[PATCH] startup.S does not find includes

[PATCH] startup.S does not find includes

[Christian Franke]

[-- Attachment #1: Type: text/plain, Size: 238 bytes --]

Hi,

some recent checkin appearently introduces the following regression:

Compilation fails if configure'd outside of $(srcdir).
#include "kern/i386/realmode.S" and "kern/i386/loader.S" fail in startup.S.

Patch is attached.

Christian


[-- Attachment #2: grub2-Makefile.in.patch --]
[-- Type: text/x-patch, Size: 497 bytes --]

--- grub2.orig/Makefile.in	2007-06-11 08:26:17.000000000 +0200
+++ grub2/Makefile.in	2007-10-21 18:43:11.421875000 +0200
@@ -64,7 +64,7 @@ CPPFLAGS = @CPPFLAGS@ -I. -Iinclude -I$(
 	 -DGRUB_LIBDIR=\"$(pkglibdir)\"
 TARGET_CC = @TARGET_CC@
 TARGET_CFLAGS = @TARGET_CFLAGS@
-TARGET_CPPFLAGS = @TARGET_CPPFLAGS@ -I. -Iinclude -I$(srcdir)/include \
+TARGET_CPPFLAGS = @TARGET_CPPFLAGS@ -I. -Iinclude -I$(srcdir) -I$(srcdir)/include \
 	-Wall -W
 TARGET_LDFLAGS = @TARGET_LDFLAGS@
 OBJCOPY = @OBJCOPY@

Re: [PATCH] startup.S does not find includes

[Robert Millan]

On Sun, Oct 21, 2007 at 06:56:21PM +0200, Christian Franke wrote:
> Hi,
> 
> some recent checkin appearently introduces the following regression:
> 
> Compilation fails if configure'd outside of $(srcdir).
> #include "kern/i386/realmode.S" and "kern/i386/loader.S" fail in startup.S.

Uhm I just noticed that previously existing #includes have relative paths,
e.g.:

  #include "lzo1x.S"

Perhaps I should have done the same instead of adding new include dirs.

Thoughts?

-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)

Re: [PATCH] startup.S does not find includes

[Christian Franke]

Robert Millan wrote:
> On Sun, Oct 21, 2007 at 06:56:21PM +0200, Christian Franke wrote:
>   
>> Hi,
>>
>> some recent checkin appearently introduces the following regression:
>>
>> Compilation fails if configure'd outside of $(srcdir).
>> #include "kern/i386/realmode.S" and "kern/i386/loader.S" fail in startup.S.
>>     
>
> Uhm I just noticed that previously existing #includes have relative paths,
> e.g.:
>
>   #include "lzo1x.S"
>
> Perhaps I should have done the same instead of adding new include dirs.
>
> Thoughts?
>
>   

For kern/i386/pc/startup.S:

#include "../realmode.S"  // include pc-independent i386 realmode code

Should work without any -I option.

Christian

Re: [PATCH] startup.S does not find includes

[walt]

Christian Franke wrote:
> Robert Millan wrote:
>> On Sun, Oct 21, 2007 at 06:56:21PM +0200, Christian Franke wrote:
>>> Hi,
>>>
>>> some recent checkin appearently introduces the following regression:
>>>
>>> Compilation fails if configure'd outside of $(srcdir).
>>> #include "kern/i386/realmode.S" and "kern/i386/loader.S" fail in 
>>> startup.S.
>>
>> Uhm I just noticed that previously existing #includes have relative 
>> paths,
>> e.g.:
>>
>> #include "lzo1x.S"
>>
>> Perhaps I should have done the same instead of adding new include dirs.
>>
>> Thoughts?
>>
> 
> For kern/i386/pc/startup.S:
> 
> #include "../realmode.S" // include pc-independent i386 realmode code
> 
> Should work without any -I option.

I considered that, but wouldn't it break compilation when building
in-tree, ie without an obj directory?

Re: [PATCH] startup.S does not find includes

[Robert Millan]

On Mon, Oct 22, 2007 at 06:31:22AM -0700, walt wrote:
> Christian Franke wrote:
> >Robert Millan wrote:
> >>On Sun, Oct 21, 2007 at 06:56:21PM +0200, Christian Franke wrote:
> >>>Hi,
> >>>
> >>>some recent checkin appearently introduces the following regression:
> >>>
> >>>Compilation fails if configure'd outside of $(srcdir).
> >>>#include "kern/i386/realmode.S" and "kern/i386/loader.S" fail in 
> >>>startup.S.
> >>
> >>Uhm I just noticed that previously existing #includes have relative 
> >>paths,
> >>e.g.:
> >>
> >>#include "lzo1x.S"
> >>
> >>Perhaps I should have done the same instead of adding new include dirs.
> >>
> >>Thoughts?
> >>
> >
> >For kern/i386/pc/startup.S:
> >
> >#include "../realmode.S" // include pc-independent i386 realmode code
> >
> >Should work without any -I option.
> 
> I considered that, but wouldn't it break compilation when building
> in-tree, ie without an obj directory?

Why would it?

-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)

Re: [PATCH] startup.S does not find includes

[walt]

Robert Millan wrote:
> On Mon, Oct 22, 2007 at 06:31:22AM -0700, walt wrote:
>> Christian Franke wrote:

>>> For kern/i386/pc/startup.S:
>>>
>>> #include "../realmode.S" // include pc-independent i386 realmode code
>>>
>>> Should work without any -I option.

>> I considered that, but wouldn't it break compilation when building
>> in-tree, ie without an obj directory?

> Why would it?

Hm.  It did yesterday when I was tired.  Today, it works okay.
Obviously black magic at work :o)

Re: [PATCH] startup.S does not find includes

[Robert Millan]

Fixed.

-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)