[Bug c/37267] New: #pragma inside structure initialization causes error

[Bug c/37267] New: #pragma inside structure initialization causes error

[neale at sinenomine dot net]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1329 bytes --]

We’ve implemented the #pragma ident which simply generates information in the
comment section of an elf object. Solaris uses it throughout its source tree.
However, it’s causing some grief when used in a manner like this:

#include <stdio.h>

typedef struct X {
        int a;
        int b;
} X_t;

int z;

X_t x[] = {
#pragma weak Y = z
{0,1},
{2,3}
};

int 
main(int argc, char **argv) 
{
        printf("%x",x[0].a);
        return x[1].b;
}


I’ve replaced ident with weak so it will run on other platforms. For gcc 4.0.1
there’s no problem but with 4.2.3 I get the following:

/tmp/test.c:11: error: expected expression before '#pragma'

The context in which I am seeing a problem is that the C code does something
like this:

tuple_t x[] = {
#include “a.h”
#include “b.h”
};

Each of the ‘.h’ files have the contents of the structure but they also contain
a #pragma ident which causes the compiler to emit the error.

Neale


-- 
           Summary: #pragma inside structure initialization causes error
           Product: gcc
           Version: 4.2.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: neale at sinenomine dot net


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2008-09-02 20:53 -------
Confirmed, this is most likely caused by the pragma changes in 4.2 as 4.1.1
works.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |rejects-valid
      Known to fail|                            |4.2.0 4.3.0 4.4.0
      Known to work|                            |4.1.1
   Last reconfirmed|0000-00-00 00:00:00         |2008-09-02 20:53:30
               date|                            |
            Summary|#pragma inside structure    |[4.2/4.3/4.4 Regression]
                   |initialization causes error |#pragma inside structure
                   |                            |initialization causes error
   Target Milestone|---                         |4.2.5


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[jakub at gcc dot gnu dot org]

------- Comment #2 from jakub at gcc dot gnu dot org  2008-09-03 09:41 -------
Similar to PR25246.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[jakub at gcc dot gnu dot org]

------- Comment #3 from jakub at gcc dot gnu dot org  2008-09-03 09:47 -------
But I'm not convinced we want to allow pragmas in this context.  While these
pragmas aren't covered by the C standard, for all the covered pragmas in there
they must appear outside of external declarations or explicit declarations and
statements inside a compound statement.  While there was a point in handling
some pragmas inside of structure/union definition, pragmas at this spot don't
serve anything useful and allowing it here would mean
basically we'd need to handle pragmas almost everywhere in the grammar.


-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rth at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[rguenth at gcc dot gnu dot org]

------- Comment #4 from rguenth at gcc dot gnu dot org  2008-09-20 14:35 -------
I agree with Jakub.  Joseph - are you ok with closing this as invalid?


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jsm28 at gcc dot gnu dot org
             Status|NEW                         |WAITING
           Keywords|rejects-valid               |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[joseph at codesourcery dot com]

------- Comment #5 from joseph at codesourcery dot com  2008-09-21 16:12 -------
Subject: Re:  [4.2/4.3/4.4 Regression] #pragma inside structure
 initialization causes error

On Sat, 20 Sep 2008, rguenth at gcc dot gnu dot org wrote:

> I agree with Jakub.  Joseph - are you ok with closing this as invalid?

Yes.

I think back ends should be able to register non-deferred pragmas, which 
should be allowed everywhere (and wouldn't reach the parser at all), like 
the "ident" pragma of the motivating case for this PR, but it's for the 
person adding such a pragma to add the facilities required.  #pragma weak 
is one I think it is entirely appropriate to allow only at limited places 
in the grammar rather than trying to define what it means in the middle 
of a declaration.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267

[Bug c/37267] [4.2/4.3/4.4 Regression] #pragma inside structure initialization causes error

[mmitchel at gcc dot gnu dot org]

------- Comment #6 from mmitchel at gcc dot gnu dot org  2008-10-22 03:50 -------
Closed as INVALID, per Comment #4 and Comment #5.


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37267