[PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Joel Brobecker]

[-- Attachment #1: Type: text/plain, Size: 523 bytes --]

I'd like more _initialize_hppa_tdep to the end of the file. This is not
absolutely needed, but it conforms a bit better with the practice I've
seen used for other targets. If nobody objects, I'd like to check this
in tomorrow.

A patch to add the gdbarch framework will shortly follow the checkin.

2002-11-05  Joel Brobecker  <brobecker@gnat.com>

        * hppa-tdep.c (_initialize_hppa_tdep): Move function body
        to end of file, to be more consistent with the pratice followed
        in other targets.

-- 
Joel

[-- Attachment #2: hppa-tdep.c.diff --]
[-- Type: text/plain, Size: 4752 bytes --]

Index: hppa-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/hppa-tdep.c,v
retrieving revision 1.29
diff -c -3 -p -r1.29 hppa-tdep.c
*** hppa-tdep.c	30 Oct 2002 20:42:54 -0000	1.29
--- hppa-tdep.c	6 Nov 2002 02:18:50 -0000
*************** hppa_skip_permanent_breakpoint (void)
*** 4674,4724 ****
    /* We can leave the tail's space the same, since there's no jump.  */
  }
  
- void
- _initialize_hppa_tdep (void)
- {
-   struct cmd_list_element *c;
-   void break_at_finish_command (char *arg, int from_tty);
-   void tbreak_at_finish_command (char *arg, int from_tty);
-   void break_at_finish_at_depth_command (char *arg, int from_tty);
- 
-   tm_print_insn = print_insn_hppa;
- 
-   add_cmd ("unwind", class_maintenance, unwind_command,
- 	   "Print unwind table entry at given address.",
- 	   &maintenanceprintlist);
- 
-   deprecate_cmd (add_com ("xbreak", class_breakpoint, 
- 			  break_at_finish_command,
- 			  concat ("Set breakpoint at procedure exit. \n\
- Argument may be function name, or \"*\" and an address.\n\
- If function is specified, break at end of code for that function.\n\
- If an address is specified, break at the end of the function that contains \n\
- that exact address.\n",
- 		   "With no arg, uses current execution address of selected stack frame.\n\
- This is useful for breaking on return to a stack frame.\n\
- \n\
- Multiple breakpoints at one place are permitted, and useful if conditional.\n\
- \n\
- Do \"help breakpoints\" for info on other commands dealing with breakpoints.", NULL)), NULL);
-   deprecate_cmd (add_com_alias ("xb", "xbreak", class_breakpoint, 1), NULL);
-   deprecate_cmd (add_com_alias ("xbr", "xbreak", class_breakpoint, 1), NULL);
-   deprecate_cmd (add_com_alias ("xbre", "xbreak", class_breakpoint, 1), NULL);
-   deprecate_cmd (add_com_alias ("xbrea", "xbreak", class_breakpoint, 1), NULL);
- 
-   deprecate_cmd (c = add_com ("txbreak", class_breakpoint, 
- 			      tbreak_at_finish_command,
- "Set temporary breakpoint at procedure exit.  Either there should\n\
- be no argument or the argument must be a depth.\n"), NULL);
-   set_cmd_completer (c, location_completer);
-   
-   if (xdb_commands)
-     deprecate_cmd (add_com ("bx", class_breakpoint, 
- 			    break_at_finish_at_depth_command,
- "Set breakpoint at procedure exit.  Either there should\n\
- be no argument or the argument must be a depth.\n"), NULL);
- }
- 
  /* Copy the function value from VALBUF into the proper location
     for a function return.
  
--- 4674,4679 ----
*************** hppa_extract_return_value (struct type *
*** 4768,4770 ****
--- 4723,4771 ----
  		: (4 - TYPE_LENGTH (type)))),
  	    TYPE_LENGTH (type));
  }
+ 
+ void
+ _initialize_hppa_tdep (void)
+ {
+   struct cmd_list_element *c;
+   void break_at_finish_command (char *arg, int from_tty);
+   void tbreak_at_finish_command (char *arg, int from_tty);
+   void break_at_finish_at_depth_command (char *arg, int from_tty);
+ 
+   tm_print_insn = print_insn_hppa;
+ 
+   add_cmd ("unwind", class_maintenance, unwind_command,
+ 	   "Print unwind table entry at given address.",
+ 	   &maintenanceprintlist);
+ 
+   deprecate_cmd (add_com ("xbreak", class_breakpoint, 
+ 			  break_at_finish_command,
+ 			  concat ("Set breakpoint at procedure exit. \n\
+ Argument may be function name, or \"*\" and an address.\n\
+ If function is specified, break at end of code for that function.\n\
+ If an address is specified, break at the end of the function that contains \n\
+ that exact address.\n",
+ 		   "With no arg, uses current execution address of selected stack frame.\n\
+ This is useful for breaking on return to a stack frame.\n\
+ \n\
+ Multiple breakpoints at one place are permitted, and useful if conditional.\n\
+ \n\
+ Do \"help breakpoints\" for info on other commands dealing with breakpoints.", NULL)), NULL);
+   deprecate_cmd (add_com_alias ("xb", "xbreak", class_breakpoint, 1), NULL);
+   deprecate_cmd (add_com_alias ("xbr", "xbreak", class_breakpoint, 1), NULL);
+   deprecate_cmd (add_com_alias ("xbre", "xbreak", class_breakpoint, 1), NULL);
+   deprecate_cmd (add_com_alias ("xbrea", "xbreak", class_breakpoint, 1), NULL);
+ 
+   deprecate_cmd (c = add_com ("txbreak", class_breakpoint, 
+ 			      tbreak_at_finish_command,
+ "Set temporary breakpoint at procedure exit.  Either there should\n\
+ be no argument or the argument must be a depth.\n"), NULL);
+   set_cmd_completer (c, location_completer);
+   
+   if (xdb_commands)
+     deprecate_cmd (add_com ("bx", class_breakpoint, 
+ 			    break_at_finish_at_depth_command,
+ "Set breakpoint at procedure exit.  Either there should\n\
+ be no argument or the argument must be a depth.\n"), NULL);
+ }
+ 

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Daniel Jacobowitz]

On Tue, Nov 05, 2002 at 06:25:55PM -0800, Joel Brobecker wrote:
> I'd like more _initialize_hppa_tdep to the end of the file. This is not
> absolutely needed, but it conforms a bit better with the practice I've
> seen used for other targets. If nobody objects, I'd like to check this
> in tomorrow.
> 
> A patch to add the gdbarch framework will shortly follow the checkin.
> 
> 2002-11-05  Joel Brobecker  <brobecker@gnat.com>
> 
>         * hppa-tdep.c (_initialize_hppa_tdep): Move function body
>         to end of file, to be more consistent with the pratice followed
>         in other targets.

And in return, I'd like to nominate Joel as the HP/PA architecture
maintainer, to reward his bravery in being the first person in some
time willing to touch the port at all (with the exception of a few
daring triages from Andrew :).

Oh, and once you've gotten the PA support cleaned up a bit perhaps we
can get the hppa-linux support properly contributed.  That'd be nice. 
I have a copy of the current patches if you want to see them.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Joel Brobecker]

> And in return, I'd like to nominate Joel as the HP/PA architecture
> maintainer, to reward his bravery in being the first person in some
> time willing to touch the port at all (with the exception of a few
> daring triages from Andrew :).

If I were you, I would tremble... Me looking at this code means that
I'll start asking questions, which in turn means you indirectly working
on it. So I'm sort of embarking you on this boat too! Welcome aboard :-).

> Oh, and once you've gotten the PA support cleaned up a bit perhaps we
> can get the hppa-linux support properly contributed.  That'd be nice. 
> I have a copy of the current patches if you want to see them.

Yes, that would be nice.  This is something we could do, once the
initial multiarching is done.  The trouble for me is that I don't have
access to such a system...

-- 
Joel

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Andrew Cagney]

>> And in return, I'd like to nominate Joel as the HP/PA architecture
>> maintainer, to reward his bravery in being the first person in some
>> time willing to touch the port at all (with the exception of a few
>> daring triages from Andrew :).
> 
> 
> If I were you, I would tremble... Me looking at this code means that
> I'll start asking questions, which in turn means you indirectly working
> on it. So I'm sort of embarking you on this boat too! Welcome aboard :-).

Daniel, Joel, just FYI.

GDB, in the past, was unfortunatly badly burnt by maintainers that 
promised the earth but didn't deliver. As a consequence, the simple rule 
of wait for that first delivery, was adopted.  This is, for instance, 
why architecture maintainers get apointed after their architecture is 
contributed.

I'll follow this up after HP has been multi-arched.

enjoy,
Andrew

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Elena Zannoni]

Joel Brobecker writes:
 > I'd like more _initialize_hppa_tdep to the end of the file. This is not
 > absolutely needed, but it conforms a bit better with the practice I've
 > seen used for other targets. If nobody objects, I'd like to check this
 > in tomorrow.

Sure.  Actually multiarching a target is considered an obvious
activity, so you have the option of just post patches, instead of
RFA's. For instance see the multiarching of the h8300 (please
do it with incremental patches).

Elena


 > 
 > A patch to add the gdbarch framework will shortly follow the checkin.
 > 
 > 2002-11-05  Joel Brobecker  <brobecker@gnat.com>
 > 
 >         * hppa-tdep.c (_initialize_hppa_tdep): Move function body
 >         to end of file, to be more consistent with the pratice followed
 >         in other targets.
 > 
 > -- 
 > Joel
 > Index: hppa-tdep.c
 > ===================================================================
 > RCS file: /cvs/src/src/gdb/hppa-tdep.c,v
 > retrieving revision 1.29
 > diff -c -3 -p -r1.29 hppa-tdep.c
 > *** hppa-tdep.c	30 Oct 2002 20:42:54 -0000	1.29
 > --- hppa-tdep.c	6 Nov 2002 02:18:50 -0000
 > *************** hppa_skip_permanent_breakpoint (void)
 > *** 4674,4724 ****
 >     /* We can leave the tail's space the same, since there's no jump.  */
 >   }
 >   
 > - void
 > - _initialize_hppa_tdep (void)
 > - {
 > -   struct cmd_list_element *c;
 > -   void break_at_finish_command (char *arg, int from_tty);
 > -   void tbreak_at_finish_command (char *arg, int from_tty);
 > -   void break_at_finish_at_depth_command (char *arg, int from_tty);
 > - 
 > -   tm_print_insn = print_insn_hppa;
 > - 
 > -   add_cmd ("unwind", class_maintenance, unwind_command,
 > - 	   "Print unwind table entry at given address.",
 > - 	   &maintenanceprintlist);
 > - 
 > -   deprecate_cmd (add_com ("xbreak", class_breakpoint, 
 > - 			  break_at_finish_command,
 > - 			  concat ("Set breakpoint at procedure exit. \n\
 > - Argument may be function name, or \"*\" and an address.\n\
 > - If function is specified, break at end of code for that function.\n\
 > - If an address is specified, break at the end of the function that contains \n\
 > - that exact address.\n",
 > - 		   "With no arg, uses current execution address of selected stack frame.\n\
 > - This is useful for breaking on return to a stack frame.\n\
 > - \n\
 > - Multiple breakpoints at one place are permitted, and useful if conditional.\n\
 > - \n\
 > - Do \"help breakpoints\" for info on other commands dealing with breakpoints.", NULL)), NULL);
 > -   deprecate_cmd (add_com_alias ("xb", "xbreak", class_breakpoint, 1), NULL);
 > -   deprecate_cmd (add_com_alias ("xbr", "xbreak", class_breakpoint, 1), NULL);
 > -   deprecate_cmd (add_com_alias ("xbre", "xbreak", class_breakpoint, 1), NULL);
 > -   deprecate_cmd (add_com_alias ("xbrea", "xbreak", class_breakpoint, 1), NULL);
 > - 
 > -   deprecate_cmd (c = add_com ("txbreak", class_breakpoint, 
 > - 			      tbreak_at_finish_command,
 > - "Set temporary breakpoint at procedure exit.  Either there should\n\
 > - be no argument or the argument must be a depth.\n"), NULL);
 > -   set_cmd_completer (c, location_completer);
 > -   
 > -   if (xdb_commands)
 > -     deprecate_cmd (add_com ("bx", class_breakpoint, 
 > - 			    break_at_finish_at_depth_command,
 > - "Set breakpoint at procedure exit.  Either there should\n\
 > - be no argument or the argument must be a depth.\n"), NULL);
 > - }
 > - 
 >   /* Copy the function value from VALBUF into the proper location
 >      for a function return.
 >   
 > --- 4674,4679 ----
 > *************** hppa_extract_return_value (struct type *
 > *** 4768,4770 ****
 > --- 4723,4771 ----
 >   		: (4 - TYPE_LENGTH (type)))),
 >   	    TYPE_LENGTH (type));
 >   }
 > + 
 > + void
 > + _initialize_hppa_tdep (void)
 > + {
 > +   struct cmd_list_element *c;
 > +   void break_at_finish_command (char *arg, int from_tty);
 > +   void tbreak_at_finish_command (char *arg, int from_tty);
 > +   void break_at_finish_at_depth_command (char *arg, int from_tty);
 > + 
 > +   tm_print_insn = print_insn_hppa;
 > + 
 > +   add_cmd ("unwind", class_maintenance, unwind_command,
 > + 	   "Print unwind table entry at given address.",
 > + 	   &maintenanceprintlist);
 > + 
 > +   deprecate_cmd (add_com ("xbreak", class_breakpoint, 
 > + 			  break_at_finish_command,
 > + 			  concat ("Set breakpoint at procedure exit. \n\
 > + Argument may be function name, or \"*\" and an address.\n\
 > + If function is specified, break at end of code for that function.\n\
 > + If an address is specified, break at the end of the function that contains \n\
 > + that exact address.\n",
 > + 		   "With no arg, uses current execution address of selected stack frame.\n\
 > + This is useful for breaking on return to a stack frame.\n\
 > + \n\
 > + Multiple breakpoints at one place are permitted, and useful if conditional.\n\
 > + \n\
 > + Do \"help breakpoints\" for info on other commands dealing with breakpoints.", NULL)), NULL);
 > +   deprecate_cmd (add_com_alias ("xb", "xbreak", class_breakpoint, 1), NULL);
 > +   deprecate_cmd (add_com_alias ("xbr", "xbreak", class_breakpoint, 1), NULL);
 > +   deprecate_cmd (add_com_alias ("xbre", "xbreak", class_breakpoint, 1), NULL);
 > +   deprecate_cmd (add_com_alias ("xbrea", "xbreak", class_breakpoint, 1), NULL);
 > + 
 > +   deprecate_cmd (c = add_com ("txbreak", class_breakpoint, 
 > + 			      tbreak_at_finish_command,
 > + "Set temporary breakpoint at procedure exit.  Either there should\n\
 > + be no argument or the argument must be a depth.\n"), NULL);
 > +   set_cmd_completer (c, location_completer);
 > +   
 > +   if (xdb_commands)
 > +     deprecate_cmd (add_com ("bx", class_breakpoint, 
 > + 			    break_at_finish_at_depth_command,
 > + "Set breakpoint at procedure exit.  Either there should\n\
 > + be no argument or the argument must be a depth.\n"), NULL);
 > + }
 > + 

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Joel Brobecker]

>  > 2002-11-05  Joel Brobecker  <brobecker@gnat.com>
>  > 
>  >         * hppa-tdep.c (_initialize_hppa_tdep): Move function body
>  >         to end of file, to be more consistent with the pratice followed
>  >         in other targets.
> 
> Sure.

Thanks for the little confirmation that this change was ok.

> Actually multiarching a target is considered an obvious activity, so
> you have the option of just post patches, instead of RFA's. For
> instance see the multiarching of the h8300 (please do it with
> incremental patches).

The trouble is that I still consider myself new to the GDB project,
and look forward to the many years of learning to come with great
spirit. In the meantime, I often feel more confortable after I asked
for a quick confirmation that my understanding was correct before
doing some checkins that seem too obvious to be correct.

The testsuite is a good tool, but is sometimes not enough. I could for
instance introduce a bug in a different target.

But I do reassure you, I will try to do as much as possible of the
multiarch conversion without asking for RFAs :-).

-- 
Joel

Re: [PATCH/RFA] Move _initialize_hppa_tdep to the end of file

[Andrew Cagney]

>  > 2002-11-05  Joel Brobecker  <brobecker@gnat.com>
>>  > 
>>  >         * hppa-tdep.c (_initialize_hppa_tdep): Move function body
>>  >         to end of file, to be more consistent with the pratice followed
>>  >         in other targets.
>> 
>> Sure.
> 
> 
> Thanks for the little confirmation that this change was ok.
> 
> 
>> Actually multiarching a target is considered an obvious activity, so
>> you have the option of just post patches, instead of RFA's. For
>> instance see the multiarching of the h8300 (please do it with
>> incremental patches).
> 
> 
> The trouble is that I still consider myself new to the GDB project,
> and look forward to the many years of learning to come with great
> spirit. In the meantime, I often feel more confortable after I asked
> for a quick confirmation that my understanding was correct before
> doing some checkins that seem too obvious to be correct.
> 
> The testsuite is a good tool, but is sometimes not enough. I could for
> instance introduce a bug in a different target.

It's an accepted risk.  Multi-arching takes priority over keeping things 
working 100%.  Besides, if you're modifying HP files, the only thing you 
can break is HP stuff :-)

> But I do reassure you, I will try to do as much as possible of the
> multiarch conversion without asking for RFAs :-).

One suggestion.  If you hit a problem (eg a method isn't yet multi-arch) 
put it to one side and work on something else.

Good luck!
Andrew