[vtk-developers] 22 new failing GPU tests on OpenGL2 dashboard

[David Cole]

Thanks very much -- we'll find out tonight when the next nightly
dashboard runs on "mug.neocisinc"

I'll ping again if there's anything further.


D


On Tue, Feb 3, 2015 at 3:40 PM, Aashish Chaudhary
<aashish.chaudhary at kitware.com> wrote:
> David,
>
> Thanks to Ken, we were able to reproduce the problem when Ken switched to
> use Intel driver on his Windows machine. We worked on
> a fix that possibly fix issues you were seeing (it worked ok Ken's machine).
> The good thing is that it became ~10 times faster  on Windows machine.
>
> I will push a fix momentarily and hopefully it will get merged tonight.
>
> - Aashish
>
>
>
>
> On Thu, Jan 29, 2015 at 4:01 PM, Ben Boeckel <ben.boeckel at kitware.com>
> wrote:
>>
>> On Thu, Jan 29, 2015 at 14:48:00 -0500, Ken Martin wrote:
>> > " I must admit, I'm a little bit disappointed that some of the commits
>> > in what was merged to VTK master don't even compile... I would hope we
>> > could be a bit less sloppy, and hold ourselves to a higher standard.
>> > Rushing something in before it's ready for prime time serves no one."
>> >
>> > I want to make sure I understand what you are saying. Are you saying in
>> > this
>> > case code was merged into master that did not compile? Or are you saying
>> > that someone did multiple commits, some of which did not compile, but
>> > they
>> > were not pushed or merged until it did compile?
>>
>> FWIW, I have a script I use to ensure this for long-lived branches:
>>
>> =======
>> #!/bin/sh
>>
>> base="${1:-master}"
>> shift
>>
>> GIT_EDITOR="sed -i -e '/^pick/p;/^pick/cexec ./build.sh'"
>> export GIT_EDITOR
>>
>> exec git rebase -i "$( git merge-base HEAD "$base" )"
>> =======
>>
>> A 'build.sh' script (provided by you) is run after each commit and stops
>> on any commit which fails to build.
>>
>> --Ben
>
>
>
>
> --
> | Aashish Chaudhary
> | Technical Leader
> | Kitware Inc.
> | http://www.kitware.com/company/team/chaudhary.html