Fix GNU build

Fix GNU build

[Alfred M. Szmidt]

Hey,

gdb changed how it handles warnings a while back, and the gnu-nat.c
code hasn't really been maintained that well over the years so it
caused some minor breakage there.  Here is a patch that fixes all
remaining errors and warnings.

Is anyone maintaining the Hurdy bits in GDB?

2006-06-10  Alfred M. Szmidt  <root@gnu.org>

	* gnu-nat.c: Undefine _process_user_ before including
	<hurd/process_request.h>.
	(gnu_resume): Supply missing argument to error().
	(gnu_read_inferior): Add extra parenthesis around arithmetic
	expression to silence warnings from GCC.
	(gnu_write_inferior): Likewise.
	(gnu_xfer_memory): Changed type of MYADDR to `gdb_byte *'.

Index: gnu-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/gnu-nat.c,v
retrieving revision 1.40
diff -u -r1.40 gnu-nat.c
--- gnu-nat.c	11 May 2006 19:37:52 -0000	1.40
+++ gnu-nat.c	10 Jun 2006 21:37:31 -0000
@@ -45,6 +45,9 @@
 #include <hurd/msg.h>
 #include <hurd/msg_request.h>
 #include <hurd/process.h>
+/* Defined by <hurd/process.h>, but we need forward declarations from
+   <hurd/process_request.h> as well, which the same name.  */
+#undef _process_user_ 
 #include <hurd/process_request.h>
 #include <hurd/signal.h>
 #include <hurd/sigpreempt.h>
@@ -1978,7 +1981,7 @@
     {
       struct proc *thread = inf_tid_to_thread (inf, PIDGET (tid));
       if (!thread)
-	error (_("Can't run single thread id %d: no such thread!"));
+	error (_("Can't run single thread id: no such thread!"), inf->pid);
       inf_debug (inf, "running one thread: %d/%d", inf->pid, thread->tid);
       inf_set_threads_resume_sc (inf, thread, 0);
     }
@@ -2257,7 +2260,7 @@
   if (err)
     return 0;
 
-  err = hurd_safe_copyin (myaddr, (void *) addr - low_address + copied, length);
+  err = hurd_safe_copyin (myaddr, (void *) (addr - low_address + copied), length);
   if (err)
     {
       warning (_("Read from inferior faulted: %s"), safe_strerror (err));
@@ -2312,7 +2315,7 @@
 
   deallocate++;
 
-  err = hurd_safe_copyout ((void *) addr - low_address + copied,
+  err = hurd_safe_copyout ((void *) (addr - low_address + copied),
 			   myaddr, length);
   CHK_GOTO_OUT ("Write to inferior faulted", err);
 
@@ -2445,7 +2448,8 @@
 /* Return 0 on failure, number of bytes handled otherwise.  TARGET
    is ignored. */
 static int
-gnu_xfer_memory (CORE_ADDR memaddr, char *myaddr, int len, int write,
+gnu_xfer_memory (CORE_ADDR memaddr, gdb_byte *myaddr, 
+		 int len, int write,
 		 struct mem_attrib *attrib,
 		 struct target_ops *target)
 {

Re: Fix GNU build

[Alfred M. Szmidt]

Sorry, the ChangeLog email address is wrong, it should be
"ams@gnu.org".  Here is a new correct ChangeLog.

2006-06-10  Alfred M. Szmidt  <ams@gnu.org>

	* gnu-nat.c: Undefine _process_user_ before including
	<hurd/process_request.h>.
	(gnu_resume): Supply missing argument to error().
	(gnu_read_inferior): Add extra parenthesis around arithmetic
	expression to silence warnings from GCC.
	(gnu_write_inferior): Likewise.
	(gnu_xfer_memory): Changed type of MYADDR to `gdb_byte *'.

Re: Fix GNU build

[Daniel Jacobowitz]

On Sat, Jun 10, 2006 at 11:45:54PM +0200, Alfred M. Szmidt wrote:
> Hey,
> 
> gdb changed how it handles warnings a while back, and the gnu-nat.c
> code hasn't really been maintained that well over the years so it
> caused some minor breakage there.  Here is a patch that fixes all
> remaining errors and warnings.
> 
> Is anyone maintaining the Hurdy bits in GDB?

No, I don't think so.

> -	error (_("Can't run single thread id %d: no such thread!"));
> +	error (_("Can't run single thread id: no such thread!"), inf->pid);

Er, is that right?

-- 
Daniel Jacobowitz
CodeSourcery

Re: Fix GNU build

[Alfred M. Szmidt]

   > Is anyone maintaining the Hurdy bits in GDB?

   No, I don't think so.

Where does one go to volunteer?

   > -	error (_("Can't run single thread id %d: no such thread!"));
   > +	error (_("Can't run single thread id: no such thread!"), inf->pid);

   Er, is that right?

No, it isn't.  Thanks for spotting it, a bit late on my side.  Here is
a new patch.

2006-06-10  Alfred M. Szmidt  <ams@gnu.org>

	* gnu-nat.c: Undefine _process_user_ before including
	<hurd/process_request.h>.
	(gnu_resume): Supply missing argument to error().
	(gnu_read_inferior): Add extra parenthesis around arithmetic
	expression to silence warnings from GCC.
	(gnu_write_inferior): Likewise.
	(gnu_xfer_memory): Changed type of MYADDR to `gdb_byte *'.

Index: gnu-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/gnu-nat.c,v
retrieving revision 1.40
diff -u -r1.40 gnu-nat.c
--- gnu-nat.c  11 May 2006 19:37:52 -0000	1.40
+++ gnu-nat.c  10 Jun 2006 21:37:31 -0000
@@ -45,6 +45,9 @@
 #include <hurd/msg.h>
 #include <hurd/msg_request.h>
 #include <hurd/process.h>
+/* Defined by <hurd/process.h>, but we need forward declarations from
+   <hurd/process_request.h> as well, which the same name.  */
+#undef _process_user_ 
 #include <hurd/process_request.h>
 #include <hurd/signal.h>
 #include <hurd/sigpreempt.h>
@@ -1978,7 +1981,7 @@
     {
       struct proc *thread = inf_tid_to_thread (inf, PIDGET (tid));
       if (!thread)
-      error (_("Can't run single thread id %d: no such thread!"));
+      error (_("Can't run single thread id %d: no such thread!"), inf->pid);
       inf_debug (inf, "running one thread: %d/%d", inf->pid, thread->tid);
       inf_set_threads_resume_sc (inf, thread, 0);
     }
@@ -2257,7 +2260,7 @@
   if (err)
     return 0;
 
-  err = hurd_safe_copyin (myaddr, (void *) addr - low_address + copied, length);
+  err = hurd_safe_copyin (myaddr, (void *) (addr - low_address + copied), length);
   if (err)
     {
       warning (_("Read from inferior faulted: %s"), safe_strerror (err));
@@ -2312,7 +2315,7 @@
 
   deallocate++;
 
-  err = hurd_safe_copyout ((void *) addr - low_address + copied,
+  err = hurd_safe_copyout ((void *) (addr - low_address + copied),
       	    myaddr, length);
   CHK_GOTO_OUT ("Write to inferior faulted", err);
 
@@ -2445,7 +2448,8 @@
 /* Return 0 on failure, number of bytes handled otherwise.  TARGET
    is ignored. */
 static int
-gnu_xfer_memory (CORE_ADDR memaddr, char *myaddr, int len, int write,
+gnu_xfer_memory (CORE_ADDR memaddr, gdb_byte *myaddr, 
+		 	     int len, int write,
 			     struct mem_attrib *attrib,
 			     struct target_ops *target)
 {

Re: Fix GNU build

[Daniel Jacobowitz]

On Sun, Jun 11, 2006 at 01:52:13AM +0200, Alfred M. Szmidt wrote:
>    > Is anyone maintaining the Hurdy bits in GDB?
> 
>    No, I don't think so.
> 
> Where does one go to volunteer?

You want the job, it's probably yours - but take a look at the text in
gdb/MAINTAINERS first, please.

Does anyone object to making Alfred maintainer for the Hurd port?


-- 
Daniel Jacobowitz
CodeSourcery

Re: Fix GNU build

[Alfred M. Szmidt]

   [...] but take a look at the text in gdb/MAINTAINERS first, please.

Done.

Re: Fix GNU build

[Jim Blandy]

Daniel Jacobowitz <drow@false.org> writes:
> On Sun, Jun 11, 2006 at 01:52:13AM +0200, Alfred M. Szmidt wrote:
>>    > Is anyone maintaining the Hurdy bits in GDB?
>> 
>>    No, I don't think so.
>> 
>> Where does one go to volunteer?
>
> You want the job, it's probably yours - but take a look at the text in
> gdb/MAINTAINERS first, please.
>
> Does anyone object to making Alfred maintainer for the Hurd port?

I ++good.

Re: Fix GNU build

[Alfred M. Szmidt]

   > Does anyone object to making Alfred maintainer for the Hurd port?

   I ++good.

Is there anything I must/should do from this point on?  (I couldn't
find any specific documents what one needs to do to become a
maintainer of a specific port)

Cheers.

Re: Fix GNU build

[Daniel Jacobowitz]

On Sat, Jun 10, 2006 at 08:00:06PM -0400, Daniel Jacobowitz wrote:
> On Sun, Jun 11, 2006 at 01:52:13AM +0200, Alfred M. Szmidt wrote:
> >    > Is anyone maintaining the Hurdy bits in GDB?
> > 
> >    No, I don't think so.
> > 
> > Where does one go to volunteer?
> 
> You want the job, it's probably yours - but take a look at the text in
> gdb/MAINTAINERS first, please.
> 
> Does anyone object to making Alfred maintainer for the Hurd port?

There was no objection.  Alfred, I've added you to MAINTAINERS.

You should probably get write access to the sources.  Please fill out
the request form (from the front of http://sourceware.org somewhere)
and list me as the approver.

After that you'll be able to commit your gnu-nat.c patch yourself.

-- 
Daniel Jacobowitz
CodeSourcery

2006-07-12  Daniel Jacobowitz  <dan@codesourcery.com>

	* MAINTAINERS: Add Alfred Szmidt for the Hurd.

Index: MAINTAINERS
===================================================================
RCS file: /cvs/src/src/gdb/MAINTAINERS,v
retrieving revision 1.347
diff -u -p -r1.347 MAINTAINERS
--- MAINTAINERS	5 Jul 2006 19:03:46 -0000	1.347
+++ MAINTAINERS	12 Jul 2006 17:35:53 -0000
@@ -332,6 +332,7 @@ their platform.
 AIX			Joel Brobecker		brobecker@adacore.com
 
 djgpp native		Eli Zaretskii		eliz@gnu.org
+GNU Hurd		Alfred M. Szmidt	ams@gnu.org
 MS Windows (NT, '00, 9x, Me, XP) host & native
 			Chris Faylor		cgf@alum.bu.edu
 GNU/Linux/x86 native & host