Conversation
isaacbrodsky
force-pushed
the
h3pg-56-test
branch
from
7a6f5ec to
0953747
Compare
|
Not sure I'm following the goal of this PR - the test behavior seems correct. Is the idea that we'll have some strategy to identify cases like this in the future and allocate less memory? |
Yes, this is to help test those cases in the C library, and also ensure we don't regress to requesting more memory in cases like this. Aside: We may want to consider partial polyfill approaches that would allow us to provide a stream (or iterator) of results to the caller rather than requiring pre-allocating all the memory. |
|
Note also that the Win32 test for this actually fails with a segfault, so this does currently fail in some configurations. |
|
Isaac Brodsky seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account. You have signed the CLA already but the status is still pending? Let us recheck it. |
4 participants
Adds a test driver for postgis/h3-pg#56 to the core library for debugging this type of polygon that currently has a very high estimation.