- Apr 09, 2025
-
-
Alan De Smet authored
This eliminates the use of os.fork that could potentially cause a race with threads and definitely genrates a DeprecatinWarning in Python 3.12. How are we using threads? That's a mystery! Maybe multiprocessing.Queue?
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
-
- Apr 08, 2025
-
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
One of the tests relies on there being a file we cannot update the update time on, which doesn't work as root.
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
-
- Mar 17, 2025
-
-
Alan De Smet authored
Don't move empty files into place if a lock attempt raises an exception https://gitlab.ssec.wisc.edu/cspp_geo/cspp-geo-aitf/-/issues/479
-
- Mar 06, 2025
-
-
Alan De Smet authored
Attempted workaround for https://gitlab.ssec.wisc.edu/cspp_geo/cspp-geo-aitf/-/issues/479
-
- Mar 05, 2025
-
-
Alan De Smet authored
-
Alan De Smet authored
Fixes warnings in output of test-example
-
- Nov 08, 2024
-
-
Alan De Smet authored
-
Alan De Smet authored
-
- Oct 01, 2024
-
-
- Aug 05, 2024
-
-
Alan De Smet authored
Handles case where a user specifies a custom URL of something like http://foo.com/ancillary which would expand to something like http://foo.com/ancillary2024-08-05-file.data https://gitlab.ssec.wisc.edu/cspp_geo/cspp-geo-aitf/-/issues/468
-
- Aug 02, 2024
-
-
Alan De Smet authored
This reverts commit b63df565.
-
- Jul 23, 2024
-
-
Alan De Smet authored
Avoids using the version in ShellB3, which isn't what we want (And is busted as a bonus)
-
- Jul 12, 2024
-
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
See #13 for details (Also tracking ticket at https://gitlab.ssec.wisc.edu/cspp_geo/cspp-geo-aitf/-/issues/402 ) A few tests had to change because error output no longer include "File not found" in the output.
-
Alan De Smet authored
It was always broken, it just happened to work by chance. It inexplicably passed a DownloadStatistics into the field where the file object should be. We only got away with it because the failure happened before the file was accessed. However, that's no longer going to be true in a world where we hand the work off to curl.
-
Alan De Smet authored
-
- Apr 19, 2024
-
-
Alan De Smet authored
Was testing the wrong variable for None-ness. The returned file from atomic.__enter__() will be None if the file already exists, not the atomic itself.
-
- Feb 06, 2024
-
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
Slightly simplifies code, eliminating piles of calls to assertMockFile and assertNoMockFile. Also will detect spuriously created files I didn't think to explicitly test for with assertNoMockFile.
-
- Feb 05, 2024
-
-
Alan De Smet authored
Chains Downloaders, allowing different data sets to fulfill a requirement.
-
- Jan 24, 2024
-
-
Alan De Smet authored
-
- Jan 23, 2024
-
-
Alan De Smet authored
-
Alan De Smet authored
-
Alan De Smet authored
fatal is an undocumented alias for critical
-
Alan De Smet authored
Ensure some code examples are flagged as ``code``
-
Alan De Smet authored
An example from a previous parameter was erroneously duplicated into the next parameter
-
- Nov 17, 2023
-
-
Alan De Smet authored
Likely typo. We have no guarantee we can write in the parent of the temporal directory.
-