Replacing the class in PCL is one option, another option is that it may be possible to override just the Start() method by registering an observer to StartEvent.<br><br>Radu, since PCL builds against VTK 5.x, PCL will need a override/workaround for some time into the future I suppose.<br>
<br>Pat<br><br><div class="gmail_quote">On Tue, Aug 7, 2012 at 1:58 PM, Radu B. Rusu <span dir="ltr"><<a href="mailto:radu.b.rusu@gmail.com" target="_blank">radu.b.rusu@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
That's fantastic, thanks to everyone for their work!<br>
<br>
Pat and I discussed that as a workaround until VTK updates would be to bring in a copy of the new Cocoa interactor class in PCL and #ifdef it for Mac, so that we can fix the outstanding issues we have at the moment in PCL. Once VTK updates, we can remove the class (it will be internal anyway).<br>
<br>
Cheers,<br>
Radu.<br>
--<br>
<a href="http://openperception.org" target="_blank">http://openperception.org</a><div class="HOEnZb"><div class="h5"><br>
<br>
On 08/07/2012 10:54 AM, Marcus D. Hanwell wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On Tue, Aug 7, 2012 at 1:50 PM, Sean McBride <<a href="mailto:sean@rogue-research.com" target="_blank">sean@rogue-research.com</a>> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On Fri, 3 Aug 2012 19:11:55 -0400, Pat Marion said:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Anyway, I appreciate you taking the time to look over the patch details!<br>
If we decide to proceed with the patch, you said some more changes might be<br>
needed, any hints what those might be?<br>
</blockquote>
<br>
Sorry for not elaborating, I was just leaving work when I wrote that. And now I'm off on vacation for two weeks. :(<br>
<br>
I don't object to you committing this as is for now. When I get back, I can expand on your change. Basically, from what you're saying we don't need any of that code that terminates the process, ever. Also, we should review when we register/unregister for window closing notifications. Currently the start and stop methods are not very symmetric, IIRC.<br>
<br>
If you do commit, I suggest you update the doxy comments for TerminateApp() at the same time.<br>
<br>
</blockquote>
I think ideally we should add new API to more cleanly accomplish a<br>
process events call, but as the other implementations all behave in a<br>
certain way consistency seems reasonable. I talked with Radu in the<br>
past about what they were doing, and how it was not necessarily the<br>
intended behavior of those functions (but it did work for them).<br>
<br>
If people are happy with the change, it seems like this will help out<br>
a set of VTK users without impacting other projects.<br>
<br>
Marcus<br>
<br>
</blockquote>
</div></div></blockquote></div><br>