Re: [Bf-committers] Node link swapping

2014-04-14 Thread gandalf3
I agree that this can be and is very annoying in some other cases, e.g. a material output node.. The only other nodes which could use this that immediately come to mind are math nodes. The wildcard is node groups. If someone makes a node group that has a mix node connected to the inputs, then i

Re: [Bf-committers] Node link swapping

2014-04-14 Thread Daniel Salazar - patazstudio.com
I'm wondering in what other nodes appart from the different mix nodes this functionality makes sense? Maybe it's not too hard to implement this in only where it makes sense On Apr 14, 2014 11:15 PM, "Greg Zaal" wrote: > I know this might not be a perfect solution, but a little while ago I added >

Re: [Bf-committers] Node link swapping

2014-04-14 Thread Greg Zaal
I know this might not be a perfect solution, but a little while ago I added something to the Node Wrangler addon that you might like - select your mix node or whatever and hit Alt+S. It'll swap the two colour inputs. On 15 Apr 2014 3:02 AM, "patrick boelens" wrote: > Aaand I should learn to read

Re: [Bf-committers] Node link swapping

2014-04-14 Thread patrick boelens
Aaand I should learn to read before commenting. =P Sorry about that! > Date: Mon, 14 Apr 2014 16:16:41 -0700 > From: zzyx...@gmail.com > To: bf-committers@blender.org > Subject: Re: [Bf-committers] Node link swapping > > Actually that was one of the ideas in my proposal :) > > On 04/14/14 16:13,

Re: [Bf-committers] Node link swapping

2014-04-14 Thread David McSween
Mostly I only have the need for socket swapping because I forget that in Blender -the bottom socket is the top layer and the top socket is the bottom layer of a mixed image :P But yeah the mix node or nodes with similar duplicated sockets are of issue, it's not ideal to just swap to the next avail

Re: [Bf-committers] Node link swapping

2014-04-14 Thread gandalf3
Actually that was one of the ideas in my proposal :) On 04/14/14 16:13, patrick boelens wrote: > Perhaps the swapping behaviour could be mapped to a key? For example, > plugging in a new image output into a Mix node makes the existing link > disappear, while doing so while holding shift could pe

Re: [Bf-committers] Node link swapping

2014-04-14 Thread Daniel Salazar - patazstudio.com
-1 On adding it back like it was before. It just made no sense most of the time. This needs to be done right Daniel Salazar patazstudio.com On Mon, Apr 14, 2014 at 5:14 PM, joe wrote: > I agree, but I think maybe we should bring it back until the new > implementation. Though it shouldn't be too

Re: [Bf-committers] Node link swapping

2014-04-14 Thread joe
I agree, but I think maybe we should bring it back until the new implementation. Though it shouldn't be too hard, just an extra flag per node. On Apr 14, 2014 3:56 PM, "Daniel Salazar - patazstudio.com" < zan...@gmail.com> wrote: > The old behavior was too broad and intrusive, maybe swapping could

Re: [Bf-committers] Node link swapping

2014-04-14 Thread patrick boelens
Perhaps the swapping behaviour could be mapped to a key? For example, plugging in a new image output into a Mix node makes the existing link disappear, while doing so while holding shift could perform the swap. Just some thoughts. Cheers! > From: zan...@gmail.com > Date: Mon, 14 Apr 2014 16:55

Re: [Bf-committers] Node link swapping

2014-04-14 Thread Daniel Salazar - patazstudio.com
The old behavior was too broad and intrusive, maybe swapping could be implemented just in the few cases where it actually makes sense like in mix nodes Daniel Salazar patazstudio.com On Mon, Apr 14, 2014 at 4:53 PM, gandalf3 wrote: > I have missed the node connection usability feature which was

[Bf-committers] Node link swapping

2014-04-14 Thread gandalf3
I have missed the node connection usability feature which was removed in r61178, and I'm convinced that my node workflow is slightly slower than it was before. I have written up a wiki page/request explaining why I think the old behavior was faster in most cases. You can find it here: http://wi

Re: [Bf-committers] Buildbot MSVC 2013

2014-04-14 Thread Martijn Berger
Hi Jeffrey, In cuda 5.5 and 6.0 there is a regression with texture memory handling that costs us significant performance I think. Cuda 5.0 requires msvc 2008 or 2010 to run but since we only use nvcc to compile a kernel and do dynamic runtime loading of libucda you can just use the 2008 install yo

Re: [Bf-committers] Buildbot MSVC 2013

2014-04-14 Thread Jeffrey H
Martijn, That would be fantastic if you could document how you configured your builders. Out of curiosity, what is the status of CUDA 5.5? I know we just updated to 5.0 to support newer cards and kernels, but I haven't seen any discussion on using 5.5 yet. I'm already using 5.0 as per Brecht's re