Closed mtbc closed 5 years ago
Code wise, everything looks okay. Is there a pre-built omero server somewhere that I can download to test this, or should I build everything locally?
(though at some point merge-ci
might appear!)
Files were deleted from binary repo. I will do further testing again today
Imported some Cellworx plates
The files are correctly deleted, the log file associated to the directory containing the data e.g. 09-59-43.839.log
is also removed.
Currently this leaves an empty directory e.g. 09-59-43.839
.
From discussion with @mtbc, this is currently expected
Assuming the above means, "good to merge" and merging. If a correction is needed, we can get a follow-up in for 5.5.0.
https://github.com/openmicroscopy/openmicroscopy/pull/3751 and https://github.com/ome/omero-server/pull/14 carefully order file deletions. The need for that latest PR was detected by the
_fs_dir_delete
database trigger however the need for this PR is probably being concealed by [trello] repo file deletion fails silently. This PR aims to ensure that the deletions from the underlying filesystem occur in the same order as from the database. This may become more of an issue when data deletion more commonly includes directories due to https://github.com/openmicroscopy/management_tools/pull/921 and its like which the extra-keen may wish to use for thorough testing.To test, try deleting complex data like SPW and watch for regressions in what is deleted from the binary repository. In normal OMERO usage thumbnails, image files, etc. should be deleted exactly as before.