0005928: serialize/unserialize unicode(CJK) string yields unkown characters sometimes.

Viewing Issue Advanced Details [ Jump to Notes ]

[ View Simple ] [ Issue History ] [ Print ]

Category

Severity

Reproducibility

Date Submitted

Last Update

0005928

[Quercus]

block

always

07-27-15 23:12

07-30-15 06:43

Reporter

weich

View Status

public

Assigned To

Priority

normal

Resolution

open

Platform

Status

new

Projection

none

OS Version

ETA

none

Fixed in Version

Product Version

4.0.36

Product Build

Summary

0005928: serialize/unserialize unicode(CJK) string yields unkown characters sometimes.

Description

I have a application runing on nginx/php5-fpm server, it can display the chines/japanese characters correctly. but after I switched to resin4.0.44 server. some page display "????" somewhere.

after long time trouble-shooting, I found there are some difference between the result generated by php5 function serialze/unserialze and the one generated by the countparts in resin4.

I guess the bug is in the class: StringBuilderValue, which get the wrong byte length when try to make a copy of a string. But I need your guys' confirmation, you are experts.

is this correct?
file : StringBuilderValue.java

  public StringBuilderValue(String s)
  {
    int len = s.length();

    _buffer = new byte[len];
    _length = len;

    for (int i = 0; i < len; i++) {
      _buffer[i] = (byte) s.charAt(i);
    }
  }

shouldn't it use "s.getBytes().length" as the following lines:

  public StringBuilderValue(String s) {
    byte[] bytes = s.getBytes();
    int len = bytes.length;

    _buffer = new byte[len];
    _length = len;

    System.arraycopy(bytes, 0, _buffer, 0, len);
}

Steps To Reproduce

Additional Information

Attached Files

api.php [^] (41 bytes) 07-30-15 06:29

test.php [^] (1,247 bytes) 07-30-15 06:41

Relationships

Notes
(0006645) weich 07-30-15 06:43	step to reproduce: open the test.php in browser. click "&26356;&26032;"&65292; and you will see the serialized string at the page end.

Issue History
Date Modified	Username	Field	Change
07-27-15 23:12	weich	New Issue
07-29-15 22:50	weich	Note Added: 0006644
07-29-15 23:31	weich	Note Deleted: 0006644
07-30-15 06:29	weich	File Added: api.php
07-30-15 06:41	weich	File Added: test.php
07-30-15 06:43	weich	Note Added: 0006645

Mantis 1.0.0rc3[^]

35 total queries executed.
28 unique queries executed.