[1/2] mingw: update gnulib: prepare the sources

From: Pedro Alves <palves@redhat.com>

  Hi guys,

We really need to move forward with importing a newer gnulib,
for several reasons...

On 12/24/2014 10:20 PM, Jan Kratochvil wrote:
> Hi Kai,
> 
> On Tue, 23 Dec 2014 12:54:50 +0100, Kai Tietz wrote:
>>> > > The whole problem is that the gnulib update (in [patch 2/2]) will cause
>>> > > (only) for build_win64 many errors like this one:
>> > Only for win64?  This is curious.  As headers are shared between 32-bit and
>> > 64-bit, and most part of Win32 API too.
> Yes, only win64 - because the change making struct timeval incompatible with
> select()'s argument
> 	http://sourceforge.net/p/mingw-w64/mailman/message/29610438/
> has there
> 	+#ifdef __LP64__
> 
> 
>> > Well, IMO the real issue here is to include windows.h instead of sys/time.h.
> The patch is now different so I am not sure if it still applies and what is
> the reason for it.
> 
> <sys/time.h> includes <winsock2.h> since:
> gnulib:
> https://lists.gnu.org/archive/html/bug-gnulib/2012-01/msg00372.html
> commit f8e84098084b3b53bc6943a5542af1f607ffd477
> Author: Bruno Haible <bruno@clisp.org>
> Date:   Sat Jan 28 18:12:10 2012 +0100
>     sys_time: Override 'struct timeval' on some native Windows platforms.
> 
> 
>> > Sadly MS decided to pollute namespace with commonly used names in their
>> > platform-headers.  So in the past we tried to avoid to include things like
>> > windows.h in bfd and related headers, and to isolate files, which actually
>> > need to include it (eg windows-unicode stuff in binutils/).  To include it
>> > now leads to a general architectural change of binutils' bfd (and other
>> > parts).  I am not in general oppose to it, but I am wondering if this is
>> > really wanted/needed.
> Both options are ugly but following the current include/coff/ assumption makes
> the patch smaller.
> 
> It builds now (not runtime tested) for mingw64 32-bit and 64-bit.
> 
> No regressions on {x86_64,x86_64-m32}-fedora21-linux-gnu.

I don't think this is correct, however.

> --- ./gdb/utils.c	2014-12-24 22:50:08.455459138 +0100
> +++ ./gdb/utils.c	2014-12-24 22:34:08.904812278 +0100
> @@ -37,8 +37,11 @@
>  #include <pc.h>
>  #endif
>  
> -#include <signal.h>
> +/* For struct timeval for timeval-utils.h.  */
> +#include <sys/time.h>
>  #include "timeval-utils.h"
> +
> +#include <signal.h>
>  #include "gdbcmd.h"
>  #include "serial.h"
>  #include "bfd.h"
> @@ -55,6 +58,7 @@
>  #include "top.h"
>  #include "main.h"
>  #include "solist.h"
> +#include "gdb_timeval.h"
>  
>  #include "inferior.h"		/* for signed_pointer_to_address */
>  

For example, further down the file we have:

static int ATTRIBUTE_PRINTF (1, 0)
defaulted_query (const char *ctlstr, const char defchar, va_list args)
{
...
  struct timeval prompt_started, prompt_ended, prompt_delta;
...
  /* Add time spend in this routine to prompt_for_continue_wait_time.  */
  gettimeofday (&prompt_ended, NULL);
  timeval_sub (&prompt_delta, &prompt_ended, &prompt_started);
  timeval_add (&prompt_for_continue_wait_time,
               &prompt_for_continue_wait_time, &prompt_delta);

and with this change above:

> -#include <signal.h>
> +/* For struct timeval for timeval-utils.h.  */
> +#include <sys/time.h>
>  #include "timeval-utils.h"

timeval_sub and timeval_add in "timeval-utils.h" will be prototyped
as if taking gnulib's "struct timeval" replacement, but, those functions
expect to be passed the native "struct timeval", as gnulib's sys/time.h
isn't used when libiberty itself is built.

The only thing gnulib's gettimeofday does is type-convert
the native struct timeval to the replaced type, like this:

int
gettimeofday (struct timeval *restrict tv, void *restrict tz)
{
#undef gettimeofday
...
# if defined timeval /* 'struct timeval' overridden by gnulib?  */
#  undef timeval
  struct timeval otv;
  int result = gettimeofday (&otv, (struct timezone *) tz);
  if (result == 0)
    {
      tv->tv_sec = otv.tv_sec;
      tv->tv_usec = otv.tv_usec;
    }
...

that is, other than POSIX compliance, gnulib's struct timeval
replacement doesn't really give us "real" 64-bit-aware gettimeofday.

So, how about this simpler approach below?  Just make sure we
continue using the native struct timeval, as today.

From 5b35653dbdd66fc91b6222e2b9147fe2f89e0d86 Mon Sep 17 00:00:00 2001
From: Pedro Alves <palves@redhat.com>
Date: Mon, 23 Mar 2015 15:20:09 +0000
Subject: [PATCH] Prepare for gnulib update

After the last gnulib import (Dec 2012), gnulib upstream started
replacing mingw's 'struct timeval' with a version with 64-bit time_t,
for POSIX compliance:

 commit f8e84098084b3b53bc6943a5542af1f607ffd477
 Author: Bruno Haible <bruno@clisp.org>
 Date:   Sat Jan 28 18:12:10 2012 +0100
     sys_time: Override 'struct timeval' on some native Windows platforms.

See:

 https://lists.gnu.org/archive/html/bug-gnulib/2012-01/msg00372.html

However, that results in conflicts with native Winsock2's 'select':

select()'s argument
	http://sourceforge.net/p/mingw-w64/mailman/message/29610438/

... and libiberty's timeval-utils.h timeval_add/timeval_sub, at the
least.

We don't really need the POSIX compliance, so this patch prepares us
to simply not use gnulib's 'struct timeval' replacement once a more
recent gnulib is imported, thus preserving the current behavior, by
adding a sys/time.h wrapper header that undefs gnulib's replacements,
and including that everywhere instead.

The SIZE -> OSIZE change is necessary because newer gnulib's
sys/time.h also includes windows.h/winsock2.h, which defines a
conflicting SIZE symbol.

Cross build-tested mingw-w64 32-bit and 64-bit.
Regtested on x86_64 Fedora 20.

gdb/ChangeLog:
2015-03-23  Pedro Alves  <palves@redhat.com>

	* Makefile.in (HFILES_NO_SRCDIR): Add common/gdb_sys_time.h.
	* common/gdb_sys_time.h: New file.
	* dsrec.c: Include gdb_sys_time.h instead of sys/time.h.
	* event-loop.c: Likewise.
	* gdb_select.h: Likewise.
	* gdb_usleep.c: Likewise.
	* m32r-rom.c: Likewise.
	* maint.c: Likewise.
	* mi/mi-main.c: Likewise.
	* mi/mi-parse.h: Likewise.
	* remote-fileio.c: Likewise.
	* remote-m32r-sdi.c: Likewise.
	* remote.c: Likewise.
	* ser-base.c: Likewise.
	* ser-pipe.c: Likewise.
	* ser-tcp.c: Likewise.
	* ser-unix.c: Likewise.
	* symfile.c: Likewise.
	* symfile.c: Likewise.  Rename OSIZE to SIZE throughout.
	* target-memory.c: Include gdb_sys_time.h instead of sys/time.h.
	* utils.c: Likewise.

gdb/gdbserver/ChangeLog:
2015-03-23  Pedro Alves  <palves@redhat.com>

	* debug.c: Include gdb_sys_time.h instead of sys/time.h.
	* gdbserver/event-loop.c: Likewise.
	* gdbserver/remote-utils.c: Likewise.
	* gdbserver/tracepoint.c: Likewise.
---
 gdb/Makefile.in              |  2 +-
 gdb/common/gdb_sys_time.h    | 38 ++++++++++++++++++++++++++++++++++++++
 gdb/dsrec.c                  |  2 +-
 gdb/event-loop.c             |  2 +-
 gdb/gdb_select.h             |  2 +-
 gdb/gdb_usleep.c             |  3 +--
 gdb/gdbserver/debug.c        |  2 +-
 gdb/gdbserver/event-loop.c   |  2 +-
 gdb/gdbserver/remote-utils.c |  2 +-
 gdb/gdbserver/tracepoint.c   |  2 +-
 gdb/m32r-rom.c               |  2 +-
 gdb/maint.c                  |  2 +-
 gdb/mi/mi-main.c             |  2 +-
 gdb/mi/mi-parse.h            |  2 +-
 gdb/remote-fileio.c          |  2 +-
 gdb/remote-m32r-sdi.c        |  2 +-
 gdb/remote.c                 |  2 +-
 gdb/ser-base.c               |  2 +-
 gdb/ser-pipe.c               |  2 +-
 gdb/ser-tcp.c                |  2 +-
 gdb/ser-unix.c               |  2 +-
 gdb/symfile.c                | 14 +++++++-------
 gdb/target-memory.c          |  2 +-
 gdb/utils.c                  |  2 +-
 24 files changed, 67 insertions(+), 30 deletions(-)
 create mode 100644 gdb/common/gdb_sys_time.h

[1/2] mingw: update gnulib: prepare the sources

Commit Message

Comments

Patch