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
If settings like encrypted_regex, unencrypted_regex, encrypted_suffix, unencrypted_suffix, encrypted_comment_regex, unencrypted_comment_regex are used, files read with the binary store often end up unencrypted. This is likely the cause for the observations in #1818 and #1805.
I think we should either:
simply ignore these settings when encrypting a file that was read with the binary store;
or maybe warn in that case, with either encrypting with or without these settings;
or fail encryption if these settings prevent encryption of the data key.
What do you think @getsops/maintainers? Simply acting as now and not encrypting the data key (and thus not encrypting anything!) is surely a bad idea.
This could also be generalized to encryption in general: if you try to encrypt a file and nothing is actually encrypted, print a warning. This might be intentional in some special cases, but more often is likely a mistake. Maybe allow to disable the warning via a command line option for the cases where it is intentional and a warning is not wanted by the user.
The text was updated successfully, but these errors were encountered:
If settings like
encrypted_regex
,unencrypted_regex
,encrypted_suffix
,unencrypted_suffix
,encrypted_comment_regex
,unencrypted_comment_regex
are used, files read with the binary store often end up unencrypted. This is likely the cause for the observations in #1818 and #1805.I think we should either:
data
key.What do you think @getsops/maintainers? Simply acting as now and not encrypting the
data
key (and thus not encrypting anything!) is surely a bad idea.This could also be generalized to encryption in general: if you try to encrypt a file and nothing is actually encrypted, print a warning. This might be intentional in some special cases, but more often is likely a mistake. Maybe allow to disable the warning via a command line option for the cases where it is intentional and a warning is not wanted by the user.
The text was updated successfully, but these errors were encountered: