Hi!
Peregrine wrote:
On Monday 18 July 2005 03:06am, Joerg Bruehe wrote:
[[...]]
You are correct. The compile was successfully completed, as in there were no
compile errors. However, I am building RPMs; that was the "build process"
that I was referring to. Sorry to have been confusing.
No confusion - I just wanted to be explicit.
Which platform are you using, or which specific features are you
combining, so that you build by yourself?
Fedora, RHEL, CentOS & SUSE distributions for i386 & AMD64. Now that these
distributions have sufficiently current packages of 4.1 available, I no
longer build those; just 5.0 (as close to the way those distributions would
probably build them) for development and testing.
Ok, that makes sense.
[[...]]
mysql [ pass ]
mysql_client_test [ fail ]
Errors are
(from
/home/lamontp/rpmbuild/BUILD/mysql-5.0.9-beta/mysql-test/var/log/mysqltes
t-time) : mysql_client_test.c:3811: check failed: 'rc == 0'
[[...]]
Yes, both look good probably.
You just ran into one reason why version 5.0 is still in "beta" state.
It is a known bug, MySQL development is working on this.
Is there any better reference to the bug available (bug ID) so that I could
follow it? I could also try contribute a fix, in that case.
No, sorry there is not. Depending on machine and test mode (especially
"embedded"), "mysql_client_test" shows different test failures. When I
first replied, I thought I had seen exactly this error, but now my
search showed it only for "embedded" tests.
If your experiments show the reason of the failure or even lead to a
patch, that would be great and very welcome!
I propose to run "make test-force", so that this failing test does not
prevent the subsequent tests from being taken.
I will do that for testing purposes. Currently, I build 5.0.x for development
and testing, not for production, so I will need to build a set of RPMs
without test-force to continue distributing for those purposes.
Ok, it is up to you in which way you run the tests; I just want to be
sure you know about this way to run them all.
Not needed. Typically, tests are skipped if they are declared to test a
component which is not included in the binary being tested. [[...]]
Which is what I expected. Thanks for the confirmation. I only offered or the
sake of completeness.
Yes, that is how I understood it.
Regards,
Jörg
--
Joerg Bruehe, Senior Production Engineer
MySQL AB, www.mysql.com
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]