[C++ Patch] Fix grokbitfield location

[C++ Patch] Fix grokbitfield location

[Paolo Carlini]

[-- Attachment #1: Type: text/plain, Size: 305 bytes --]

Hi,

as mentioned in one of my last patches, we can now improve this 
location. Note: in the same function there are a few further issues 
which I mean to incrementally fix (eg, the diagnostics for 
warn_if_not_aligned ICEs for unnamed bit-fields). Tested x86_64-linux.

Thanks, Paolo.

////////////////


[-- Attachment #2: CL_grokbitfield1 --]
[-- Type: text/plain, Size: 231 bytes --]

/cp
2018-12-05  Paolo Carlini  <paolo.carlini@oracle.com>

	* decl2.c (grokbitfield): USe DECL_SOURCE_LOCATION in error message.

/testsuite
2018-12-05  Paolo Carlini  <paolo.carlini@oracle.com>

	* g++.dg/parse/bitfield6b.C: New.

[-- Attachment #3: patch_grokbitfield1 --]
[-- Type: text/plain, Size: 868 bytes --]

Index: cp/decl2.c
===================================================================
--- cp/decl2.c	(revision 266818)
+++ cp/decl2.c	(working copy)
@@ -1027,7 +1027,8 @@ grokbitfield (const cp_declarator *declarator,
       && (INDIRECT_TYPE_P (value)
           || !dependent_type_p (TREE_TYPE (value))))
     {
-      error ("bit-field %qD with non-integral type", value);
+      error_at (DECL_SOURCE_LOCATION (value),
+		"bit-field %qD with non-integral type", value);
       return error_mark_node;
     }
 
Index: testsuite/g++.dg/parse/bitfield6b.C
===================================================================
--- testsuite/g++.dg/parse/bitfield6b.C	(nonexistent)
+++ testsuite/g++.dg/parse/bitfield6b.C	(working copy)
@@ -0,0 +1,4 @@
+typedef void a();
+struct A {
+a a1: 1;  // { dg-error "3:bit-field .void A::a1\\(\\). with non-integral type" }
+};

Re: [C++ Patch] Fix grokbitfield location

[Jason Merrill]

On 12/5/18 7:45 AM, Paolo Carlini wrote:
> Hi,
> 
> as mentioned in one of my last patches, we can now improve this 
> location. Note: in the same function there are a few further issues 
> which I mean to incrementally fix (eg, the diagnostics for 
> warn_if_not_aligned ICEs for unnamed bit-fields). Tested x86_64-linux.

As long as we're messing with this diagnostic, let's also print the type 
in question.

Jason

Re: [C++ Patch] Fix grokbitfield location

[Paolo Carlini]

[-- Attachment #1: Type: text/plain, Size: 557 bytes --]

Hi,

On 05/12/18 20:31, Jason Merrill wrote:
> On 12/5/18 7:45 AM, Paolo Carlini wrote:
>> Hi,
>>
>> as mentioned in one of my last patches, we can now improve this 
>> location. Note: in the same function there are a few further issues 
>> which I mean to incrementally fix (eg, the diagnostics for 
>> warn_if_not_aligned ICEs for unnamed bit-fields). Tested x86_64-linux.
> As long as we're messing with this diagnostic, let's also print the 
> type in question.

Agreed. Thus I tested on x86_64-linux the below.

Thanks, Paolo.

/////////////////////



[-- Attachment #2: patch_grokbitfield1_2 --]
[-- Type: text/plain, Size: 2248 bytes --]

Index: cp/decl2.c
===================================================================
--- cp/decl2.c	(revision 266818)
+++ cp/decl2.c	(working copy)
@@ -1016,7 +1016,9 @@ grokbitfield (const cp_declarator *declarator,
 
   if (value == error_mark_node)
     return NULL_TREE; /* friends went bad.  */
-  if (TREE_TYPE (value) == error_mark_node)
+
+  tree type = TREE_TYPE (value);
+  if (type == error_mark_node)
     return value;
 
   /* Pass friendly classes back.  */
@@ -1023,11 +1025,12 @@ grokbitfield (const cp_declarator *declarator,
   if (VOID_TYPE_P (value))
     return void_type_node;
 
-  if (!INTEGRAL_OR_ENUMERATION_TYPE_P (TREE_TYPE (value))
-      && (INDIRECT_TYPE_P (value)
-          || !dependent_type_p (TREE_TYPE (value))))
+  if (!INTEGRAL_OR_ENUMERATION_TYPE_P (type)
+      && (INDIRECT_TYPE_P (value) || !dependent_type_p (type)))
     {
-      error ("bit-field %qD with non-integral type", value);
+      error_at (DECL_SOURCE_LOCATION (value),
+		"bit-field %qD with non-integral type %qT",
+		value, type);
       return error_mark_node;
     }
 
@@ -1048,7 +1051,7 @@ grokbitfield (const cp_declarator *declarator,
       return NULL_TREE;
     }
 
-  if (width && TYPE_WARN_IF_NOT_ALIGN (TREE_TYPE (value)))
+  if (width && TYPE_WARN_IF_NOT_ALIGN (type))
     {
       error ("cannot declare bit-field %qD with %<warn_if_not_aligned%> type",
 	     DECL_NAME (value));
Index: testsuite/g++.dg/parse/bitfield3.C
===================================================================
--- testsuite/g++.dg/parse/bitfield3.C	(revision 266818)
+++ testsuite/g++.dg/parse/bitfield3.C	(working copy)
@@ -5,5 +5,5 @@ typedef void (func_type)();
 
 struct A
 {
-  friend func_type f : 2; /* { dg-error "with non-integral type" } */
+  friend func_type f : 2; /* { dg-error "20:bit-field .void f\\(\\). with non-integral type .func_type." } */
 };
Index: testsuite/g++.dg/parse/bitfield6b.C
===================================================================
--- testsuite/g++.dg/parse/bitfield6b.C	(nonexistent)
+++ testsuite/g++.dg/parse/bitfield6b.C	(working copy)
@@ -0,0 +1,4 @@
+typedef void a();
+struct A {
+a a1: 1;  // { dg-error "3:bit-field .void A::a1\\(\\). with non-integral type .void \\(A::\\)\\(\\)." }
+};

Re: [C++ Patch] Fix grokbitfield location

[Jason Merrill]

On 12/5/18 5:34 PM, Paolo Carlini wrote:
> Hi,
> 
> On 05/12/18 20:31, Jason Merrill wrote:
>> On 12/5/18 7:45 AM, Paolo Carlini wrote:
>>> Hi,
>>>
>>> as mentioned in one of my last patches, we can now improve this 
>>> location. Note: in the same function there are a few further issues 
>>> which I mean to incrementally fix (eg, the diagnostics for 
>>> warn_if_not_aligned ICEs for unnamed bit-fields). Tested x86_64-linux.
>> As long as we're messing with this diagnostic, let's also print the 
>> type in question.
> 
> Agreed. Thus I tested on x86_64-linux the below.
> 
> Thanks, Paolo.
> 
> /////////////////////
> 
> 
> -  if (!INTEGRAL_OR_ENUMERATION_TYPE_P (TREE_TYPE (value))
> -      && (INDIRECT_TYPE_P (value)
> -          || !dependent_type_p (TREE_TYPE (value))))
> +  if (!INTEGRAL_OR_ENUMERATION_TYPE_P (type)
> +      && (INDIRECT_TYPE_P (value) || !dependent_type_p (type)))

Hmm, surely this should be INDIRECT_TYPE_P (type).  OK with that change.

Jason