Hema,
I think this patch should move the backup_charsets test to the new
backup_charset suite.
Patch looks good so far. Just a few comments. Most are general and apply to
the entire patch (things to watch out for mainly).
I will wait for your last patch before finalizing my review approval.
Chuck
> 2683 Hema Sridharan 2008-08-16
> WL#4227(Test of identifiers and objects for different
> character sets and
> collations). The testcases for this WL is committied
> according to changes requested by Chuck and Rafal.
> There is a seperate backup_charsets suite created which
> will have list of
> all combinations of storage engine + character set + collation.
> Each test case from this suite will run against all
> these combinations.
> Right now this suite includes the following tests:
> backup_triggers.test
> backup_functions.test
> backup_procedures.test
> backup_datatypes.test
> backup_partitions.test
> added:
...
> === modified file 'mysql-test/suite/backup/t/backup_functions.test'
...
> #Again calling function to check if Sf are backed up and
> restored properly or
Or what?
> === modified file 'mysql-test/suite/backup/t/backup_procedures.test'
...
> #Excercise the objects after restore to make sure if objects
> are backedup and
> #restored properly.
Spelling errrros. Technically, there is no verb 'backedup' but we can make
new one! :P
...
> +SELECT * FROM d2;
> +region summary data details queries query2 extract paras
> +xxxxxxxx Testofonline backup aaaaaaaaaa
> bbbbbbbbbbb hhhhhhhhhhh kkkkkkkkkkkkk
> zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz
> zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz onlinebackup1
Just a suggestion -- no need to change it, but we want to get away from
'online backup' as we are calling the project 'MySQL backup'.
...
> +UPDATE bup_ts.t2 SET a='#' WHERE a='w';
> +SELECT * FROM bup_ts.t4;
> +a
> +J
> +O
> +B
> +J
> +O
> +B
Is there any significance to the choice of letters in the output? Shouldn't
we include some unusual characters?
...
> +if(`SELECT '$coll'='latin2_czech_cs'`)
> +{
> + skip "This test does'nt support latin2_czech_cs collation
> BUG#37854" ;
> +}
Doesn't no does'nt.
...
> +colours SET('red','blue','yellow'),
Using the King's English are we? Colors is US. No need to change this as
technically it isn't wrong.
...
> +--echo **backup data**
Why do you use ** <message> ** in some tests but not others? I don't care
about the formatting but I would like to see consistency.
> +#############################################################
> ##############
> +# Author: Hema
> +# Date: 2008-06-28
> +# Purpose: To test the backup and Restore of different
> partitions using
> +# Reserved words as identifiers for all Character sets, coll and SE
> +#############################################################
> ##################
Spacing issues in text.
...
> +--echo Bug #34391 Character sets: crash if char(), utf32, innodb
> +--echo Bug #33566 Backup: crash with partitions and Falcon
> +--echo Bug #37551 Junk detected in data contents sometimes
> when utf8mb3
> +--echo character set is used.
> +--echo Bug #37554 Use of character set and collate as
> 'filename' shows
> +--echo unusual behaviour.
> +--echo Bug #35499 View when created with swe7 character set fails.
> +--echo Bug#38784 Mysql server crash if table is altered by
> partition changes.
> +
> +#Bug #33566 Backup: crash with partitions and Falcon.
> +#Remove this condition once the bug#33566 is fixed.
Wow! Is this WL addressing all of these bugs or is this patch just tests for
confirmation of the bugs? It isn't clear from the comment.
...
> +#Bug#20129
> +# We can include these operations in the test case once this
> bug is fixed.
There are several of these statements. Would be very good to have them
documented in the worklog as well.
Chuck
