Description

lukashov-artem

opened on Sep 13, 2019

Issue body actions

Describe the bug
String of multibyte characters overflows variable when converting ut_varchar2_list to ut_varchar2_rows

Provide version info
utPLSQL v3.1.2.2134
Oracle 12.2 charset UTF8

To Reproduce
Steps to reproduce the behavior:

1. generate multibyte character string rpad('№',1334,'№')||'№'
2. make it hit the reporter
3. See error

Expected behavior
the error is not raised and the log lines are displayed correctly

Example code

create or replace package utestoq.test_package as
  --%suite(Test multibyte strings)

  --%test
  procedure pTest1;
end;
/
create or replace package body utestoq.test_package as
  procedure pTest1 as
  begin
    utest.ut.fail(rpad('№',1332,'№'));
  end;
end;
/
select * from table(utest.ut.run('utestoq.test_package'))

[1]: ORA-12899: значение для столбца "UTEST"."UT_OUTPUT_BUFFER_TMP$"."TEXT" слишком велико (фактическое: 4002, максимальное: 4000)
[1]: ORA-06512: на  "UTEST.UT_RUNNER", line 150
[1]: ORA-06512: на  "UTEST.UT_OUTPUT_TABLE_BUFFER", line 61
[1]: ORA-06512: на  "UTEST.UT_OUTPUT_REPORTER_BASE", line 46
[1]: ORA-06512: на  "UTEST.UT_DOCUMENTATION_REPORTER", line 38
[1]: ORA-06512: на  "UTEST.UT_CONSOLE_REPORTER_BASE", line 26
[1]: ORA-06512: на  "UTEST.UT_DOCUMENTATION_REPORTER", line 101
[1]: ORA-06512: на  "UTEST.UT_DOCUMENTATION_REPORTER", line 11

Fix
We tried to fix it. I'm new to github, so I did not understand where to click to offer a fix.
lukashov-artem@2673e70#diff-b081c1c6f662c3d212eb2be99d4d58bd
lukashov-artem@be17067

Activity

1. pesse commented on Sep 13, 2019

   

   pesse

   on Sep 13, 2019

   Member

   More actions

   Thank you for bringing this up!
   Seems the bug might still exist in the latest version
2. jgebal commented on Sep 13, 2019

   

   jgebal

   on Sep 13, 2019

   Member

   More actions

   Open a pull request for that change into our repo.

   
3. jgebal

   added

   bug

   on Sep 13, 2019
4. jgebal

   added this to the 3.1.9 milestone on Sep 13, 2019
5. jgebal commented on Sep 13, 2019

   

   jgebal

   on Sep 13, 2019

   Member

   More actions

   Could you also add a unit test to cover that situation?
   Our tests are in test folder.
   The test will run on our Travis once you open a Pull Request.
   For details see the contributing guide.

   Thanks again for finding the issue, reporting it and fixing it 👍
6. jgebal commented on Sep 22, 2019

   

   jgebal

   on Sep 22, 2019

   Member

   More actions

   There definitely is a problem with muti-byte characters.
   I'm not sure if the fix solves them all.

   @lukashov-artem - I've tried applying you fix on my AL32UTF8 databases and encountered problems in other parts of code.

   Most probably your fix is not addressing all the issues.
7. jgebal

   modified the milestones: 3.1.9, 3.1.10 on Nov 30, 2019
8. jgebal

   removed this from the 3.1.10 milestone on Feb 23, 2020
9. jgebal

   pinned this issue on Jun 15, 2026

Issue actions