Just finished testing the new code with all the default
settings in the rc1 perldl.conf but forced the PROJ4
build for cygwin (which had been forced to 0 as a
work around). The result was that everything built and
t/gis_proj.t ran but t/proj_transform.t failed due to out
of bounds values.
I then enabled the bad values option and all PROJ4
tests passed. In fact, all tests were either passed
or skipped and the 4 skipped were because I haven't
gotten HDF/NDF built for cygwin yet.
I added a check for WITH_BADVAL to the proj_transform.t
test to skip running if bad values were not enabled which
seems to work for cygwin. Hope it works for others.
With luck this will fix the outstanding build/test problem
with the PROJ4 modules.
On a side note: I saw a number of warnings during
the proj_transform.t test for undef values in various TriD
module functions. These may be bad values from the
proj_transform.t test being passed on to TriD. I don't know
if TriD is supposed to be badvalue enabled although it
doesn't seem to die. This'll have to wait for the TriD
revisit TBD after 2.4.3...
Thanks for the good work!
Chris
Judd Taylor wrote:
> Ok, I've committed those changes into the rel_2_4_3rc1
> branch, and now all of the tests and make stuff is
> working (for me, at least), just like it is for the main
> branch. Please test away.
>
> -Judd
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
