last session” is not coherent and sometimes give me a file that is not the last that I had.
So, I want to ask by a little features
Multiple autosave files for same file, to have the last two or three autosave
Quit file with the filename, to not overwrite with other blender session
I came to report exactly this. I just lost (thankfully only) 1 hour of work because there is missing DNA block in the file. I was very lucky I saved previously as I don’t actually save all that often because Blender is soooooooooo reliable when it comes to autosaving.
It’s amazing to have blend and blend1 but there should also be blend and blend1 for autosaves because I have it set to save every 2 minutes or something like that and I don’t feel it at all I only feel it once the file is corrupto ://
I have many autosaves of this file but not one of them is useful. The one that’s corrupt is highlighted.
If it could be added that would greatly ease the pain. Although it needs to be somehow communicated to the users so they know it’s actually there. Maybe if opening the autosave returns “Missing DNA Block” error then the blend1 is loaded.
@sergey The issue is that Blender crashes during the actual autosaving process so the file ends up corrupted.
The issue is that Blender crashes during the actual autosaving process so the file ends up corrupted.
This seems to be another deviation from regular .blend file writing which I wouldn’t personally expect. For regular .blend files we write to a temp file next to the destination, and swap file over once we know the save was successful. This allows to prevent data loss if code crashes or when you run out of disk space. Kind of expect same approach to be used for autosave files as well.
Technically, I can not think of anything what would make it hard to make autosave to have same logic for versions and temp file. But you’re rising a good point about good UI. Not sure auto-opening is good, but at least all available versions are to be presented IMO.