GNOME Bugzilla – Bug 355456
Selection outline randomly vanishes.
Last modified: 2006-09-12 18:03:23 UTC
Recently when I've been using GIMP, after a while (maybe only a few minutes.. it varies, I haven't figured out what triggers it yet) The selection outline disappears; any changes to the selection take effect but they are not shown. Toggling the visibility of the selection outline via the View menu has no visible effect. It seems to be per view. Creating a new view makes the selection outline show up on the new view (but still not on the old view.) Other pertinent information: * Usually, I just paint with the Ink tool or Pencil tool, adjusting the brush parameters a lot. Often when I use a selection tool, it's the first selection tool I've used on this image this session at all. * Marching ants speed is set to 500ms * I have the selection outline invisible much of the time * It has occurred on all types of images, with one layer or many layers. It seems independent of the image. * This has only ever happened when I originally had a blank selection. I've never had a non-empty selection 'frozen' like this. * My best guess at the cause of this is that the selection's GC is being cached and not invalidated appropriately (ie. the old empty selection is being drawn rather than the current non-empty selection)
It would be really helpful if you could give a reproduceable way to getting this to happen.
I already, have as far as I can; I just had a short session which ended in this bug, and all I used was the free select tool (a lot), edit->fill ( a lot) and the ink tool (a little). Also used QMask, but that's probably not significant, as I've encountered the problem mainly without the use of QMask. I can reproduce it fairly reliably currently, just by using those tools for a while. (took me a few minutes before it would occur here) Another note: File->Revert does nothing (ie the selection stays invisible.). So maybe it is a problem that belongs to the view, but is not propagated when creating new views (which sounds like a re/initialization mistake)
I'll add a pointer to bug #355456, which probably describes the same problem, but I'm not certain enough to mark either as a dup yet.
... What is the actual bug number?
Gah, I meant bug #355395 -- sorry.
*** This bug has been marked as a duplicate of 355395 ***