[MSVC 8-x64] 0.printf test fails with number of assertions
----------------------------------------------------------
Key: STDCXX-798
URL: https://issues.apache.org/jira/browse/STDCXX-798
Project: C++ Standard Library
Issue Type: Bug
Components: Test Driver
Affects Versions: 4.2.0
Environment: All
Reporter: Farid Zaripov
Fix For: 4.2.1
The 0.printf test runs successfully on 32-bit MSVC, but fails on 64-bit MSVC.
The failed assertions listed below:
{noformat}
# Assertion failed on line 419: rw_sprintf("%s", ...) == "(invalid address
0x0000000000000020)", got "(invalid address 0x00000020:00000000)"
# Assertion failed on line 429: rw_sprintf("%ls", ...) == "(invalid address
0x0000000000000020)", got "(invalid address 0x00000020:00000000)"
# Assertion failed on line 432: rw_sprintf("%ls", ...) == "(misaligned address
0x000000014011e6e9)", got "(misaligned address 0x4011e6e9:00000001)"
# Assertion failed on line 507: rw_sprintf("%{#ls}", ...) == "(invalid address
0x0000000000000020)", got "(invalid address 0x00000020:00000000)"
# Assertion failed on line 510: rw_sprintf("%{#ls}", ...) == "(misaligned
address 0x000000014011e6e9)", got "(misaligned address 0x4011e6e9:00000001)"
# Assertion failed on line 1994: rw_printf("%p", 0, 0, 0) ==
"0000000000000000", got "00000000:00000000"
# Assertion failed on line 1995: rw_printf("%p", 1, 0, 0) ==
"0000000000000001", got "00000001:00000000"
# Assertion failed on line 1996: rw_printf("%p", -1, 0, 0) ==
"00000000ffffffff", got "ffffffff:00000000"
# Assertion failed on line 1998: rw_printf("%#p", 0, 0, 0) ==
"0x0000000000000000", got "0x00000000:00000000"
# Assertion failed on line 1999: rw_printf("%#p", 291, 0, 0) ==
"0x0000000000000123", got "0x00000123:00000000"
# Assertion failed on line 2000: rw_printf("%#p", -1, 0, 0) ==
"0x00000000ffffffff", got "0xffffffff:00000000"
# Assertion failed on line 2047: rw_sprintf("%{f}", ...) == "0000000000000000",
got "00000000:00000000"
# Assertion failed on line 2048: rw_sprintf("%{f}", ...) == "0000000000000001",
got "00000001:00000000"
# Assertion failed on line 2049: rw_sprintf("%{f}", ...) == "00000000ffffffff",
got "ffffffff:00000000"
# Assertion failed on line 2051: rw_sprintf("%{#f}", ...) ==
"0x0000000000000000", got "0x00000000:00000000"
# Assertion failed on line 2052: rw_sprintf("%{#f}", ...) ==
"0x0000000000000123", got "0x00000123:00000000"
# Assertion failed on line 2053: rw_sprintf("%{#f}", ...) ==
"0x00000000ffffffff", got "0xffffffff:00000000"
# Assertion failed on line 2614: rw_sprintf("[%{.3P}]", ...) == "[12345]", got
"[123]"
# Assertion failed on line 2615: rw_sprintf("[%{10.3P}]", ...) == "[
12345]", got "[ 123]"
# Assertion failed on line 2616: rw_sprintf("[%{+10.3P}]", ...) == "[
12345]", got "[ 123]"
# Assertion failed on line 2617: rw_sprintf("[%{-10.3P}]", ...) == "[12345
]", got "[123 ]"
# Assertion failed on line 2702: rw_sprintf("%{t}", ...) == "(invalid address
0x0000000000000020)", got "(invalid address 0x00000020:00000000)"
# Assertion failed on line 2705: rw_sprintf("%{t}", ...) == "(misaligned
address 0x000000014011e6e9)", got "(misaligned address 0x4011e6e9:00000001)"
Failed 23 out of 1743 assertions.
{noformat}
The first problem is difference between implementations of the
format_bad_address() in 0.printf.cpp file and _rw_fmtbadaddr() in rwtest
library. The _rw_fmtbadaddr() uses _rw_fmtpointer() to format the pointer value
(the pointer divided to parts with size = sizeof (long) and these parts are
separated with colon ':') while format_bad_address() formats the pointer as
integer by casting it to size_t. The same problem when "%p" and "%{f}" formats
are tested: the expected value hardcoded in the test is a string with 16
numerals
I'm not sure what should be corrected: the format_bad_address() and expected
results for pointer formatting in 0.printf test of the _rw_fmtpointer() in test
driver?
The second problem is difference between implementations of the _rw_fmtlong()
and _rw_fmtllong(). I think they should be implemented the same way.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
