[
https://issues.apache.org/jira/browse/CALCITE-2273?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16457488#comment-16457488
]
yuqi edited comment on CALCITE-2273 at 4/28/18 9:29 AM:
--------------------------------------------------------
As far as i see, unicode that have length more than 4 will be converted to two
normal unicode with length of 4
{code:java}
U&'\01CCCC'{code}
will be covert to '\uD833\uDCCC'.
was (Author: yuqi):
As far as i see, unicode that have length more than 4 will be converted to two
normal unicode with length of 4
U&'\01CCCC'
will be covert to '\uD833\uDCCC'.
> <Unicode 6 digit escape value> misinterpreted
> ----------------------------------------------
>
> Key: CALCITE-2273
> URL: https://issues.apache.org/jira/browse/CALCITE-2273
> Project: Calcite
> Issue Type: Bug
> Reporter: Zhong Yu
> Assignee: Julian Hyde
> Priority: Major
>
>
>
> The following string literal is accepted by Calcite parser, but the result
> value in Java is incorrect
> {code:java}
> U&'\+01F600'{code}
>
> Cause: currently, SqlLiteral.unescapeUnicode() only intends to handle 4-digit
> unicode escape value in the form of "\xyzw" . When given the 6-digit form
> "\+xyzwrs", it parses the four chars "+xyz" as a hexadecimal, which succeeds
> too. Therefore the result value contains 4 characters: [\u0xyz, w, r, s]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
