Re: insighttoolki4, (was gccxml is uploaded)
>>
>> Bumping to v4.6.0 is a good option -- freshly tagged. :-)
>
> Damn, you ;)
>
> The thing is, I can not upload 4.6 because it will introduce a new
> package by replacing libinsighttoolkit4.5 with libinsighttoolkit4.6,
> and hence, it has to go through the NEW pipeline (At least that's how I
> understand it). And as a DM I can only upload packages that don't go
> through NEW, which means for this upload we will need a DD (like
> Steve).
>
> Anyway, I think we will get 4.6 into jessie before the freeze.
>
Awesome!
>>
>> Otherwise, the hashes are:
>>
>> 9c54724 BUG: Remove division in inner loop for otsu threshold calculator
>> defdd99 BUG: Use tolerant float compare when choosing better otsu threshold
>> b3ca416 BUG: Increased OtsuThreshold computation precision
>> 76c4580 BUG: Increase itkVoronoiPartitioningImageFilterTest1 for 32-bit builds.
>>
>
> Is "b3ca416 BUG: Increased OtsuThreshold computation precision" really
> needed? I thought the first two patches were actually there to avoid the
> need for a higher tolerance?
For the Otsu test, 9c54724 will make the test pass. The other patches
can be added optionally for good measure.
> On the bright side the Debian i386 build wet through in a cowbuilder
> environment with patches 1,2, and 4, and now I'm starting to run the
> amd64 build just to make sure. If this goes well I should be able to do
> the upload by the end of the day.
Excellent!
Thanks,
Matt
Reply to: