[RFA,1/2] Fix two regressions in scalar printing

Message ID 871sow9scp.fsf@tromey.com
State New, archived
Headers

Commit Message

Tom Tromey July 31, 2017, 10:03 p.m. UTC
  >>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:

Pedro> Yeah, that seems OK to me GDB-output-wise.  "You get what you ask
Pedro> for".

Here is an updated version of this patch, that (I think) implements what
was discussed in this thread.

I regtested it on the buildbot.

Let me know what you think.

Tom

commit 32d9b636591021a7d2a31ce53da6ba1db9f85689
Author: Tom Tromey <tom@tromey.com>
Date:   Tue Jul 11 06:40:40 2017 -0600

    Fix two regressions in scalar printing
    
    PR gdb/21675 points out a few regressions in scalar printing.
    
    One type of regression is due to not carrying over the old handling of
    floating point printing -- where a format like "/d" causes a floating
    point number to first be cast to a signed integer.  This patch restores
    this behavior.
    
    The other regression is a longstanding bug in print_octal_chars: one of
    the constants was wrong.  This patch fixes the constant and adds static
    asserts to help catch this sort of error.
    
    gdb/ChangeLog
    2017-07-31  Tom Tromey  <tom@tromey.com>
    
            PR gdb/21675
            * valprint.c (LOW_ZERO): Change value to 034.
            (print_octal_chars): Add static_asserts for octal constants.
            * printcmd.c (print_scalar_formatted): Add 'd' case.
    
    gdb/testsuite/ChangeLog
    2017-07-31  Tom Tromey  <tom@tromey.com>
    
            PR gdb/21675:
            * gdb.base/printcmds.exp (test_radices): New function.
            * gdb.dwarf2/var-access.exp: Use p/u, not p/d.
            * gdb.base/sizeof.exp (check_valueof): Use p/d.
            * lib/gdb.exp (get_integer_valueof): Use p/d.
  

Comments

Tom Tromey Aug. 17, 2017, 2:23 a.m. UTC | #1
Tom> Here's the updated patch.

My latest try run on the buildbot shows this regressing again.
I will try to fix it soon.

Tom
  
Jan Kratochvil Dec. 10, 2017, 2:14 p.m. UTC | #2
On Thu, 17 Aug 2017 04:23:49 +0200, Tom Tromey wrote:
> Tom> Here's the updated patch.
> 
> My latest try run on the buildbot shows this regressing again.
> I will try to fix it soon.

I am not sure which regression did you mean here but:

d6382fffde99214ce4aee99a208ddb703c647008 is the first bad commit
commit d6382fffde99214ce4aee99a208ddb703c647008
Author: Tom Tromey <tom@tromey.com>
Date:   Tue Jul 11 06:40:40 2017 -0600
    Fix two regressions in scalar printing

Running /home/jkratoch/redhat/gdb-test-guile/gdb/testsuite/gdb.guile/scm-ports.exp ...
FAIL: gdb.guile/scm-ports.exp: buffered: seek to $sp
FAIL: gdb.guile/scm-ports.exp: buffered: seek to $sp for restore
FAIL: gdb.guile/scm-ports.exp: unbuffered: seek to $sp
FAIL: gdb.guile/scm-ports.exp: unbuffered: seek to $sp for restore

Tested on Fedora Rawhide i386.  It happens also on x86_64 with -m32.


Jan


 guile (print (seek rw-mem-port (value->integer sp-reg) SEEK_SET))^M
-= 4294949960^M
-(gdb) PASS: gdb.guile/scm-ports.exp: buffered: seek to $sp
+= 4294949832^M
+(gdb) FAIL: gdb.guile/scm-ports.exp: buffered: seek to $sp

 guile (print (seek rw-mem-port (value->integer sp-reg) SEEK_SET))^M
-= 4294949960^M
-(gdb) PASS: gdb.guile/scm-ports.exp: buffered: seek to $sp for restore
+= 4294949832^M
+(gdb) FAIL: gdb.guile/scm-ports.exp: buffered: seek to $sp for restore

 guile (print (seek rw-mem-port (value->integer sp-reg) SEEK_SET))^M
-= 4294949960^M
-(gdb) PASS: gdb.guile/scm-ports.exp: unbuffered: seek to $sp
+= 4294949832^M
+(gdb) FAIL: gdb.guile/scm-ports.exp: unbuffered: seek to $sp

 guile (print (seek rw-mem-port (value->integer sp-reg) SEEK_SET))^M
-= 4294949960^M
-(gdb) PASS: gdb.guile/scm-ports.exp: unbuffered: seek to $sp for restore
+= 4294949832^M
+(gdb) FAIL: gdb.guile/scm-ports.exp: unbuffered: seek to $sp for restore
  

Patch

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index dd66a45..afdfb16 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,10 @@ 
+2017-07-31  Tom Tromey  <tom@tromey.com>
+
+	PR gdb/21675
+	* valprint.c (LOW_ZERO): Change value to 034.
+	(print_octal_chars): Add static_asserts for octal constants.
+	* printcmd.c (print_scalar_formatted): Add 'd' case.
+
 2017-07-31  Xavier Roirand  <roirand@adacore.com>
 
 	* solib-darwin.c (DYLD_VERSION_MAX): Increase value.
diff --git a/gdb/printcmd.c b/gdb/printcmd.c
index a8cc052..f5ed513 100644
--- a/gdb/printcmd.c
+++ b/gdb/printcmd.c
@@ -413,7 +413,9 @@  print_scalar_formatted (const gdb_byte *valaddr, struct type *type,
       && (options->format == 'o'
 	  || options->format == 'x'
 	  || options->format == 't'
-	  || options->format == 'z'))
+	  || options->format == 'z'
+	  || options->format == 'd'
+	  || options->format == 'u'))
     {
       LONGEST val_long = unpack_long (type, valaddr);
       converted_float_bytes.resize (TYPE_LENGTH (type));
@@ -427,11 +429,13 @@  print_scalar_formatted (const gdb_byte *valaddr, struct type *type,
     case 'o':
       print_octal_chars (stream, valaddr, len, byte_order);
       break;
+    case 'd':
+      print_decimal_chars (stream, valaddr, len, true, byte_order);
+      break;
     case 'u':
       print_decimal_chars (stream, valaddr, len, false, byte_order);
       break;
     case 0:
-    case 'd':
       if (TYPE_CODE (type) != TYPE_CODE_FLT)
 	{
 	  print_decimal_chars (stream, valaddr, len, !TYPE_UNSIGNED (type),
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 2a777a8..f763661 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,11 @@ 
+2017-07-31  Tom Tromey  <tom@tromey.com>
+
+	PR gdb/21675:
+	* gdb.base/printcmds.exp (test_radices): New function.
+	* gdb.dwarf2/var-access.exp: Use p/u, not p/d.
+	* gdb.base/sizeof.exp (check_valueof): Use p/d.
+	* lib/gdb.exp (get_integer_valueof): Use p/d.
+
 2017-07-26  Yao Qi  <yao.qi@linaro.org>
 
 	* gdb.gdb/unittest.exp: Invoke command
diff --git a/gdb/testsuite/gdb.base/printcmds.exp b/gdb/testsuite/gdb.base/printcmds.exp
index 323ca73..03275c3 100644
--- a/gdb/testsuite/gdb.base/printcmds.exp
+++ b/gdb/testsuite/gdb.base/printcmds.exp
@@ -155,6 +155,13 @@  proc test_float_rejected {} {
     test_print_reject "p 1.1ll"
 }
 
+# Regression test for PR gdb/21675
+proc test_radices {} {
+    gdb_test "print/o 16777211" " = 077777773"
+    gdb_test "print/d 1.5" " = 1\[^.\]"
+    gdb_test "print/u 1.5" " = 1\[^.\]"
+}
+
 proc test_print_all_chars {} {
     global gdb_prompt
 
@@ -981,3 +988,4 @@  test_printf
 test_printf_with_dfp
 test_print_symbol
 test_repeat_bytes
+test_radices
diff --git a/gdb/testsuite/gdb.base/sizeof.exp b/gdb/testsuite/gdb.base/sizeof.exp
index d7ada65..5d89407 100644
--- a/gdb/testsuite/gdb.base/sizeof.exp
+++ b/gdb/testsuite/gdb.base/sizeof.exp
@@ -81,7 +81,7 @@  check_sizeof "long double" ${sizeof_long_double}
 
 proc check_valueof { exp val } {
     gdb_test "next" "" ""
-    gdb_test "p value" " = ${val}" "check valueof \"$exp\""
+    gdb_test "p /d value" " = ${val}" "check valueof \"$exp\""
 }
 
 # Check that GDB and the target agree over the sign of a character.
diff --git a/gdb/testsuite/gdb.dwarf2/var-access.exp b/gdb/testsuite/gdb.dwarf2/var-access.exp
index 8ebad6a..9180c88 100644
--- a/gdb/testsuite/gdb.dwarf2/var-access.exp
+++ b/gdb/testsuite/gdb.dwarf2/var-access.exp
@@ -282,16 +282,16 @@  gdb_test_no_output "set var \$[lindex $regname 0] = 81" \
     "init reg for s2.a"
 gdb_test_no_output "set var \$[lindex $regname 1] = 28" \
     "init reg for s2.c"
-gdb_test "print/d s2" " = \\{a = 81, b = 4, c = 28, d = 5\\}" \
+gdb_test "print/u s2" " = \\{a = 81, b = 4, c = 28, d = 5\\}" \
     "initialized s2 from mem and regs"
 gdb_test_no_output "set var s2.c += s2.a + s2.b - s2.d"
-gdb_test "print/d s2" " = \\{a = 81, b = 4, c = 108, d = 5\\}" \
+gdb_test "print/u s2" " = \\{a = 81, b = 4, c = 108, d = 5\\}" \
     "verify s2.c"
-gdb_test "print/d \$[lindex $regname 1]" " = 108" \
+gdb_test "print/u \$[lindex $regname 1]" " = 108" \
     "verify s2.c through reg"
 gdb_test_no_output "set var s2 = {191, 73, 231, 123}" \
     "re-initialize s2"
-gdb_test "print/d s2"  " = \\{a = 191, b = 73, c = 231, d = 123\\}" \
+gdb_test "print/u s2"  " = \\{a = 191, b = 73, c = 231, d = 123\\}" \
     "verify re-initialized s2"
 
 # Unaligned bitfield access through byte-aligned pieces.
diff --git a/gdb/valprint.c b/gdb/valprint.c
index 1667882..9e216cf 100644
--- a/gdb/valprint.c
+++ b/gdb/valprint.c
@@ -1593,15 +1593,21 @@  print_octal_chars (struct ui_file *stream, const gdb_byte *valaddr,
    */
 #define BITS_IN_OCTAL 3
 #define HIGH_ZERO     0340
-#define LOW_ZERO      0016
+#define LOW_ZERO      0034
 #define CARRY_ZERO    0003
+  static_assert (HIGH_ZERO + LOW_ZERO + CARRY_ZERO == 0xff,
+		 "cycle zero constants are wrong");
 #define HIGH_ONE      0200
 #define MID_ONE       0160
 #define LOW_ONE       0016
 #define CARRY_ONE     0001
+  static_assert (HIGH_ONE + MID_ONE + LOW_ONE + CARRY_ONE == 0xff,
+		 "cycle one constants are wrong");
 #define HIGH_TWO      0300
 #define MID_TWO       0070
 #define LOW_TWO       0007
+  static_assert (HIGH_TWO + MID_TWO + LOW_TWO == 0xff,
+		 "cycle two constants are wrong");
 
   /* For 32 we start in cycle 2, with two bits and one bit carry;
      for 64 in cycle in cycle 1, with one bit and a two bit carry.  */