[MDEV-6194] unaligned memory access with pcretest Created: 2014-05-01  Updated: 2014-08-12  Due: 2014-07-10  Resolved: 2014-08-12

Status: Closed
Project: MariaDB Server
Component/s: None
Affects Version/s: 10.0.10
Fix Version/s: 10.0.13

Type: Bug Priority: Minor
Reporter: Brad Smith Assignee: Unassigned
Resolution: Fixed Votes: 0
Labels: None
Environment:

OpeBSD, alpha / sparc64



 Description   

Not sure if this is just the test or if MariaDB is affected but running the test suite shows unaligned memory access on strict alignment architectures via pcretest and the test fails.

===>  Regression tests for mariadb-10.0.10
[1/1] cd /usr/obj/pobj/mariadb-10.0.10/build-alpha &&
/usr/local/bin/ctest --force-new-ctest-process
pid 27013 (pcretest): unaligned access: va=0x73dab092 pc=0x2d23bbcc ra=0x2d23b5f4 op=ldl



 Comments   
Comment by Sergei Golubchik [ 2014-05-10 ]

We aren't officially supporting alpha or sparc yet — I cannot even repeat this issue, as we don't use alpha/sparc builders.
But I'd be happy to apply a bugfix patch if there is one.

Anyway, unless MariaDB server fails similarly, I'd expect it to be just the test issue. If this is the case you can simply ignore it.

Comment by Brad Smith [ 2014-05-10 ]

This isn't about officially "supporting" any particular architecture. Memory alignment is a pretty fundamental issue with programming period.

I can easily provide access to a sparc64 system.

If I had a bug fix patch I would have provided it.

How do I know if it is working properly within the server or not? That's the whole point of unit tests. It looks like pcreteset tests the library on its own and not the functionality within the server itself.

Comment by Brad Smith [ 2014-08-12 ]

The test is now passing with the 10.0.13 release.

Comment by Sergei Golubchik [ 2014-08-12 ]

Thanks! I'm closing this issue then.

Generated at Thu Feb 08 07:10:04 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.