[MDEV-21976] main.udf fails on Ubuntu Bionic on Travis-CI Created: 2020-03-18  Updated: 2020-05-18  Resolved: 2020-05-18

Status: Closed
Project: MariaDB Server
Component/s: Tests
Affects Version/s: 10.5
Fix Version/s: 10.5.4, 10.1.46, 10.2.33, 10.3.24, 10.4.14

Type: Bug Priority: Minor
Reporter: Otto Kekäläinen Assignee: Marko Mäkelä
Resolution: Fixed Votes: 0
Labels: None

Attachments: PNG File image-2020-05-15-08-56-42-223.png     PNG File screenshot-1.png    

 Description   

From https://travis-ci.org/github/ottok/mariadb/jobs/663987331

main.udf                                 w1 [ fail ]
 
        Test ended at 2020-03-18 15:56:11
 
CURRENT_TEST: main.udf
 
--- /home/travis/build/ottok/mariadb/mysql-test/main/udf.result	2020-03-18 15:30:33.031262216 +0000
 
+++ /home/travis/build/ottok/mariadb/mysql-test/main/udf.reject	2020-03-18 15:56:11.234936942 +0000
 
@@ -32,7 +32,7 @@
 
 ERROR HY000: Can't initialize function 'lookup'; Wrong arguments to lookup;  Use the source
 
 select lookup("localhost");
 
 lookup("localhost")
 
-127.0.0.1
 
+127.0.1.1
 
 select reverse_lookup();
 
 ERROR HY000: Can't initialize function 'reverse_lookup'; Wrong number of arguments to reverse_lookup;  Use the source
 
 select reverse_lookup("127.0.0.1");
 
mysqltest: Result content mismatch

Seems the Travis-CI builders occasionally have 127.0.1.1 as localhost, which is a valid value. Do we want the test to allow that value as well?

 Comments   
Comment by Otto Kekäläinen [ 2020-03-19 ]

I did some more testing and can confirm, that the test suite 'main' fails on Ubuntu Bionic with this error no matter if using gcc-5, 6, 7, 8 or 9. So it is not a GCC specific issue

Also clang builds on Bionic fail with this same error, e.g. https://travis-ci.org/github/ottok/mariadb/jobs/664165657

Maybe it is related to some build dependency that interacts with main.udf?

Comment by Elena Stepanova [ 2020-04-05 ]

Looks like specifics of network configuration, is it not?
Probably the possibility needs to be taken into account in the test.

Comment by Otto Kekäläinen [ 2020-04-07 ]

I disabled this temporarily in https://github.com/MariaDB/server/commit/a135f0ab88d63b9a8976d6b3010f27766c38873d, otherwise all other QA work would be stalled. Please re-enable it once this issue has been solved.

Comment by Otto Kekäläinen [ 2020-05-15 ]

Current 10.5 head on Travis-CI is failing occasionally on this issue. Seems disabling the test in a135f0ab88d63b9a8976d6b3010f27766c38873d did affect only Debian builds, not MTR in general?

This also shows up as failing CI results in pull requests:

sanja Do you have any estimate when you could look into this?

Comment by Daniel Black [ 2020-05-16 ]

trivial fixes:
https://github.com/MariaDB/server/pull/1543 (10.1)
https://github.com/MariaDB/server/pull/1542 (10.5)

Comment by Marko Mäkelä [ 2020-05-18 ]

I merged danblack’s fix to 10.1 and will merge it up.

Generated at Thu Feb 08 09:11:15 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.