[Truffle] Privatise ByteList #4372
Conversation
chrisseaton
force-pushed
the
truffle-bytelist
branch
from
7180f26
to
e2a4664
Compare
| this.stringValue = decoded = decode(bytes, begin, realSize, "ISO-8859-1"); | ||
| } | ||
| return decoded; | ||
| return decode(bytes, begin, realSize, "ISO-8859-1"); |
There was a problem hiding this comment.
"ISO-8859-1" is wrong here (it should just use the ByteList encoding to decode), so we could take this opportunity to fix it as well.
…mply for everyone else.
chrisseaton
force-pushed
the
truffle-bytelist
branch
from
e2a4664
to
3180e2b
Compare
|
If we have our own copy of |
|
|
||
| public ByteListKey(ByteList bytes) { | ||
| this.bytes = bytes; | ||
| hash = Arrays.hashCode(bytes.bytes()); |
There was a problem hiding this comment.
It's possible that the ByteList changes and this would invalidate the hash. It looks like this is only used for regexp caching and that might be fine.
The second potential issue is ByteList#bytes always makes a copy. You could use ByteList#unsafeBytes instead, but you may have two equivalent ByteLists with different underyling byte arrays due to changing offsets or lengths. In which case, you'd want to call ByteList#unsafeBytes and loop through from ByteList#begin to ByteList#length.
|
|
||
| ByteListKey that = (ByteListKey) o; | ||
|
|
||
| return bytes.equals(that.bytes); |
There was a problem hiding this comment.
Note that ByteList#equals ignores encoding. This has been problematic in the past with things like our symbol table.
We can drop logic such as invalidating the hash, and space such as holding the hash and string value, if we privatise this.