Re: What about gcc-3.0.1-20010813 libiberty testresults failure on cygwin?

[Christian Jönsson <christian dot jonsson at foi dot se>]

OK, I don't understand how to run the tests individually and when I manually
run
the testsuite (without the /bin/sh) I get the same result:

$ pwd
/usr/local/src/objdir-gcc-3.0.1-20010813/libiberty/testsuite
Administrator@BOTTEN
/usr/local/src/objdir-gcc-3.0.1-20010813/libiberty/testsuite
$ ../../../gcc-3.0.1-20010813/libiberty/testsuite/regress-demangle
../../../gcc-3.0.1-20010813/libiberty/testsuite/demangle-expected
FAIL: --format=gnu _10PageButton\$__both
   result: _10PageButton\$__both
 expected: PageButton::__both
FAIL: --format=gnu _3RNG\$singleMantissa
   result: _3RNG\$singleMantissa
 expected: RNG::singleMantissa
FAIL: --format=gnu _5IComp\$_release
   result: _5IComp\$_release
 expected: IComp::_release
FAIL: --format=gnu _\$_10BitmapComp
   result: _\$_10BitmapComp
 expected: BitmapComp::~BitmapComp(void)
FAIL: --format=gnu _\$_9__io_defs
   result: _\$_9__io_defs
 expected: __io_defs::~__io_defs(void)
FAIL: --format=gnu _\$_Q23foo3bar
   result: _\$_Q23foo3bar
 expected: foo::bar::~bar(void)
FAIL: --format=gnu _\$_Q33foo3bar4bell
   result: _\$_Q33foo3bar4bell
 expected: foo::bar::bell::~bell(void)
FAIL: --format=gnu _vt\$foo
   result:  virtual table\$foo
 expected: foo virtual table
FAIL: --format=gnu _vt\$foo\$bar
   result:  virtual table\$foo\$bar
 expected: foo::bar virtual table
FAIL: --format=gnu _2T4\$N
   result: _2T4\$N
 expected: T4::N
FAIL: --format=gnu _Q22T42t1\$N
   result: _Q22T42t1\$N
 expected: T4::t1::N
FAIL: --format=gnu _\$_t6vector1Zdi
   result: _\$_t6vector1Zdi
 expected: vector<double>::~vector(int)
FAIL: --format=gnu _\$_t6vector1Zii
   result: _\$_t6vector1Zii
 expected: vector<int>::~vector(int)
FAIL: --format=gnu _GLOBAL_\$D\$set
   result: _GLOBAL_\$D\$set
 expected: global destructors keyed to set
FAIL: --format=gnu _GLOBAL_\$I\$set
   result: _GLOBAL_\$I\$set
 expected: global constructors keyed to set
FAIL: --format=gnu _vt\$t8BDDHookV1ZPc
   result:  virtual table\$t8BDDHookV1ZPc
 expected: BDDHookV<char *> virtual table
FAIL: --format=gnu _26_GLOBAL_\$N\$_tmp_n.iilg4Gya\$app_instance
   result: _26_GLOBAL_\$N\$_tmp_n.iilg4Gya\$app_instance
 expected: {anonymous}::app_instance
17 of 645 tests failed

Administrator@BOTTEN
/usr/local/src/objdir-gcc-3.0.1-20010813/libiberty/testsuite

Cheers,

/ChJ
----- Original Message -----
From: "Billinghurst, David (CRTS)" <David.Billinghurst@riotinto.com>
To: "'Christian Jönsson'" <christian.jonsson@foi.se>; "gcc"
<gcc@gcc.gnu.org>
Sent: Wednesday, August 15, 2001 2:31 PM
Subject: RE: What about gcc-3.0.1-20010813 libiberty testresults failure on
cygwin?


There is a PR on this, from some time ago. Ah yes - PR 1752   I picked it up
a month or so ago, but got side tracked again.

>From memory, the tests thatfail are exactly those that have a $ in the name.
I seemto recall that the tests pass if run manually, but a shell quoting
"bug" (or feature) meant that the test script passed "\$" in the tests
directly, while on irix the "\$" was passed as "$".

> -----Original Message-----
> From: Christian Jönsson [SMTP:christian.jonsson@foi.se]
> Sent: Wednesday, August 15, 2001 9:43 PM
> To: gcc
> Subject: What about gcc-3.0.1-20010813 libiberty testresults failure
> on cygwin?
>
> In the recent gcc-3.0.1-20010813/libiberty/testsuite/regress-demangle
> I get the following "failures" on a win2k/sp2 with an up to date cygwin
> install:
>
> /bin/sh ../../../gcc-3.0.1-20010813/libiberty/testsuite/regress-demangle
> .../../../gcc-3.0.1-20010813/libiberty/testsuite/demangle-expected
> FAIL: --format=gnu _10PageButton\$__both
>    result: _10PageButton\$__both
>  expected: PageButton::__both
> FAIL: --format=gnu _3RNG\$singleMantissa
>    result: _3RNG\$singleMantissa
>  expected: RNG::singleMantissa
> FAIL: --format=gnu _5IComp\$_release
>    result: _5IComp\$_release
>  expected: IComp::_release
> FAIL: --format=gnu _\$_10BitmapComp
>    result: _\$_10BitmapComp
>  expected: BitmapComp::~BitmapComp(void)
> FAIL: --format=gnu _\$_9__io_defs
>    result: _\$_9__io_defs
>  expected: __io_defs::~__io_defs(void)
> FAIL: --format=gnu _\$_Q23foo3bar
>    result: _\$_Q23foo3bar
>  expected: foo::bar::~bar(void)
> FAIL: --format=gnu _\$_Q33foo3bar4bell
>    result: _\$_Q33foo3bar4bell
>  expected: foo::bar::bell::~bell(void)
> FAIL: --format=gnu _vt\$foo
>    result:  virtual table\$foo
>  expected: foo virtual table
> FAIL: --format=gnu _vt\$foo\$bar
>    result:  virtual table\$foo\$bar
>  expected: foo::bar virtual table
> FAIL: --format=gnu _2T4\$N
>    result: _2T4\$N
>  expected: T4::N
> FAIL: --format=gnu _Q22T42t1\$N
>    result: _Q22T42t1\$N
>  expected: T4::t1::N
> FAIL: --format=gnu _\$_t6vector1Zdi
>    result: _\$_t6vector1Zdi
>  expected: vector<double>::~vector(int)
> FAIL: --format=gnu _\$_t6vector1Zii
>    result: _\$_t6vector1Zii
>  expected: vector<int>::~vector(int)
> FAIL: --format=gnu _GLOBAL_\$D\$set
>    result: _GLOBAL_\$D\$set
>  expected: global destructors keyed to set
> FAIL: --format=gnu _GLOBAL_\$I\$set
>    result: _GLOBAL_\$I\$set
>  expected: global constructors keyed to set
> FAIL: --format=gnu _vt\$t8BDDHookV1ZPc
>    result:  virtual table\$t8BDDHookV1ZPc
>  expected: BDDHookV<char *> virtual table
> FAIL: --format=gnu _26_GLOBAL_\$N\$_tmp_n.iilg4Gya\$app_instance
>    result: _26_GLOBAL_\$N\$_tmp_n.iilg4Gya\$app_instance
>  expected: {anonymous}::app_instance
> 17 of 645 tests failed
>
> Is this to be expected? (It's been like this all the time I have tried...)
> Is this something that *could* be made work? (Is it "needed"?)
>
> Cheers,
>
> /ChJ