mk-table-checksum test is too finicky

[GoogleCodeExporter]

This test is driving me nuts while releasing :-)

The stuff that pipes through awk and then back to diff just hangs about
every time.  I have to ctrl-C it and then I get

ok 45 - --schema --algorithm=BIT_XOR
ok 46 - --schema --algorithm=ACCUM
fetch() without execute() at line 1540 while doing mysql.func on
P=12345,h=127.0.0.1
not ok 47 - --schema --checksum
#   Failed test '--schema --checksum'
#   in mk-table-checksum.t at line 222.
#          got: 256
#     expected: 0
ok 48 - --schema --chunksize=1M
ok 49 - --schema --count
ok 50 - --schema --crc
not ok 51 - --schema --emptyrepltbl
#   Failed test '--schema --emptyrepltbl'
#   in mk-table-checksum.t at line 222.
#          got: 256
#     expected: 0
ok 52 - --schema --float-precision=3
not ok 53 - --schema --function=FNV_64
#   Failed test '--schema --function=FNV_64'
#   in mk-table-checksum.t at line 222.

I think we ought to do

mk-table-checksum --checksum ..... > tmpfile
diff tmpfile samplefile

no pipes involved.  When I strace'd the hung process, it just said it was
stuck inside of wait4(....) so I think we need to ditch the fancies.

Tried to start this, but saw the need for fixing issue 103 then :)

Original issue reported on code.google.com by baron.schwartz on 18 Oct 2008 at 9:16

[GoogleCodeExporter]

Original comment by baron.schwartz on 18 Oct 2008 at 9:17

• Added labels: Type-Task
• Removed labels: Type-Defect

[GoogleCodeExporter]

Where's my pipe wrench...

Original comment by dan...@percona.com on 25 Oct 2008 at 5:51

• Changed state: Started
• Added labels: Tool-mk_table_checksum

[GoogleCodeExporter]

There it is: r2451

Original comment by dan...@percona.com on 25 Oct 2008 at 5:57

• Changed state: Fixed