MySQL Bugs: #36787: Test funcs_1.charset_collation

Bug #36787	Test funcs_1.charset_collation_1 failing
Submitted:	18 May 2008 21:18	Modified:	28 Jul 2008 15:35
Reporter:	Patrick Crews	Email Updates:
Status:	Closed	Impact on me:	None
Category:	Tests: Server	Severity:	S3 (Non-critical)
Version:	5.1+	OS:	Any
Assigned to:	Matthias Leich	CPU Architecture:	Any
Tags:	charset, collation, funcs_1, Tests

Description:
Test funcs_1.charset_collation_1 is failing.

It appears that a number of charsets / collations that were a part of source builds are no longer included when executing BUILD/compile-dist.

In addition, several charsets / collations have changed characteristics.

Still researching to see if this is expected behavior or not.

How to repeat:
Build a version of 5.1 with any of the following characteristics:
 Enterprise, Classic , Pushbuild or Source-without-max build

Execute the test:

./mysql-test-run --suite=funcs_1 --do-test charset

You will observe that a number of expected charset / collations aren't in the build and that some existing charset / collations have changed from the expected results.

example output:
@@ -109,11 +86,10 @@
 latin1_german2_ci	latin1	31		Yes	2
 latin1_spanish_ci	latin1	94		Yes	1
 latin1_swedish_ci	latin1	8	Yes	Yes	1
-latin2_bin	latin2	77		Yes	1
-latin2_croatian_ci	latin2	27		Yes	1
-latin2_czech_cs	latin2	2		Yes	4
-latin2_general_ci	latin2	9	Yes	Yes	1
-latin2_hungarian_ci	latin2	21		Yes	1
+latin2_bin	latin2	77			0
+latin2_croatian_ci	latin2	27			0
+latin2_general_ci	latin2	9	Yes		0
+latin2_hungarian_ci	latin2	21			0

Suggested fix:
Determine if these changes are expected.

If they are, re-record the .result file and verify that the existing charsets / collations are proper for the build.

If not, determine what is causing the change from the expected set of charsets / collations and rectify the problem.

Result of discussion with Joerg (build team):
---------------------------------------------
Although the "vanilla" build does not correspond to
- any package given to customers
- packages created during development work
  (BUILD/compile-....-max and similar)
it is a variant which is most probably used by some
customers (Example: development of third party
storage engines etc.).
Build steps:
   BUILD(autorun.sh
   ./configure
   make
This means the funcs_1 tests must pass.
But there is no strict need to have a dedicated
charset_collation variant for such a package.

Result with experiments on downloadable packages
and typical compile from source packages 
(all MySQL5.1):
------------------------------------------------
The charset_collation variants pass on all
variants except "BUILD/compile_dist" which is
to some extend similar to "vanilla".
All charset_collation variants get skipped in
case of "mysql-advanced". Maybe one of the
existing variant/result combination fits
and only the detection within the toplevel
script has to be adjusted.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48662

2660 Konstantin Osipov	2008-06-27
      Disable funcs_1.charset_collation_1 (Bug#36787)

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48667

Ok to push.

Tested on 5.1 with and without charset ucs2_bin and behaves as expected.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48758

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48803

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48819

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48822

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48823

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48973

Pushed into 5.0.66

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/48974

Pushed into 5.0.66

Patch pushed to
mysql-5.0
mysql-5.1
mysql-6.0

Pushed into 5.1.28 and 6.0.7-alpha

Pushed into 5.0.68  (revid:kpettersson@mysql.com-20080715152926-s5kgnqhtu93b3c8v) (pib:2)
(Retry automatic marking, to ensure nothing is missed. cm01)

Test case changes. No changelog entry needed.

Pushed into 5.1.28  (revid:joerg@mysql.com-20080709112531-2yipmftwu8lqrc0n) (version source revid:joerg@mysql.com-20080709112531-2yipmftwu8lqrc0n) (pib:3)