MySQL Bugs: #26561: Test 'execution_constants' fails on 64 bit Mac OS X PPC and HP-UX 11.23 IA64

Bug #26561	Test 'execution_constants' fails on 64 bit Mac OS X PPC and HP-UX 11.23 IA64
Submitted:	22 Feb 2007 10:39	Modified:	16 Apr 2007 17:44
Reporter:	Kent Boortz	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Server: Tests	Severity:	S7 (Test Cases)
Version:	5.1.16, 5.0.37	OS:	MacOS (Mac OS X PPC64, HP-UX 11.23 IA64, Linux IA64)
Assigned to:	Assigned Account	CPU Architecture:	Any

Description:
Test 'execution_constants' fails the same on the two platforms

 mysqltest: At line 68: query '$query_head 0 $query_tail'
 failed with wrong errno 2013: 'Lost connection to MySQL server during query',
 instead of 0...

How to repeat:
Run test case on the specified platforms

Note that this test case, as far as can be seen
right now, only fails running tests with the
"mysqld-debug" server.

Also observed in building 5.0.37 on HP/UX 11.23 and Linux IA64.

Intermediate result from the 5.0.38 release builds:

Failure described above occurs in several runs on IA64 CPUs,
using different compilers (Intel, HP).
List of failing runs (till now - "cluster" build still active):

rx2620b-icc-glibc23-5.0-enterprise-gpl.log   ps
rx2620b-icc-glibc23-5.0-enterprise.log   ps
rx2620b-icc-glibc23-5.0-cluster.log   ps
Do-rpm-sles9-ia64-5.0-enterprise-gpl.sles9-ia64.log   normal
rx2620b-icc-glibc23-5.0-enterprise-gpl.log   debug
rx2620b-icc-glibc23-5.0-enterprise-gpl.log   normal
rx2620b-icc-glibc23-5.0-enterprise.log   debug
rx2620b-icc-glibc23-5.0-enterprise.log   normal
Do-rpm-rhas4-ia64-5.0-cluster.rhel4-ia64.log   normal
Do-rpm-rhas4-ia64-5.0-enterprise-gpl.rhel4-ia64.log   normal
Do-rpm-rhas4-ia64-5.0-enterprise.rhel4-ia64.log   normal
Do-rpm-rx2620b-5.0-cluster.rhel3-ia64.log   normal
Do-rpm-rx2620b-5.0-enterprise-gpl.rhel3-ia64.log   normal
Do-rpm-rx2620b-5.0-enterprise.rhel3-ia64.log   normal
Do-rpm-sles9-ia64-5.0-cluster.sles9-ia64.log   normal
Do-rpm-sles9-ia64-5.0-enterprise.sles9-ia64.log   normal
rx2620b-icc-glibc23-5.0-cluster.log   debug
rx2620b-icc-glibc23-5.0-cluster.log   normal

hpita2-64bit-5.0-enterprise.log   debug
hpita2-64bit-5.0-cluster.log   debug

The failure on PPC-64 has not been observer in 5.0.38 builds,
this is in line with the above remark about 5.0.37 builds.
The restriction to "debug" runs (of the original report) does not apply to 5.0.

I confirm that this bug also exists in 5.0.37 on Linux PPC64 (Gentoo Linux, 64-bit kernel, 32-bit userland). This is blocking the release of 5.0.37 on  Gentoo now. All the arches here are big-endian perhaps there's something to that.
The bug does NOT occur on Linux AMD64.

a duplicate of (reopened) Bug#21476