> An even more mature, and hopefully even more final, release candidate
> of DBD::Oracle 1.16 is available for testing at:
>  
> http://homepage.eircom.net/~timbunce/DBD-Oracle-1.16-rc7-20040826.tar.gz
[snip]
> Please help by taking a moment to at least check it builds 
> and tests okay
> in your environment. Especially if you had any problems with RC6.

Results here: http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/

http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/dbd_oracle_MSWin32_summary.l
og
http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/dbd_oracle_cygwin_summary.lo
g
http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/dbd_oracle_linux_summary.log

24implicit_utf8 tests 33,34,70,71 (fetching nvarchar2 unicode "smiley face"
character) have some fails; they result in the following output but only on
databases with a single-byte basic character set and NLS_LANG set to a
Unicode encoding.

not ok 33 - byte_string test of row 4: column: nch  (uft8) smiley face
#     Failed test (t/nchar_test_lib.pl at line 312)
#          got: '194|191'
#     expected: '226|152|186'
not ok 34 - nice_string test of row 4: column: nch  (uft8) smiley face
#     Failed test (t/nchar_test_lib.pl at line 316)
#          got: '┐'
#     expected: '\x{263A}'
ok 35 - byte_string test of row 4: column: descr  (uft8) smiley face
ok 36 - nice_string test of row 4: column: descr  (uft8) smiley face
#    row 4: DUMP(nch) = Typ=1 Len=2: 194,191
ok 37 - number of rows fetched

>From the DUMP output it looks like the problem's on insert rather than
select since the wrong data is in the table already?
194,191 is the utf8 encoding of U+00BF "INVERTED QUESTION MARK" in case
that's any help.

Full output of the test:
http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/24implicit_utf8.t.log
And with DBI_TRACE=9:
http://www.andyh.uklinux.net/dbd-oracle-1.16rc7/24implicit_utf8.t.trace9.log

30long.t no longer hangs under any configuration, but it fails some subtests
with 8i client with NLS_LANG=UTF8
With 8i client -> 9i database it's just one test:

result is non-UTF8
not ok 322 - CLOB should be UTF8
#     Failed test (t/30long.t at line 275)

8i -> 10g fails 314->368:

read via blob_read_all
DBD::Oracle::st blob_read failed: ORA-01403: no data found (DBD NEED_DATA:
OCILo
bRead) [for Statement "select idx, lng, dt from dbd_ora__drop_me order by
idx"]
at t/30long.t line 429.
not ok 314 - blob_read_all = length($long_data[0])
#     Failed test (t/30long.t at line 260)
#          got: 0
#     expected: 10240
not ok 315 - Strings are of different lengths (0 vs 10240)
#     Failed test (t/30long.t at line 261)
DBD::Oracle::st fetchrow_arrayref failed: ORA-03118: two-task coroutine has
inva
lid state (DBD ERROR: OCIStmtFetch) [for Statement "select idx, lng, dt from
dbd
_ora__drop_me order by idx"] at t/30long.t line 264.
not ok 316 - fetchrow_arrayref 2: select idx, lng, dt from dbd_ora__drop_me
orde
r by idx

Since it's now apparently trashed the connection the rest of the tests fail.

Cygwin reports many Unicode-related failures - I don't know where to start
with these, I don't tend to use Perl on Cygwin much.

-- 
Andy Hassall <[EMAIL PROTECTED]> / Space: disk usage analysis tool
<http://www.andyh.co.uk> / <http://www.andyhsoftware.co.uk/space> 

Reply via email to