As noted by some of you managed to get it done.
Most problematic thing seems to be "Texas" geometry which causes
problems in Oracle and sometimes with SpatiaLite as well.
I'll try to investigate if that data could be improved to create more
consistent test results.
On 10.11.2015 21:23, Tim
Well I fix those algorithm based errors, and do deeper analysis for
those yet unknown ones. If I can't figure out why that happens I create
ticket for each and mark them expected failures.
On 10.11.2015 21:23, Tim Graham wrote:
The changes to get it running were merged in September. Absent a
p
The changes to get it running were merged in September. Absent a proper
solution for those 6 failures, I agree it would be a good idea to mark them
expectedFailure so we can get the build running on CI before 1.9 final.
On Tuesday, November 10, 2015 at 1:54:42 PM UTC-5, Marc Tamlyn wrote:
>
> Th
Thanks for getting it running at all!
It looks to me like those tests just need to be less accurate for Oracle,
good plan.
Did you have any specific changes to make to get it to run at all? If so we
should include those soon, irrespective of whether the fixes for these
other tests make it into 1.
Hi, I finally had time to get back on Oracle GIS issues.
I ran test suite against 1.9.x and 7 tests fails.
3 tests are most probably due different algorithms used to calculate
geographical distance and areas. My proposal for fix is to use backend
spesific values.
Failing testcases are:
=
