[RFA-v3] Avoid invalid parameter warnings in C runtime function for mingw built GDB

["Pierre Muller" <pierre.muller@ics-cnrs.unistra.fr>]

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Pedro Alves
> Envoyé : mercredi 14 août 2013 14:02
> À : Pierre Muller
> Cc : 'Tom Tromey'; 'Eli Zaretskii'; gdb-patches@sourceware.org
> Objet : Re: [RFA-v2] Avoid invalid parameter warnings in C runtime
function
> for mingw built GDB
> 
> On 08/14/2013 12:38 PM, Pierre Muller wrote:
> 
> > Is this OK to commit?
> > Maybe some comments on the ChangeLog entry?
> 
> > 	* common/filestuff.c (gdb_fopen_cloexec): Do not try to use "e"
> > 	mode if operating system doesn't know O_CLOEXEC, this allows to
> > 	avoid getting a output debug string warning for mingw hosted
> > 	GDB executables.
> 
> This comment should really be in the sources instead.  That here you'd
> have:
> 
> 	* common/filestuff.c (gdb_fopen_cloexec): Do not try to use "e"
> 	mode if operating system doesn't know O_CLOEXEC.

  I knew that the ChangeLog entry was not what is usually expected...
But your version seems perfect!
 
> and in the source, where you have:
> 
> > +  /* If O_CLOEXEC is zero, the operating system doesn't
> > +     know about close on exec mode "e", so don't even try to use it.
*/
> > +  static int fopen_e_ever_failed = O_CLOEXEC == 0;
> 
> I suggest:
> 
>   /* Probe for "e" support once.  But, if we can tell the operating
>      system doesn't know about close on exec mode "e" without probing,
>      skip it.  E.g., the Windows runtime issues an "Invalid parameter
>      passed to C runtime function" OutputDebugString warning for
>      unknown modes.  Assume that if O_CLOEXEC is zero, then "e" isn't
>      supported.  */
>   static int fopen_e_ever_failed;
  Your comment is much better indeed, but let's still keep the 
static int fopen_e_ever_failed = O_CLOEXEC == 0;
line, which is the only real change to the code...

> --
> Pedro Alves

Thanks Pedro!

Is this third version OK?


2013-08-14  Pierre Muller  <muller@sourceware.org>
	    Tom Tromey  <tromey@redhat.com>
	    Pedro Alves  <palves@redhat.com>

	* common/filestuff.c (gdb_fopen_cloexec): Do not try to use "e"
	mode if operating system doesn't know O_CLOEXEC.


Index: common/filestuff.c
===================================================================
RCS file: /cvs/src/src/gdb/common/filestuff.c,v
retrieving revision 1.7
diff -u -p -r1.7 filestuff.c
--- common/filestuff.c	26 Jun 2013 08:01:55 -0000	1.7
+++ common/filestuff.c	14 Aug 2013 12:07:02 -0000
@@ -311,7 +311,13 @@ FILE *
 gdb_fopen_cloexec (const char *filename, const char *opentype)
 {
   FILE *result = NULL;
-  static int fopen_e_ever_failed;
+  /* Probe for "e" support once.  But, if we can tell the operating
+     system doesn't know about close on exec mode "e" without probing,
+     skip it.  E.g., the Windows runtime issues an "Invalid parameter
+     passed to C runtime function" OutputDebugString warning for
+     unknown modes.  Assume that if O_CLOEXEC is zero, then "e" isn't
+     supported.  */
+  static int fopen_e_ever_failed = O_CLOEXEC == 0;
 
   if (!fopen_e_ever_failed)
     {