Stupid Question

Stupid Question

[Michael A. Thompson]

1. Which rpm contains the standard include headers?
I come from mklinux and IRIX... I cant seem to find stdio.h stdlib.h
etc... or sys/*.h directory. I installed the egcs c/c++ rpms,
kernal-devel rpms and other gnu develope apps libs utils and headers but
can find the standard stuff....
2. is it normal for c++ includes to be here: /usr/include/g++-2?
3. What is the normal file structure for linuxppc 5 include directory?
is it like x86 linux, BSD, etc....

Thanks for any help on this.  I have a very limited ppc linux system and
only want a working devel system without the other stuff....

Michael

--
----------------------------------
Michael A. Thompson
[IRIX - NeXTStep - Linux - MacOS - Windows]

Home: (940)382-2086
E-Mail: mat0001@jove.acs.unt.edu
----------------------------------


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: Stupid Question

[Martin Costabel]

"Michael A. Thompson" wrote:
> 
> 1. Which rpm contains the standard include headers?
> I come from mklinux and IRIX... I cant seem to find stdio.h stdlib.h
> etc... or sys/*.h directory. I installed the egcs c/c++ rpms,
> kernal-devel rpms and other gnu develope apps libs utils and headers but
> can find the standard stuff....

They are in the glibc-devel package.

> 2. is it normal for c++ includes to be here: /usr/include/g++-2?

Yes. Depends on your compiler. 

> 3. What is the normal file structure for linuxppc 5 include directory?
> is it like x86 linux, BSD, etc....

I don't know the others. One thing: Make sure that /usr/include/asm is a
link to /usr/src/linux/include/asm which in turn points to
/usr/src/linux/include/asm-ppc and /usr/include/linux points to
/usr/src/linux/include/linux (you do have the kernel headers installed,
don't you?)

--
Martin

[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Stupid Question.

[John T. Williams]

I know I should know this, but I don't, and haven't been able to find a
good answer.

What does declaring a function static do?    ie 
What is the difference between

static int blue() {
	return 0;
}

and 

int blue() {
	return 0;
}

Re: Stupid Question.

[Jeff Woods]

At 5/12/2004 02:53 PM -0400, John T. Williams wrote:
>What does declaring a function static do?    ie
>What is the difference between
>
>static int blue() {
>         return 0;
>}
>
>and
>
>int blue() {
>         return 0;
>}

It makes the function visible only within the comilation unit it's in.  See 
also:

http://www.phim.unibe.ch/comp_doc/c_manual/C/SYNTAX/static.htm

As the website points out static functions and static variables seem like 
very different meanings of the word "static".  In both cases I think of it 
as "global but private/hidden".

--
Jeff Woods <kazrak+kernel@cesmail.net> 

Re: Stupid Question.

[Jan-Benedict Glaw]

[-- Attachment #1: Type: text/plain, Size: 1076 bytes --]

On Wed, 2004-05-12 13:02:25 -0600, Jeff Woods <Kazrak+kernel@cesmail.net>
wrote in message <6.0.1.1.0.20040512125919.02e8cec0@no.incoming.mail>:
> At 5/12/2004 02:53 PM -0400, John T. Williams wrote:
> >What does declaring a function static do?    ie
> >What is the difference between
> >static int blue() {
> >        return 0;
> >}
> >and
> >int blue() {
> >        return 0;
> >}
> 
> It makes the function visible only within the comilation unit it's in.  See 
> also:

A static function (or variable) will only be "visible" within exactly
that .c file which it is defined in.

Static automatic variables (those that are within functions) do have a
different semantic: their contents is preserved during multiple function
calls.

MfG, JBG

-- 
   Jan-Benedict Glaw       jbglaw@lug-owl.de    . +49-172-7608481
   "Eine Freie Meinung in  einem Freien Kopf    | Gegen Zensur | Gegen Krieg
    fuer einen Freien Staat voll Freier Bürger" | im Internet! |   im Irak!
   ret = do_actions((curr | FREE_SPEECH) & ~(NEW_COPYRIGHT_LAW | DRM | TCPA));

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Stupid Question.

[Glynn Clements]

Jan-Benedict Glaw wrote:

> Static automatic variables (those that are within functions) do have a
> different semantic: their contents is preserved during multiple function
> calls.

"Static automatic variables" is a self-contradiction. "Static local
variables" (as opposed to "automatic local variables") would be more
accurate.

Local variables (those whose scope is limited to the block in which
they are declared) are automatic (stored on the stack) by default. The
"static" keyword makes them static (stored in the data segment).

This overloading of the "static" keyword is a common source of
confusion. They really should have added another keyword, IMHO. [And
dispensed with the (entirely redundant) "automatic" keyword, which
I've never actually seen used in nearly 20 years of C programming.]

And the distinction between automatic and static local variables is
somewhat deeper than the fact that the value is preserved across
calls. It's highly significant when dealing with reentrancy (e.g. 
recursive functions, multi-threaded code).

For instance, an automatic variable has a separate instance for each
call, while a static variable only has a single instance. OTOH, you
can legitimately return or store a pointer to a static variable and
reference it from outside of the function, whereas an automatic
variable ceases to exist once the function returns.

-- 
Glynn Clements <glynn.clements@virgin.net>

Stupid question

[Marcel Hilzinger]

What, if someone likes to create a directory called "metas" under reiser4? 

Actually this is not possible, as the system will complain about: "there is 
already a directory metas".

Marcel
-- 
Üdvözlettel -- Mit freundlichen Grüssen,
Marcel Hilzinger

Re: Stupid question

[mjt]

On Thu, May 27, 2004 at 11:38:20AM +0200, Marcel Hilzinger wrote:
>What, if someone likes to create a directory called "metas" under reiser4? 

That's a can of worms opened many times too often.

>Actually this is not possible, as the system will complain about: "there is 
>already a directory metas".

Yes.

I recommend patching it to ..metas/

NEEXT!

-- 
mjt

Stupid question

[Gene Heskett]

Greetings;

Silly Q of the day probably, but what do I set in a Makefile for the 
-march=option for building on a 233 mhz Pentium 2?

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[Jan Engelhardt]

>Greetings;
>
>Silly Q of the day probably, but what do I set in a Makefile for the
>-march=option for building on a 233 mhz Pentium 2?

Now that's really stupid, but here's the answer:

You run `make menuconfig` (or whichever you like) and choose Processor Type
"Pentium II".


Jan Engelhardt
-- 
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, www.gwdg.de

Re: Stupid question

[Gene Heskett]

On Monday 22 November 2004 03:33, Jan Engelhardt wrote:
>>Greetings;
>>
>>Silly Q of the day probably, but what do I set in a Makefile for
>> the -march=option for building on a 233 mhz Pentium 2?
>
>Now that's really stupid, but here's the answer:
>
>You run `make menuconfig` (or whichever you like) and choose
> Processor Type "Pentium II".
>
>
>Jan Engelhardt

If I was building a kernel, then yes my question was stupid.

Except I'm not building a kernel, I'm tryng to compile a module to 
drive some truely dumb hardware, reusing code that was last touched 
in 1999 when the Makefile could use the -mi486 syntax with gcc-2.95.  
Now we have gcc 3.3.3, even on a Brain Dead Install of emc 
(LinuxCNC), which has a 2.4 kernel with the rtai patch kit for real 
time.  I need to put this directly into the Makefile using the 
currently valid  -march= & -mcpu= syntax's.

So far, the closest I've come to getting it right on that box bails 
out because its being mis-interpreted as -march=mips on that box.  
Those includes aren't part of an x86 install.  Hence the 
question. :-)

It appears I've got more problems that this one though, lots of "hey 
you can't do that anymore" warnings when I try to build it on this 
box (bleeding edge kernel 2.6.10-rc2-bk4-kjt1 etc) and it eventually 
bails out or course.  It apparently plays mix-n-match with kernel vs 
glibc headers too.  As a loadable module, I'd assume it should be 
using kernel headers only?

Yeah, I'm stupid.  Virtually all of my original C coding has been done 
on much smaller architectures, and 15 to 20 years ago.  Terminal rust 
has set in on those skills I once had well honed when I was only 55.  
Now I'm 70, semi-retired, and I'd like to drive some stepper motors 
on a micromill.

While composing this, I also read the man page for gcc and while it 
has a bunch of options, it doesn't always list/define the valid ones, 
particularly for x86 stuffs.  Lots of other architectures are covered 
in much better detail. I guess they figure everyone knows x86 stuff.

Sadly, for those of us who came up thru the motorola ranks, such is 
not the case.  Heck, I'd do this in assembly IF I knew intel asm as 
well as I know 6x09 asm.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[Jan Engelhardt]

>>>Silly Q of the day probably, but what do I set in a Makefile for
>>> the -march=option for building on a 233 mhz Pentium 2?
>>
>>Now that's really stupid, but here's the answer:
>>
>>You run `make menuconfig` (or whichever you like) and choose
>> Processor Type "Pentium II".
>
>If I was building a kernel, then yes my question was stupid.
>
>Except I'm not building a kernel, I'm tryng to compile a module to
>drive some truely dumb hardware, reusing code that was last touched
>[...]

Well, take a fresh kernel tree, set the desired CPU type, and then look at the
.config which is generated. Voilà -- in theory ;-)

>Yeah, I'm stupid.  Virtually all of my original C coding has been done
>on much smaller architectures, and 15 to 20 years ago.  Terminal rust

Never hurts trying to compile a 2.6 for 386SX if the will is strong. :-)



Jan Engelhardt
-- 
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, www.gwdg.de

Re: Stupid question

[Gene Heskett]

On Monday 22 November 2004 14:02, Jan Engelhardt wrote:
>>>>Silly Q of the day probably, but what do I set in a Makefile for
>>>> the -march=option for building on a 233 mhz Pentium 2?
>>>
>>>Now that's really stupid, but here's the answer:
>>>
>>>You run `make menuconfig` (or whichever you like) and choose
>>> Processor Type "Pentium II".
>>
>>If I was building a kernel, then yes my question was stupid.
>>
>>Except I'm not building a kernel, I'm tryng to compile a module to
>>drive some truely dumb hardware, reusing code that was last touched
>>[...]
>
>Well, take a fresh kernel tree, set the desired CPU type, and then
> look at the .config which is generated. Voilà -- in theory ;-)
>
I'm glad you put that caveat in there, Jan. :)

>>Yeah, I'm stupid.  Virtually all of my original C coding has been
>> done on much smaller architectures, and 15 to 20 years ago. 
>> Terminal rust
>
>Never hurts trying to compile a 2.6 for 386SX if the will is strong.
> :-)

On an SX?  No way Jose.  That would take more will power than I have & 
I quit a 2 pack a day habit cold turkey 15+ years ago.  Theres one of 
those things out in a storage shed, won't even run winderz > 95SR1, 
its been tried a couple of times.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[Jan Engelhardt]

>>Well, take a fresh kernel tree, set the desired CPU type, and then
>> look at the .config which is generated. Voilà -- in theory ;-)

>I'm glad you put that caveat in there, Jan. :)

What? There are no bugs! They're features.

>>Never hurts trying to compile a 2.6 for 386SX if the will is strong.
>> :-)
>
>On an SX?  No way Jose.  That would take more will power than I have &

Read more closely. I said "compile for" not "compile on".


Jan Engelhardt
-- 
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, www.gwdg.de

Re: Stupid question

[Gene Heskett]

On Monday 22 November 2004 14:32, Jan Engelhardt wrote:
[...]
>>
>>On an SX?  No way Jose.  That would take more will power than I
>> have &
>
>Read more closely. I said "compile for" not "compile on".
>
Touche'

>
>Jan Engelhardt

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[P]

Gene Heskett wrote:
> Greetings;
> 
> Silly Q of the day probably, but what do I set in a Makefile for the 
> -march=option for building on a 233 mhz Pentium 2?

http://www.pixelbeat.org/scripts/gcccpuopt

Re: Stupid question

[Gene Heskett]

On Monday 22 November 2004 09:32, P@draigBrady.com wrote:
>Gene Heskett wrote:
>> Greetings;
>>
>> Silly Q of the day probably, but what do I set in a Makefile for
>> the -march=option for building on a 233 mhz Pentium 2?
>
>http://www.pixelbeat.org/scripts/gcccpuopt

Thanks very much.  Obviously someone else needed to scratch this itch 
too.  This should produce the correct results when running on the 
target machine.  Here, it produces this:
[root@coyote CIO-DIO96]# sh ../gcccpuopt
 -march=athlon-xp -mfpmath=sse -msse -mmmx -m3dnow

-- 
Cheers & thanks again, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[Nix]

On 23 Nov 2004, Gene Heskett yowled:
> On Monday 22 November 2004 09:32, P@draigBrady.com wrote:
>>Gene Heskett wrote:
>>> Greetings;
>>>
>>> Silly Q of the day probably, but what do I set in a Makefile for
>>> the -march=option for building on a 233 mhz Pentium 2?
>>
>>http://www.pixelbeat.org/scripts/gcccpuopt
> 
> Thanks very much.  Obviously someone else needed to scratch this itch 
> too.  This should produce the correct results when running on the 
> target machine.  Here, it produces this:
> [root@coyote CIO-DIO96]# sh ../gcccpuopt
>  -march=athlon-xp -mfpmath=sse -msse -mmmx -m3dnow

... which is peculiar, as -mmmx -msse is redundant, as is -mmmx -m3dnow,
and all three of those flags are the end are implied by -march=athlon-xp
anyway.

(-mfpmath=sse *is* useful on non-64-bit platforms, though.)

-- 
`The sword we forged has turned upon us
 Only now, at the end of all things do we see
 The lamp-bearer dies; only the lamp burns on.'

Re: Stupid question

[P]

Nix wrote:
> On 23 Nov 2004, Gene Heskett yowled:
> 
>>On Monday 22 November 2004 09:32, P@draigBrady.com wrote:
>>
>>>Gene Heskett wrote:
>>>
>>>>Greetings;
>>>>
>>>>Silly Q of the day probably, but what do I set in a Makefile for
>>>>the -march=option for building on a 233 mhz Pentium 2?
>>>
>>>http://www.pixelbeat.org/scripts/gcccpuopt
>>
>>Thanks very much.  Obviously someone else needed to scratch this itch 
>>too.  This should produce the correct results when running on the 
>>target machine.  Here, it produces this:
>>[root@coyote CIO-DIO96]# sh ../gcccpuopt
>> -march=athlon-xp -mfpmath=sse -msse -mmmx -m3dnow
> 
> 
> ... which is peculiar, as -mmmx -msse is redundant, as is -mmmx -m3dnow,
> and all three of those flags are the end are implied by -march=athlon-xp
> anyway.
> 
> (-mfpmath=sse *is* useful on non-64-bit platforms, though.)

I added those in so that they were explicit.
They do no harm. I had a version that didn't print
these redundant options but got many requests
about whether they were needed. You can't win.

Pádraig.

Re: Stupid question

[Gene Heskett]

On Tuesday 23 November 2004 07:19, P@draigBrady.com wrote:
>Nix wrote:
>> On 23 Nov 2004, Gene Heskett yowled:
>>>On Monday 22 November 2004 09:32, P@draigBrady.com wrote:
>>>>Gene Heskett wrote:
>>>>>Greetings;
>>>>>
>>>>>Silly Q of the day probably, but what do I set in a Makefile for
>>>>>the -march=option for building on a 233 mhz Pentium 2?
>>>>
>>>>http://www.pixelbeat.org/scripts/gcccpuopt
>>>
>>>Thanks very much.  Obviously someone else needed to scratch this
>>> itch too.  This should produce the correct results when running
>>> on the target machine.  Here, it produces this:
>>>[root@coyote CIO-DIO96]# sh ../gcccpuopt
>>> -march=athlon-xp -mfpmath=sse -msse -mmmx -m3dnow
>>
>> ... which is peculiar, as -mmmx -msse is redundant, as is -mmmx
>> -m3dnow, and all three of those flags are the end are implied by
>> -march=athlon-xp anyway.
>>
>> (-mfpmath=sse *is* useful on non-64-bit platforms, though.)
>
>I added those in so that they were explicit.
>They do no harm. I had a version that didn't print
>these redundant options but got many requests
>about whether they were needed. You can't win.
>
>Pádraig.

I've heard that expression before :-)

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.

Re: Stupid question

[Gene Heskett]

On Tuesday 23 November 2004 07:08, Nix wrote:
>On 23 Nov 2004, Gene Heskett yowled:
>> On Monday 22 November 2004 09:32, P@draigBrady.com wrote:
>>>Gene Heskett wrote:
>>>> Greetings;
>>>>
>>>> Silly Q of the day probably, but what do I set in a Makefile for
>>>> the -march=option for building on a 233 mhz Pentium 2?
>>>
>>>http://www.pixelbeat.org/scripts/gcccpuopt
>>
>> Thanks very much.  Obviously someone else needed to scratch this
>> itch too.  This should produce the correct results when running on
>> the target machine.  Here, it produces this:
>> [root@coyote CIO-DIO96]# sh ../gcccpuopt
>>  -march=athlon-xp -mfpmath=sse -msse -mmmx -m3dnow
>
>... which is peculiar, as -mmmx -msse is redundant, as is -mmmx
> -m3dnow, and all three of those flags are the end are implied by
> -march=athlon-xp anyway.
>
Humm, as I see it, the choices presented in a make xconfig do not 
allow the choice of the athlon-xp, just the athlon.  Are there any 
worthwhile optimizations the added '-xp' would bring into play?  
Making it enough faster to make that choice a worthwhile choice?

Said another way, what file would I edit after the .config has been 
saved in order to put that into effect for the subsequent build?

>(-mfpmath=sse *is* useful on non-64-bit platforms, though.)

Oh?  As that stuff is pretty invisible during a make today, how would 
one go about determining if thats in effect on this machine?

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.29% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.