On Apr 19, 2013, at 2:02 PM, Robin Sommer <[email protected]>
 wrote:

>> Provided that &default for tables is not supposed to modify table
>> membership when accessing indices that don't exist,
> 
> I'm wondering if it should modify the table here. I can see how the
> current behaviour is misleading, it violates the "principle of least
> surprise". :)

Does seem more intuitive to me.

> Can we change tables so that if &default is a non-constant, the first
> time one accesses a non-existing index, that slot gets assigned a
> deep-copy of the &default value?

Probably.

> The downside would be that if
> somebody is relying on the current behaviour, he might access lots of
> non-existing entries with the assumption that the table won't change
> (i.e., he won't run into memory trouble).

Maybe we can provide a script-layer flag that, when on, generates warnings for 
script locations that end up assigning &default values to non-existing indices? 
 That at least would help someone pinpoint locations they need to change to 
work w/ new behavior.

- Jon
_______________________________________________
bro-dev mailing list
[email protected]
http://mailman.icsi.berkeley.edu/mailman/listinfo/bro-dev

Reply via email to