Michael Hudson wrote: > Me neither. Are you sure you regenerated cookielib? Oops, no. Or, rather, yes - even the first number is already regenerated. > What would it cost to check if all strings could be stored via > TYPE_STRINGREF? (Almost nothing in code terms...) It would cause additional dictionary lookups on marshalling, which in turn would cause hash computation for all strings. Currently, as only interned strings are marshalled through TYPE_STRINGREF, we know that hash values are already computed for all of them, so we only pay a dict lookup per string (and perhaps an allocation if the string is new and the dict needs resizing). In addition, there is a slight semantical change, in that equal strings become shared on unmarshalling. I don't think this is relevant, though, since marshal never worried about shared strings in the first place. > I'd be happy with any other (I mentioned not doing > refcounting on dictobject.c's dummy; maybe I should actually see if > that's possible :-). Feel free to change the marshalling to try sharing all strings; it might be easier to revert the change to compile.c, though. Regards, Martin
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4