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
This was motivated by a cache being written using "" for the name and breaking tools further down the line. We should also consider disallowing "/" in a name since we use it as a path separator. Need to consider carefully at which level of IO we do this - perhaps IndexedIO should be left alone?
The text was updated successfully, but these errors were encountered:
I'm just going to play devil's advocate here, and suggest an alternative approach. In a vector<InternedString> there is no ambiguity at all with any choice of name (even dumb names like "" and "/"), so why should we restrict it? The other option is to allow "/" to be escaped in the string->vector<InternedString> conversion, to disambiguate. For instance :
/a/b/c -> [ a, b, c ]
/a\/b/c -> [ a, /b, c ]
There's a related discussion on GafferHQ/gaffer#2070, where we're considering restricting even further to just alphanumerics. I think that's probably too much, so I'm proposing the above as an example of how we could remain totally general if we really wanted to.
Just so I have mentioned this: The solution to use just alphanumerics that I have brought up before actually wouldn't work for us because we have nodes that create objects in an array-like format:
This was motivated by a cache being written using "" for the name and breaking tools further down the line. We should also consider disallowing "/" in a name since we use it as a path separator. Need to consider carefully at which level of IO we do this - perhaps IndexedIO should be left alone?
The text was updated successfully, but these errors were encountered: