See http://viewvc.slicer.org/viewvc.cgi/Slicer4?view=revision&revision=18326

It works fine, thanks !
There is a typo though. It says that it's the "ITK" directory that is too long, instead of "Slicer".

Typo fixed - see https://github.com/Slicer/Slicer/commit/93cbb032b228e2113fa68b6359bab3acd560fb09

ITKv4 limit is 50

ITKv4 being a dependency of Slicer, I guess it should even be smaller. [for example 40 = 50 - length(/ITK-build)]

Ideally, we could scan a successful build of windows, and find out what's the longest path. From there we would be able to do a more educated guess.

Let's continue the discussion on that mantis issue: http://na-mic.org/Mantis/view.php?id=1480

Thanks
Jc

• Hide quoted text -

On Fri, Oct 14, 2011 at 3:39 PM, Steve Pieper <pieper@ibility.net> wrote:

Jc -

It looks like the windows build failed due to the new check on the directory depth - that bit me on my local builds too.  Did the # of 50 come from an empirical test or is it just a guess?  Could we get by with a number more like 75 or 100?  My path is 67 and it always worked until the new tests were committed.

-Steve

Sometime, the length test can pass for the superbuild but fail for the inner build. Unfortunately, it is not caught at configure time.
Can you test superbuild is not 11 chars too long (sizeof("Slicer-Build"))?

The option is not passed to the inner build neither.

See http://viewvc.slicer.org/viewvc.cgi/Slicer4?view=revision&amp;revision=18466