You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently files when they are registered with an automatically generated name have the version appended to the end of the file name.
This in theory messes up the filename suffix, if an application was to use that.
Could:
Put dataset files also into directories first with the auto generated name and version, then put the file into that directory
Name the directories based on the executions and have the files go into their execution folders (however as you can append datasets to executions after, theres no guarantee of filename conflicts later)
The text was updated successfully, but these errors were encountered:
I'd like to also make a somewhat related change to ensure that automatically-generated paths don't clash with user-supplied paths: we make a subdirectory, say called .gen_paths, under the owner level, and all generated names go under there. We would also need to disallow any user paths starting with that string.
I'd like to also make a somewhat related change to ensure that automatically-generated paths don't clash with user-supplied paths: we make a subdirectory, say called .gen_paths, under the owner level, and all generated names go under there. We would also need to disallow any user paths starting with that string.
I have added this to the PR. Now auto generated relative paths start with a .gen_paths directory. There is also a check making sure users cannot start their relative paths with that directory
Currently files when they are registered with an automatically generated name have the version appended to the end of the file name.
This in theory messes up the filename suffix, if an application was to use that.
Could:
The text was updated successfully, but these errors were encountered: