[commit] Fix up return values in ctti.exp

[commit] Fix up return values in ctti.exp

[Daniel Jacobowitz]

This test is bit-rotten because it is disabled for GCC.  It still
doesn't pass with GCC, so I left it disabled, but I hope that Keith's
branch will resolve this.

Meanwhile, while trying it with local patches for RealView compiler
support, I discovered that the exp file expects some functions to
return -126 when their return type is "unsigned char".  It should be
130 instead.

Checked in.

-- 
Daniel Jacobowitz
CodeSourcery

2009-11-02  Daniel Jacobowitz  <dan@codesourcery.com>

	* gdb.cp/ctti.exp: Correct return values for unsigned char functions.

---
 gdb/testsuite/gdb.cp/ctti.exp |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Index: gdb-mainline/gdb/testsuite/gdb.cp/ctti.exp
===================================================================
--- gdb-mainline.orig/gdb/testsuite/gdb.cp/ctti.exp	2009-01-02 21:58:04.000000000 -0800
+++ gdb-mainline/gdb/testsuite/gdb.cp/ctti.exp	2009-10-31 07:02:05.000000000 -0700
@@ -112,7 +112,7 @@ gdb_test_multiple "print add<float>(2.25
 }
 
 gdb_test_multiple "print add<unsigned char>('A','A')" "print add<unsigned char>('A','A')" {
-    -re "\\$\[0-9\]+ = -126 .*\r\n$gdb_prompt $" {
+    -re "\\$\[0-9\]+ = 130 .*\r\n$gdb_prompt $" {
 	pass "print add<unsigned char>('A','A')"
     }
     -re "No symbol \"add<unsigned char>\" in current context.\r\n$gdb_prompt $" {
@@ -142,7 +142,7 @@ gdb_test_multiple "print add2<float>(2.2
 }
 
 gdb_test_multiple "print add2<unsigned char>('A','A')" "print add2<unsigned char>('A','A')" {
-    -re "\\$\[0-9]+ = -126 .*$gdb_prompt $" {
+    -re "\\$\[0-9]+ = 130 .*$gdb_prompt $" {
 	pass "print add2<unsigned char>('A','A')"
     }
     -re "No symbol \"add2<unsigned char>\" in current context.\r\n$gdb_prompt $" {
@@ -172,7 +172,7 @@ gdb_test_multiple "print add3<float>(2.2
 }
 
 gdb_test_multiple "print add3<unsigned char>('A','A')" "print add3<unsigned char>('A','A')" {
-    -re "\\$\[0-9]+ = -126 .*$gdb_prompt $" {
+    -re "\\$\[0-9]+ = 130 .*$gdb_prompt $" {
 	pass "print add3<unsigned char>('A','A')"
     }
     -re "No symbol \"add3<unsigned char>\" in current context.\r\n$gdb_prompt $" {
@@ -202,7 +202,7 @@ gdb_test_multiple "print add4<float>(2.2
 }
 
 gdb_test_multiple "print add4<unsigned char>('A','A')" "print add4<unsigned char>('A','A')" {
-    -re "\\$\[0-9]+ = -126 .*$gdb_prompt $" {
+    -re "\\$\[0-9]+ = 130 .*$gdb_prompt $" {
 	pass "print add4<unsigned char>('A','A')"
     }
     -re "No symbol \"add4<unsigned char>\" in current context.\r\n$gdb_prompt $" {

Re: [commit] Fix up return values in ctti.exp

[Keith Seitz]

On 11/02/2009 07:55 AM, Daniel Jacobowitz wrote:
> This test is bit-rotten because it is disabled for GCC.  It still
> doesn't pass with GCC, so I left it disabled, but I hope that Keith's
> branch will resolve this.

I commented out the bit in the test that causes the majority of tests to 
get skipped with gcc and ran ctti.exp through both FSF gdb and my 
expr-cumulative archer branch:

FSF gdb:
Running ../../../src/gdb/testsuite/gdb.cp/ctti.exp ...
FAIL: gdb.cp/ctti.exp: print add<int>(2,2)
FAIL: gdb.cp/ctti.exp: print add<float>(2.25,2.25)
FAIL: gdb.cp/ctti.exp: print add<unsigned char>('A','A')
FAIL: gdb.cp/ctti.exp: print add2<int>(2,2)
FAIL: gdb.cp/ctti.exp: print add2<float>(2.25,2.25)
FAIL: gdb.cp/ctti.exp: print add2<unsigned char>('A','A')
FAIL: gdb.cp/ctti.exp: print add3<int>(2,2)
FAIL: gdb.cp/ctti.exp: print add3<float>(2.25,2.25)
FAIL: gdb.cp/ctti.exp: print add3<unsigned char>('A','A')
FAIL: gdb.cp/ctti.exp: print add4<int>(2,2)
FAIL: gdb.cp/ctti.exp: print add4<float>(2.25,2.25)
FAIL: gdb.cp/ctti.exp: print add4<unsigned char>('A','A')

		=== gdb Summary ===

# of expected passes		4
# of unexpected failures	12
/home/keiths/sources/gdb/linux/gdb/testsuite/../../gdb/gdb version 
7.0.50.20091109-cvs -nw -nx

Archer/expr-cumulative:
Running ../../../archer/gdb/testsuite/gdb.cp/ctti.exp ...

		=== gdb Summary ===

# of expected passes		16
/home/keiths/work/archer/expr-cumulative/linux/gdb/testsuite/../../gdb/gdb 
version  7.0.50.20090917-cvs -nw -nx

I am concentrating my efforts to get the contents of expr-cumulative 
submitted, reviewed, and committed upstream now.

Keith

Re: [commit] Fix up return values in ctti.exp

[Daniel Jacobowitz]

On Mon, Nov 09, 2009 at 03:50:55PM -0800, Keith Seitz wrote:
> On 11/02/2009 07:55 AM, Daniel Jacobowitz wrote:
> >This test is bit-rotten because it is disabled for GCC.  It still
> >doesn't pass with GCC, so I left it disabled, but I hope that Keith's
> >branch will resolve this.
> 
> I commented out the bit in the test that causes the majority of tests
> to get skipped with gcc and ran ctti.exp through both FSF gdb and my
> expr-cumulative archer branch:

> Archer/expr-cumulative:
> Running ../../../archer/gdb/testsuite/gdb.cp/ctti.exp ...
> 
> 		=== gdb Summary ===
> 
> # of expected passes		16

Excellent!!!  I was hoping that would be the case.  If you haven't
already, I suggest enabling it on your branch :-)

-- 
Daniel Jacobowitz
CodeSourcery

Re: [commit] Fix up return values in ctti.exp

[Keith Seitz]

On 11/10/2009 11:14 AM, Daniel Jacobowitz wrote:
> If you haven't already, I suggest enabling it on your branch :-)

I've committed your patch and enabled the test on the 
archer-keiths-expr-cumulative branch. I'll be sure that this gets into 
HEAD at the appropriate time.

Keith