This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
[RFC] Decimal Floating Point support for GDB (Part 2: testcase -Revised)
- From: Wu Zhou <woodzltc at cn dot ibm dot com>
- To: Daniel Jacobowitz <drow at false dot org>
- Cc: bje at au1 dot ibm dot com, gdb-patches at sources dot redhat dot com
- Date: Tue, 18 Oct 2005 18:20:27 +0800 (CST)
- Subject: [RFC] Decimal Floating Point support for GDB (Part 2: testcase -Revised)
- References: <Pine.LNX.4.63.0509281256030.1231@linux.site><20051002210053.GA32664@nevyn.them.org>
Hi Daniel,
Appended is the testcase for newly rewrited DFP patch. I had seperated
the testcases into two parts: dfp-exprs.exp which don't need any DFP code
to be run, dfp-test.exp which need to compile a source file dfp-test.c.
In dfp-test.exp, I added a few tests to verify that GDB could correctly
handle the DFP members in C struct and also some tests to handle dfp value
setting by "dfp1=dfp2". I had tested it on x86 using dfp-branch gcc. All
passed. And no regression was found with any other testcases too.
Please review and comment. Thanks a lot!
2005-10-18 Wu Zhou <woodzltc@cn.ibm.com>
* dfp-exprs.exp: New testcase to evaluate simple DFP expressions.
* dfp-test.exp: New testcase to verify the printing and setting of
DFP values.
* dfp-test.c: New source file.
Index: gdb.base/dfp-exprs.exp
===================================================================
RCS file: gdb.base/dfp-exprs.exp
diff -N gdb.base/dfp-exprs.exp
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- gdb.base/dfp-exprs.exp 18 Oct 2005 07:52:59 -0000
***************
*** 0 ****
--- 1,65 ----
+ # Copyright (C) 2005 Free Software Foundation, Inc.
+
+ # This program is free software; you can redistribute it and/or modify
+ # it under the terms of the GNU General Public License as published by
+ # the Free Software Foundation; either version 2 of the License, or
+ # (at your option) any later version.
+ #
+ # This program is distributed in the hope that it will be useful,
+ # but WITHOUT ANY WARRANTY; without even the implied warranty of
+ # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ # GNU General Public License for more details.
+ #
+ # You should have received a copy of the GNU General Public License
+ # along with this program; if not, write to the Free Software
+ # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+ # This file was written by Wu Zhou. (woodzltc@cn.ibm.com)
+
+ # This file is part of the gdb testsuite. It contains test for evaluating
+ # simple decimal floating point (DFP) expression.
+
+ if $tracelevel then {
+ strace $tracelevel
+ }
+
+ proc test_dfp_literals_accepted {} {
+
+ # Test various dfp values.
+
+ gdb_test "p 1.2df" " = 1.2df"
+ gdb_test "p -1.2df" " = -1.2df"
+ gdb_test "p 1.234567df" " = 1.234567df"
+ gdb_test "p -1.234567df" " = -1.234567df"
+ gdb_test "p 1234567.df" " = 1234567.df"
+ gdb_test "p -1234567.df" " = -1234567.df"
+
+ gdb_test "p 1.2dd" " = 1.2dd"
+ gdb_test "p -1.2dd" " = -1.2dd"
+ gdb_test "p 1.234567890123456dd" " = 1.234567890123456dd"
+ gdb_test "p -1.234567890123456dd" " = -1.234567890123456dd"
+ gdb_test "p 1234567890123456.dd" " = 1234567890123456.dd"
+ gdb_test "p -1234567890123456.dd" " = -1234567890123456.dd"
+
+ gdb_test "p 1.2dl" " = 1.2dl"
+ gdb_test "p -1.2dl" " = -1.2dl"
+ gdb_test "p 1.234567890123456789012345678901234dl" " = 1.234567890123456789012345678901234dl"
+ gdb_test "p -1.234567890123456789012345678901234dl" " = -1.234567890123456789012345678901234dl"
+ gdb_test "p 1234567890123456789012345678901234.dl" " = 1234567890123456789012345678901234.dl"
+ gdb_test "p -1234567890123456789012345678901234.dl" " = -1234567890123456789012345678901234.dl"
+
+ }
+
+ proc test_arithmetic_expressions {} {
+
+ # Arithmetic operations for DFP types are not yet implemented in GDB.
+
+ }
+
+ # Start with a fresh gdb.
+
+ gdb_exit
+ gdb_start
+ gdb_reinitialize_dir $srcdir/$subdir
+
+ test_dfp_literals_accepted
Index: gdb.base/dfp-test.c
===================================================================
RCS file: gdb.base/dfp-test.c
diff -N gdb.base/dfp-test.c
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- gdb.base/dfp-test.c 18 Oct 2005 07:52:59 -0000
***************
*** 0 ****
--- 1,95 ----
+ /* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2005 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software
+ Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ 02111-1307, USA. */
+
+ #include <stdio.h>
+ #include <stdlib.h>
+
+ volatile _Decimal32 d32;
+ volatile _Decimal64 d64;
+ volatile _Decimal128 d128;
+
+ struct decstruct
+ {
+ int int4;
+ long long8;
+ _Decimal32 dec32;
+ _Decimal64 dec64;
+ _Decimal128 dec128;
+ } ds;
+
+ static _Decimal32
+ arg0_32 (_Decimal32 arg0, _Decimal32 arg1, _Decimal32 arg2,
+ _Decimal32 arg3, _Decimal32 arg4, _Decimal32 arg5)
+ {
+ return arg0;
+ }
+
+ static _Decimal64
+ arg0_64 (_Decimal64 arg0, _Decimal64 arg1, _Decimal64 arg2,
+ _Decimal64 arg3, _Decimal64 arg4, _Decimal64 arg5)
+ {
+ return arg0;
+ }
+
+ static _Decimal128
+ arg0_128 (_Decimal128 arg0, _Decimal128 arg1, _Decimal128 arg2,
+ _Decimal128 arg3, _Decimal128 arg4, _Decimal128 arg5)
+ {
+ return arg0;
+ }
+
+ int main()
+ {
+ /* An finite 32-bits decimal floating point. */
+ d32 = 1.2345df; /* Initialize d32. */
+
+ /* Non-finite 32-bits decimal floating point: infinity and NaN. */
+ d32 = __builtin_infd32(); /* Positive infd32. */
+ d32 = -__builtin_infd32(); /* Negative infd32. */
+ d32 = __builtin_nand32("");
+
+ /* An finite 64-bits decimal floating point. */
+ d64 = 1.2345dd; /* Initialize d64. */
+
+ /* Non-finite 64-bits decimal floating point: infinity and NaN. */
+ d64 = __builtin_infd64(); /* Positive infd64. */
+ d64 = -__builtin_infd64(); /* Negative infd64. */
+ d64 = __builtin_nand64("");
+
+ /* An finite 128-bits decimal floating point. */
+ d128 = 1.2345dl; /* Initialize d128. */
+
+ /* Non-finite 128-bits decimal floating point: infinity and NaN. */
+ d128 = __builtin_infd128(); /* Positive infd128. */
+ d128 = -__builtin_infd128(); /* Negative infd128. */
+ d128 = __builtin_nand128("");
+
+ /* Functions with decimal floating point as parameter and return value. */
+ d32 = arg0_32 (0.1df, 1.0df, 2.0df, 3.0df, 4.0df, 5.0df);
+ d64 = arg0_64 (0.1dd, 1.0dd, 2.0dd, 3.0dd, 4.0dd, 5.0dd);
+ d128 = arg0_128 (0.1dl, 1.0dl, 2.0dl, 3.0dl, 4.0dl, 5.0dl);
+
+ ds.int4 = 1;
+ ds.long8 = 2;
+ ds.dec32 = 1.2345df;
+ ds.dec64 = 1.2345dd;
+ ds.dec128 = 1.2345dl;
+
+ return 0; /* Exit point. */
+ }
Index: gdb.base/dfp-test.exp
===================================================================
RCS file: gdb.base/dfp-test.exp
diff -N gdb.base/dfp-test.exp
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- gdb.base/dfp-test.exp 18 Oct 2005 07:53:00 -0000
***************
*** 0 ****
--- 1,243 ----
+ # Copyright 2005 Free Software Foundation, Inc.
+
+ # This program is free software; you can redistribute it and/or modify
+ # it under the terms of the GNU General Public License as published by
+ # the Free Software Foundation; either version 2 of the License, or
+ # (at your option) any later version.
+ #
+ # This program is distributed in the hope that it will be useful,
+ # but WITHOUT ANY WARRANTY; without even the implied warranty of
+ # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ # GNU General Public License for more details.
+ #
+ # You should have received a copy of the GNU General Public License
+ # along with this program; if not, write to the Free Software
+ # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+ # This file was written by Wu Zhou. (woodzltc@cn.ibm.com)
+
+ # This file is part of the gdb testsuite. It is intended to test that
+ # gdb could correctly handle decimal floating point introduced in IEEE 754R.
+
+ proc d32_set_tests {} {
+
+ gdb_test "p d32=123.45df" " = 123.45df"
+ gdb_test "p d32=12345.df" " = 12345.df"
+ gdb_test "p d32=12345.67df" " = 12345.67df"
+ gdb_test "p d32=1234567.df" " = 1234567.df"
+
+ gdb_test "p d32=1.234567E0df" " = 1.234567df"
+ gdb_test "p d32=1.234567E10df" " = 1.234567E\\+10df"
+ gdb_test "p d32=1.234567E+96df" " = 1.234567E\\+96df"
+
+ # Test that gdb could handle the max, normalized min and subnormalized min.
+ gdb_test "p d32=9.999999E+96df" " = 9.999999E\\+96df"
+ gdb_test "p d32=1.0E-95df" " = 1.E\\-95df"
+ gdb_test "p d32=1.E-101df" " = 1.E\\-101df"
+ gdb_test "p d32=0.000001E-95df" " = 1.E\\-101df"
+
+ # Test that gdb could detect coefficient/exponent out of range.
+ gdb_test "p d32=1.2345678df" ".*coefficient is out of range.*" "1.2345678df is out of range (coefficient)"
+ gdb_test "p d32=1.0E-101df" ".*exponent is out of range.*" "1.0E-101df is out of range (exponent)"
+ gdb_test "p d32=1.234567E+97df" ".*exponent is out of range.*" "1.234567E+97df is out of range (exponent)"
+
+ # Test that gdb could detect the errors in the string representation of _Decimal32
+ gdb_test "p d32=12345df" ".*Invalid number.*" "12345df is an invalid number"
+ gdb_test "p d32=1.23Edf" ".*Digit expected.*" "p d32=1.23Edf"
+ gdb_test "p d32=1.23E45Adf" ".*Invalid char found.*" "p d32=1.23E45Adf"
+ }
+
+ proc d64_set_tests {} {
+
+ gdb_test "p d64=123.45dd" " = 123.45dd"
+ gdb_test "p d64=12345.dd" " = 12345.dd"
+ gdb_test "p d64=12345.67dd" " = 12345.67dd"
+ gdb_test "p d64=1.234567890123456dd" " = 1.234567890123456dd"
+
+ gdb_test "p d64=1.234567890123456E10dd" " = 12345678901.23456dd"
+ gdb_test "p d64=1.234567890123456E100dd" " = 1.234567890123456E\\+100dd"
+ gdb_test "p d64=1.234567890123456E384dd" " = 1.234567890123456E\\+384dd"
+
+ # Test that gdb could handle the max, normalized min and subnormalized min.
+ gdb_test "p d64=9.999999999999999E384dd" " = 9.999999999999999E\\+384dd"
+ gdb_test "p d64=1.E-383dd" " = 1.E\\-383dd"
+ gdb_test "p d64=1.E-398dd" " = 1.E\\-398dd"
+ gdb_test "p d64=0.000000000000001E-383dd" " = 1.E\\-398dd"
+
+ # Test that gdb could detect coefficient/exponent out of range.
+ gdb_test "p d64=1.2345678901234567dd" ".*coefficient is out of range.*" "1.2345678901234567dd is out of range (coefficient)"
+ gdb_test "p d64=9.9999999999999999E384dd" ".*coefficient is out of range.*" "d64=9.9999999999999999E384dd is out of range (coefficient)"
+ gdb_test "p d64=1.234567890123456E385dd" ".*exponent is out of range.*" "d64=1.234567890123456E385dd is out of range (exponent)"
+ gdb_test "p d64=1.E-399dd" ".*exponent is out of range.*" "d64=1.E-399dd is out of range (exponent)"
+
+ # Test that gdb could detect the errors in the string representation of _Decimal32
+ gdb_test "p d64=12345dd" ".*Invalid number.*" "12345dd is an invalid number"
+ gdb_test "p d64=1.23Edd" ".*Digit expected.*" "p d64=1.23Edd"
+ gdb_test "p d64=1.23E45Add" ".*Invalid char found.*" "p d64=1.23E45Add"
+ }
+
+ proc d128_set_tests {} {
+
+ gdb_test "p d128=123.45dl" " = 123.45dl"
+ gdb_test "p d128=12345.dl" " = 12345.dl"
+ gdb_test "p d128=12345.67dl" " = 12345.67dl"
+ gdb_test "p d128=1.234567890123456789012345678901234dl" "= 1.234567890123456789012345678901234dl"
+
+ gdb_test "p d128=1.234567890123456E10dl" " = 12345678901.23456dl"
+ gdb_test "p d128=1.234567890123456E100dl" " = 1.234567890123456E\\+100dl"
+ gdb_test "p d128=1.234567890123456E1000dl" " = 1.234567890123456E\\+1000dl"
+
+ # Test that gdb could handle the max, normalized min and subnormalized min.
+ gdb_test "p d128=9.99999999999999999999999999999999E6144dl" " = 9.99999999999999999999999999999999E\\+6144dl"
+ gdb_test "p d128=1.E-6143dl" " = 1.E\\-6143dl"
+ gdb_test "p d128=1.E-6176dl" " = 1.E\\-6176dl"
+ gdb_test "p d128=0.000000000000000000000000000000001E-6143dl" " = 1.E\\-6176dl"
+
+ # Test that gdb could detect coefficient/exponent out of range.
+ gdb_test "p d128=1.2345678901234567890123456789012345dl" ".*coefficient is out of range.*" "1.2345678901234567890123456789012345dl is out of range (coefficient)"
+ gdb_test "p d128=9.9999999999999999999999999999999999dl" ".*coefficient is out of range.*" "d128=9.9999999999999999999999999999999999 is out of range (coefficient)"
+ gdb_test "p d128=1.234567890123456E6145dl" ".*exponent is out of range.*" "d128=1.234567890123456E6145dl is out of range (exponent)"
+ gdb_test "p d128=1.E-6177dl" ".*exponent is out of range.*" "d128=1.E-6177dl is out of range (exponent)"
+
+ # Test that gdb could detect the errors in the string representation of _Decimal32
+ gdb_test "p d128=12345dl" ".*Invalid number.*" "12345dl is an invalid number"
+ gdb_test "p d128=1.23Edl" ".*Digit expected.*" "p d128=1.23Edl"
+ gdb_test "p d128=1.23E45Adl" ".*Invalid char found.*" "p d128=1.23E45Adl"
+ }
+
+
+ if $tracelevel {
+ strace $tracelevel
+ }
+
+ set testfile "dfp-test"
+ set srcfile ${testfile}.c
+ set binfile ${objdir}/${subdir}/${testfile}
+ if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable {debug}] != "" } {
+ untested "Couldn't compile ${srcfile}"
+ return -1
+ }
+
+ gdb_exit
+ gdb_start
+ gdb_reinitialize_dir $srcdir/$subdir
+ gdb_load ${binfile}
+
+ if ![runto_main] then {
+ perror "couldn't run to breakpoint"
+ continue
+ }
+
+ # Different tests on 32-bits decimal floating point, including the printing
+ # of finite numbers, infinite and NaN, and also the setting of different
+ # decimal value.
+
+ if [gdb_test "next" \
+ ".*Positive infd32.*" \
+ "next after initializing d32"] then { gdb_suppress_tests }
+ gdb_test "print d32" "1.2345df" "d32 is initialized to 1.2345df"
+
+ if [gdb_test "next" \
+ ".*Negative infd32.*" \
+ "next after assigning builtin infinity to d32"] then { gdb_suppress_tests }
+ gdb_test "print d32" "Infinity" "d32 is positive Infinity"
+
+ if [gdb_test "next" \
+ ".*__builtin_nand32.*" \
+ "next after assigning negative builtin infinity to d32"] then { gdb_suppress_tests }
+ gdb_test "print d32" "-Infinity" "d32 is negative Infinity"
+
+ if [gdb_test "next" \
+ ".*d64 = 1.2345dd.*" \
+ "next after assigning builtin NaN to d32"] then { gdb_suppress_tests }
+ gdb_test "print d32" "NaN" "d32 is NaN"
+
+ d32_set_tests
+
+
+ # Different tests on 64-bits decimal floating point, including the display
+ # of finite number, infinite and NaN, and also the setting of different
+ # decimal value.
+
+ if [gdb_test "next" \
+ ".*Positive infd64.*" \
+ "next after initializing d64"] then { gdb_suppress_tests }
+ gdb_test "print d64" "1.2345dd" "d64 is initialized to 1.2345dd"
+
+ if [gdb_test "next" \
+ ".*Negative infd64.*" \
+ "next after assigning builtin infinity to d64"] then { gdb_suppress_tests }
+ gdb_test "print d64" "Infinity" "d64 is positive Infinity"
+
+ if [gdb_test "next" \
+ ".*__builtin_nand64.*" \
+ "next after assigning negative builtin infinity to d64"] then { gdb_suppress_tests }
+ gdb_test "print d64" "-Infinity" "d64 is negative Infinity"
+
+ if [gdb_test "next" \
+ ".*d128 = 1.2345dl.*" \
+ "next after assigning builtin NaN to d64"] then { gdb_suppress_tests }
+ gdb_test "print d64" "NaN" "d64 is NaN"
+
+ d64_set_tests
+
+
+ # Different tests on 128-bits decimal floating point, including the display
+ # of finite number, infinite and NaN, and also the setting of different
+ # decimal value.
+
+ if [gdb_test "next" \
+ ".*Positive infd128.*" \
+ "next after initializing d128"] then { gdb_suppress_tests }
+ gdb_test "print d128" "1.2345dl" "d128 is initialized to 1.2345dl"
+
+ d128_set_tests
+
+ if [gdb_test "next" \
+ ".*Negative infd128.*" \
+ "next after assigning builtin infinity to d128"] then { gdb_suppress_tests }
+ gdb_test "print d128" "Infinity" "d128 is positive Infinity"
+
+ if [gdb_test "next" \
+ ".*__builtin_nand128.*" \
+ "next after assigning negative builtin infinity to d128"] then { gdb_suppress_tests }
+ gdb_test "print d128" "-Infinity" "d128 is negative Infinity"
+
+ if [gdb_test "next" \
+ ".*arg0_32.*" \
+ "next after assigning builtin NaN to d128"] then { gdb_suppress_tests }
+ gdb_test "print d128" "NaN" "d128 is NaN"
+
+ # The following tests are intended to verify that gdb can correctly handle
+ # DFP types in function arguments.
+
+ gdb_breakpoint arg0_32
+ gdb_continue_to_breakpoint "entry to arg0_32"
+ gdb_test "backtrace" ".*arg0_32 \\(arg0=0.1df, arg1=1.df, arg2=2.df, arg3=3.df, arg4=4.df, arg5=5.df\\).*" "backtrace at arg0_32"
+
+ gdb_breakpoint arg0_64
+ gdb_continue_to_breakpoint "entry to arg0_64"
+ gdb_test "backtrace" ".*arg0_64 \\(arg0=0.1dd, arg1=1.dd, arg2=2.dd, arg3=3.dd, arg4=4.dd, arg5=5.dd\\).*" "backtrace at arg0_64"
+
+ gdb_breakpoint arg0_128
+ gdb_continue_to_breakpoint "entry to arg0_128"
+ gdb_test "backtrace" ".*arg0_128 \\(arg0=0.1dl, arg1=1.dl, arg2=2.dl, arg3=3.dl, arg4=4.dl, arg5=5.dl\\).*" "backtrace at arg0_128"
+
+ # The following tests are intended to verify that gdb can handle DFP types
+ # correctly in struct.
+
+ gdb_breakpoint [gdb_get_line_number "Exit point"]
+ gdb_continue_to_breakpoint "Setting a decimal struct"
+ gdb_test "print ds.dec32" " = 1.2345df"
+ gdb_test "print ds.dec64" " = 1.2345dd"
+ gdb_test "print ds.dec128" " = 1.2345dl"
+
+ # The following tests are intended to verify that gdb can handle "dfp1=dfp2"
+ # and "dfp1=-dfp2" correctly.
+
+ gdb_test "print ds.dec32=d32" " = 0.1df"
+ gdb_test "print ds.dec64=d64" " = 0.1dd"
+ gdb_test "print ds.dec128=d128" " = 0.1dl"
+ gdb_test "print ds.dec32 = -d32" " = -0.1df"
+ gdb_test "print ds.dec64 = -d64" " = -0.1dd"
+ gdb_test "print ds.dec128 = -d128" " = -0.1dl"
Best Regards
- Wu Zhou
On Sun, 2 Oct 2005, Daniel Jacobowitz wrote:
> On Wed, Sep 28, 2005 at 01:28:07PM +0800, Wu Zhou wrote:
> > In this testcase, there are totally 83 tests. Most of them are about
> > displaying and setting of dfp types, which are partitioned into
> > three groups, one for each dfp types (_Decimal32, _Decimal64 and
> > _Decimal128). In each group, we first begin with the printing of
> > a normal finite dfp value, then Infinity and NaN. Then we call
> > d32_set_tests (d64_set_tests or d128_set_tests) to test that GDB
> > could handle the setting different values of variable d32 (d64 or
> > d128).
> >
> > At the end of the testcase, we test that GDB could print out the dfp
> > values correctly in the function arguments and backtrace.
>
> Just a few comments on the tests:
>
> - This should be at least two separate tests, since the "p 1.5df" tests
> can be run without a GCC which supports DFP, even though dfp-test.c
> can't be compiled.
>
> - C files need to have copyright headers.
>
> In this:
>
> > + gdb_test "p d32=123.45df" "123.45df" "p d32=123.45df"
>
> - The third argument to gdb_test defaults to the first argument, so you
> don't need it.
>
> - The second argument should be more specific. Something like
> "\$${decimal} = 123.45df". I went to look at this test to see if my
> understanding of the use of string types was right, and the matching
> patterns were so loose I couldn't tell.
>
> Otherwise I'll wait to look at this further until we're agreed on the
> GDB support.
>
> --
> Daniel Jacobowitz
> CodeSourcery, LLC
>
>