http://i.imgur.com/9bomRPA.jpg

This occurs on fills of any colour. Multiple fills ontop of each other will create that pattern of random colours, otherwise its mostly the fill colour, with that pattern in some other colour. Red fill gives that pattern in black, Black fill gives that pattern in White, so on so forth.

Running Windows 7 64bit
GTX760 for graphics card if thats relevant...

EDIT: That squiggle about the third of the way up was a random brush stroke, in red. I filled around it and thats whats left of the brush stroke.

Hi,

I think I need some more information:

* os version and whether it's 64 or 32 bits and whether your version of krita is 64 or 32 bits
* the git number in the about box
* size of the image
* size of your system's memory

I've tried to reproduce on 64 bits linux, but didn't manage yet.

Os is windows 64, Krita is x64 as well.
Git is: e2f65ad

It was an A4 page

I have 16 gigs of ram.

I have exactly the same problem on Windows 8.1 Pro 64 Bit, 16 Gigs RAM, NVIDIA GTX 645 with 3 Gigs RAM

same git number... (occurs with all project sizes I tried)

thanks for fixing this in advance! Krita is an awesome program!

marco

I wish I could reproduce it... What CPU do you have? Intel or AMD?

Intel (NVIDIA GTX 645 as I already wrote )

thanks
Marco

I've just tried the latest build (Version 2.8 Beta 3 (git d0d3125)) and the bug is still present...

Also the whole app crashes completely and immediately when trying to paint with several "hatch" brushes (e.g. hatch_crispy) (But that was also already the same with the version I had installed before...)

Marco

Okay, Could you follow the instructions at viewtopic.php?f=281&t=119719 and see whether you can actually reproduce the crash with the debug build and get a stacktrace for me?

If you want, we can use teamviewer so I can "look over your shoulder" to see what happens. I really want to get to the bottom of these crashes...

Hey,
just wanted to let you know that version 2.8 solved the fill problem and the crashes with certain brushes I had before! Great stuff! Congrats to the new release! Rock solid so far

Marco

Ha, yes, we figured out what was going wrong... Basically, a bug in MSVC 2012 caused our AVX2 optimizations to go haywire, so we had to remove those.