Re: [cmake-developers] Export files in the build tree have a total path length which is unnecessarily long

2016-08-09 Thread Alan W. Irwin
On 2016-08-09 10:11-0400 Brad King wrote: On 08/08/2016 02:54 PM, Alan W. Irwin wrote: For all such staging areas, would it be straightforward to hash the paths for the cmake-generated EXPORT files in the build tree to substantially reduce their pathlengths for the absolute install location

Re: [cmake-developers] Export files in the build tree have a total path length which is unnecessarily long

2016-08-09 Thread Brad King
On 08/08/2016 02:54 PM, Alan W. Irwin wrote: > For all such staging areas, would it be straightforward to hash the > paths for the cmake-generated EXPORT files in the build tree to > substantially reduce their pathlengths for the absolute install > location case while still avoiding name

Re: [cmake-developers] Export files in the build tree have a total path length which is unnecessarily long

2016-08-08 Thread Alan W. Irwin
On 2016-08-08 13:40-0400 Brad King wrote: On 08/07/2016 01:39 AM, Alan W. Irwin wrote: why is the install location of these export files included in the pathname for both export files? That significantly increases the pathlength of these files to no purpose since the install location of these

Re: [cmake-developers] Export files in the build tree have a total path length which is unnecessarily long

2016-08-08 Thread Brad King
On 08/07/2016 01:39 AM, Alan W. Irwin wrote: > why is the install location of these export files included in the > pathname for both export files? > That significantly increases the pathlength of these files to no purpose > since the > install location of these files is already known to CMake by

[cmake-developers] Export files in the build tree have a total path length which is unnecessarily long

2016-08-06 Thread Alan W. Irwin
I have a simple project called test_fortran which somone else is testing for me on MinGW-w64/MSYS2, and it failed because the following file