Closed oscbyspro closed 1 week ago
Additionally, the TextInt/encode(...)
methods now allocate the prefix in a U8
or U16
, which I think is kind of cute. It makes the new implementation a bit faster for small inputs. Big integer inputs obviously don't care about such optimizations, however. Edit: I suppose all small prefix sequences (max: "-0x&"
) fit in a U32
, so I may make this part non-generic too.
132
This patch addresses the above issue by removing the neutral mask (
"#"
) from the current binary integer description format. It also updates internal methods to ease support of future formats. In particular, I may want to encode and decode radix indicator prefixes ("0x"
) at some point. The internal algorithms now accept an arbitrary prefix sequence.