Segmentation fault with dreich algorithm

[jmdelvecchio]

Perhaps related to #16 but I'm running on my institution's cluster so no sudo for me.

Installed from conda forge fresh this month on Linux cluster. Wiener and Pelletier algorithms perform as expected, but dreich fails with segmentation fault after printing number of source nodes. Ran it on 128GB and still had segmentation fault. Ran it on example .bils instead of my own data and problem persists.

Any advice in light of #16 which may or may not be my problem too?

[simon-m-mudd]

Hi Joanmarie. This is a bug in the code, not a compatibility issue. I've located the offending lines but to fix it there will need to be a new release. Which might not happen all that soon.
Our testing suggests the most reliable method is the Wiener method, so in the meantime I would just extract the networks using that method.

[jmdelvecchio]

Good to know, thanks!

[jmdelvecchio]

@simon-m-mudd actually now getting a segmentation fault with the hilltop flow routing too. Works if I use the legacy routine but I get a segmentation fault after it prints a ymax when using "the new and refactored version of the routing code." Independent of channel head type, still giving it 128GB to crank. Advice? Open a new issue?

[simon-m-mudd]

I will test. Might take a few days.
Do you have a small DEM and a driver file that has the issue that you can send (via email)?