Very large datasets are very slow to bring up save dialog
Very large datasets are very slow to bring up any of the save dialogs. This approaches fatal for extremely large datasets. Examples where this issue shows up is Save State, Save Screenshot and Save Animation. On a large (but not huge) dataset that I have, the wait is about 45 to 90 seconds. This is with around 10,000 files. On larger datasets, my user reports enough time in a remote Linux terminal to start File/ Save Screenshot, leave his house, bicycle to work, open his Linux terminal and ParaView is still hanging. Within a few minutes ParaView becomes responsive again. This is with around 100,000 files.
- 5.12.0, Linux, remote server.
- Open dataset.
- File/ Save State or File/ Save Screenshot
The file dialog won't appear for up to half an hour for huge datasets.
What Cory and I think is happening is that the code that allows saving to the remote disk is doing an "ls" type operation. These are famously slow for large numbers of files. File/ Open uses different logic/commands and is very fast. We should be using that logic.
As this is fatal for huge datasets, I am marking this as showstopper for 5.13.0. Please talk to Cory or me before slipping.