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
The file overwrite policy should be clear in the usage.
Also, any errors encountered should be very explicit, and even non-errors. Particularly, these should be called out at the parfu level so that they do not generate MPI_File_Open errors.
create mode, in non-overwrite-archive mode, user specifies archive file that already exists
create mode, in overwrite mode, user specifies file that exists, code should confirm
extract mode, in non-overwrite mode, some target file exists, code should complain before extracting ANY files.
extract mode, in overwrite mode, code should confirm there are files that will be overwritten.
The text was updated successfully, but these errors were encountered:
Now looking at this during the 0.6 build-up:
I think the extract is clear. If the extract is going to put a file on disk, but there's already a file there, there's a question what to do.
But for the create case, what's the worry? parfu isn't designed to add to archives, it just creates a single archive. It's not supposed to do progressive backups; that's not its schtick. However, I suppose with weird linking, it could get in a situation where a file was going to be archived twice, which I supposed should be flagged or mentioned because it's going to create an ambiguity upon extraction.
The file overwrite policy should be clear in the usage.
Also, any errors encountered should be very explicit, and even non-errors. Particularly, these should be called out at the parfu level so that they do not generate MPI_File_Open errors.
The text was updated successfully, but these errors were encountered: