[Qemu-devel] another patch against head

[Qemu-devel] another patch against head

[Hotmail]

The following is another patch against the head that lets me build it on Windows, as well as makes the adlib.c file debuggable.

Index: Makefile
===================================================================
RCS file: /sources/qemu/qemu/Makefile,v
retrieving revision 1.136
diff -r1.136 Makefile
17a18,20
> ifdef CONFIG_WIN32
> LIBS= -lmingw32 -mno-cygwin
> else
18a22
> endif
Index: cocoa.m
===================================================================
RCS file: /sources/qemu/qemu/cocoa.m,v
retrieving revision 1.14
diff -r1.14 cocoa.m
754c754
<         [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"cow",@"cloop",@"vmdk",nil]
---
>         [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"qcow2",@"cow",@"cloop",@"vmdk",nil]
Index: monitor.c
===================================================================
RCS file: /sources/qemu/qemu/monitor.c,v
retrieving revision 1.87
diff -r1.87 monitor.c
31a32
> #include "qemu-timer.h"
Index: hw/adlib.c
===================================================================
RCS file: /sources/qemu/qemu/hw/adlib.c,v
retrieving revision 1.8
diff -r1.8 adlib.c
25a26
> #include "audio/audio.h"
26a28
> #include "qemu-timer.h"

Re: [Qemu-devel] another patch against head

[Johannes Schindelin]

Hi Hotmail,

On Tue, 27 Nov 2007, Hotmail wrote:

> The following is another patch against the head that lets me build it on 
> Windows, as well as makes the adlib.c file debuggable.
> 
> Index: Makefile
> ===================================================================
> RCS file: /sources/qemu/qemu/Makefile,v
> retrieving revision 1.136
> diff -r1.136 Makefile
> 17a18,20
> > ifdef CONFIG_WIN32
> > LIBS= -lmingw32 -mno-cygwin
> > else
> 18a22
> > endif

So you build it in Cygwin, not "on Windows".

> Index: cocoa.m
> ===================================================================
> RCS file: /sources/qemu/qemu/cocoa.m,v
> retrieving revision 1.14
> diff -r1.14 cocoa.m
> 754c754
> <         [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"cow",@"cloop",@"vmdk",nil]
> ---
> >         [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"qcow2",@"cow",@"cloop",@"vmdk",nil]

How does this help compiling on Windows?

> Index: monitor.c
> ===================================================================
> RCS file: /sources/qemu/qemu/monitor.c,v
> retrieving revision 1.87
> diff -r1.87 monitor.c
> 31a32
> > #include "qemu-timer.h"
> Index: hw/adlib.c
> ===================================================================
> RCS file: /sources/qemu/qemu/hw/adlib.c,v
> retrieving revision 1.8
> diff -r1.8 adlib.c
> 25a26
> > #include "audio/audio.h"
> 26a28
> > #include "qemu-timer.h"

Why are these needed?  It appears that all other platforms build QEmu just 
fine without these #includes.

But then, the answer to that riddle could be in the context.  Since you 
did not use the unified diff format, I have no chance to know.

Ciao,
Dscho

Re: [Qemu-devel] another patch against head

[Samuel Thibault]

Johannes Schindelin, le Wed 28 Nov 2007 12:49:34 +0000, a écrit :
> On Tue, 27 Nov 2007, Hotmail wrote:
> 
> > The following is another patch against the head that lets me build it on 
> > Windows, as well as makes the adlib.c file debuggable.
> > 
> > Index: Makefile
> > ===================================================================
> > RCS file: /sources/qemu/qemu/Makefile,v
> > retrieving revision 1.136
> > diff -r1.136 Makefile
> > 17a18,20
> > > ifdef CONFIG_WIN32
> > > LIBS= -lmingw32 -mno-cygwin
> > > else
> > 18a22
> > > endif
> 
> So you build it in Cygwin, not "on Windows".

Not exactly. -mno-cygwin means that he used the cygwin-provided gcc, but
provide a non-cygwin binary that can be shipped without extra dlls.

Samuel

Re: [Qemu-devel] another patch against head

[Johannes Schindelin]

Hi,

On Wed, 28 Nov 2007, Samuel Thibault wrote:

> Johannes Schindelin, le Wed 28 Nov 2007 12:49:34 +0000, a ?crit :
> > On Tue, 27 Nov 2007, Hotmail wrote:
> > 
> > > The following is another patch against the head that lets me build it on 
> > > Windows, as well as makes the adlib.c file debuggable.
> > > 
> > > Index: Makefile
> > > ===================================================================
> > > RCS file: /sources/qemu/qemu/Makefile,v
> > > retrieving revision 1.136
> > > diff -r1.136 Makefile
> > > 17a18,20
> > > > ifdef CONFIG_WIN32
> > > > LIBS= -lmingw32 -mno-cygwin
> > > > else
> > > 18a22
> > > > endif
> > 
> > So you build it in Cygwin, not "on Windows".
> 
> Not exactly. -mno-cygwin means that he used the cygwin-provided gcc, but
> provide a non-cygwin binary that can be shipped without extra dlls.

And how exactly does that differ from what I said?  It is building in 
Cygwin.  Maybe I should have added for completeness' sake "as opposed to 
MinGW".

Thanks,
Dscho

Re: [Qemu-devel] another patch against head

[Samuel Thibault]

Johannes Schindelin, le Wed 28 Nov 2007 14:01:21 +0000, a écrit :
> On Wed, 28 Nov 2007, Samuel Thibault wrote:
> 
> > Johannes Schindelin, le Wed 28 Nov 2007 12:49:34 +0000, a ?crit :
> > > On Tue, 27 Nov 2007, Hotmail wrote:
> > > 
> > > > The following is another patch against the head that lets me build it on 
> > > > Windows, as well as makes the adlib.c file debuggable.
> > > > 
> > > > Index: Makefile
> > > > ===================================================================
> > > > RCS file: /sources/qemu/qemu/Makefile,v
> > > > retrieving revision 1.136
> > > > diff -r1.136 Makefile
> > > > 17a18,20
> > > > > ifdef CONFIG_WIN32
> > > > > LIBS= -lmingw32 -mno-cygwin
> > > > > else
> > > > 18a22
> > > > > endif
> > > 
> > > So you build it in Cygwin, not "on Windows".
> > 
> > Not exactly. -mno-cygwin means that he used the cygwin-provided gcc, but
> > provide a non-cygwin binary that can be shipped without extra dlls.
> 
> And how exactly does that differ from what I said?  It is building in 
> Cygwin.

With the MinGW headers. That makes a big difference.
Compiling in cygwin with -mno-cygwin is almost the same as compiling
with MinGW tools (the difference mostly lies in Makefile and shell
script subtleties)

Samuel

Re: [Qemu-devel] another patch against head

[Johannes Schindelin]

Hi,

On Wed, 28 Nov 2007, Samuel Thibault wrote:

> Johannes Schindelin, le Wed 28 Nov 2007 14:01:21 +0000, a ?crit :
> > On Wed, 28 Nov 2007, Samuel Thibault wrote:
> > 
> > > Johannes Schindelin, le Wed 28 Nov 2007 12:49:34 +0000, a ?crit :
> > > > On Tue, 27 Nov 2007, Hotmail wrote:
> > > > 
> > > > > The following is another patch against the head that lets me build it on 
> > > > > Windows, as well as makes the adlib.c file debuggable.
> > > > > 
> > > > > Index: Makefile
> > > > > ===================================================================
> > > > > RCS file: /sources/qemu/qemu/Makefile,v
> > > > > retrieving revision 1.136
> > > > > diff -r1.136 Makefile
> > > > > 17a18,20
> > > > > > ifdef CONFIG_WIN32
> > > > > > LIBS= -lmingw32 -mno-cygwin
> > > > > > else
> > > > > 18a22
> > > > > > endif
> > > > 
> > > > So you build it in Cygwin, not "on Windows".
> > > 
> > > Not exactly. -mno-cygwin means that he used the cygwin-provided gcc, but
> > > provide a non-cygwin binary that can be shipped without extra dlls.
> > 
> > And how exactly does that differ from what I said?  It is building in 
> > Cygwin.
> 
> With the MinGW headers. That makes a big difference.
> Compiling in cygwin with -mno-cygwin is almost the same as compiling
> with MinGW tools (the difference mostly lies in Makefile and shell
> script subtleties)

Okay, I failed again to make myself understood.

If you say "it fixes building on Windows", while many people have no 
problem whatsoever building it with MinGW, and other people who might want 
to build it with Visual Studio or Borland or whatever will still have a 
ton of problems, is ever so _slightly_ misleading.

Okay?

Thanks,
Dscho

Re: [Qemu-devel] another patch against head

[Hotmail]

----- Original Message ----- 
From: "Johannes Schindelin" <Johannes.Schindelin@gmx.de>
To: "Hotmail" <computers57@hotmail.com>
Cc: "qemu Devel" <qemu-devel@nongnu.org>
Sent: Wednesday, November 28, 2007 5:49 AM
Subject: Re: [Qemu-devel] another patch against head


> Hi Hotmail,
> 
Hello.  Please call me C.W.
> ...
> > (patch to cocoa.m)
> How does this help compiling on Windows?
It doesn't.  It's just a little patch should someone decide to build it on cocoa and are confused
> 
> > (patch against monitor.c and hw/adlib.c)
> Why are these needed?  It appears that all other platforms build QEmu just 
> fine without these #includes.
I also seem to have this problem on my Linux box.  It could be my cvs files has become disjoined from the server, or both compilers somehow don't include the needed headers elsewhere in the code.
> But then, the answer to that riddle could be in the context.  Since you 
> did not use the unified diff format, I have no chance to know.
How would I make a unified diff from a command line?

Re: [Qemu-devel] another patch against head

[Johannes Schindelin]

On Wed, 28 Nov 2007, Hotmail wrote:

> > But then, the answer to that riddle could be in the context.  Since 
> > you did not use the unified diff format, I have no chance to know.
>
> How would I make a unified diff from a command line?

cvs diff -u

Hth,
Dscho

Re: [Qemu-devel] another patch against head

[Hotmail]

[-- Attachment #1.1: Type: text/plain, Size: 1376 bytes --]

Also, if you disable profiling, the monitor.c file doesnt compile.
And due to request, the unified diff is on this message (as an attachment)
----- Original Message ----- 
From: "Sylvain Petreolle" <spetreolle@yahoo.fr>
To: <qemu-devel@nongnu.org>
Sent: Wednesday, November 28, 2007 10:28 AM
Subject: Re : [Qemu-devel] another patch against head


Hi Johannes,
----- Message d'origine ----
> De : Johannes Schindelin <Johannes.Schindelin@gmx.de>
> À : Hotmail <computers57@hotmail.com>
> Cc : qemu Devel <qemu-devel@nongnu.org>
> Envoyé le : Mercredi, 28 Novembre 2007, 13h49mn 34s
> Objet : Re: [Qemu-devel] another patch against head
> > Index: hw/adlib.c
> > ===================================================================
> > RCS file: /sources/qemu/qemu/hw/adlib.c,v
> > retrieving revision 1.8
> > diff -r1.8 adlib.c
> > 25a26
> > > #include "audio/audio.h"
> > 26a28
> > > #include "qemu-timer.h"
> 
> Why are these needed?  It appears that all other platforms build
> QEmu
> 
 just 
> fine without these #includes.
> 
> But then, the answer to that riddle could be in the context.  Since
> you
> 
 
> did not use the unified diff format, I have no chance to know.
> 
> Ciao,
> Dscho
> 
Try enabling the adlib emulation, it doesnt build either for me on llinux now.
this include was forgotten in vl.h refactoring.





[-- Attachment #1.2: Type: text/html, Size: 2844 bytes --]

[-- Attachment #2: udiff.txt --]
[-- Type: text/plain, Size: 2513 bytes --]

? cocoa-patch.m
? cvsdiff.diff
? dlls.txt
? error.txt
? set.txt
? test-powerpc.cpp
? test-ppc
? udiff.diff
? pc-bios/PM.ROM
? pc-bios/g3.rom
? target-i386/cpu.h.bad
? target-i386/ops_sse.h.bad
? target-i386/translate.c.bad
Index: Makefile
===================================================================
RCS file: /sources/qemu/qemu/Makefile,v
retrieving revision 1.136
diff -u -r1.136 Makefile
--- Makefile	24 Nov 2007 23:35:07 -0000	1.136
+++ Makefile	30 Nov 2007 15:29:08 -0000
@@ -15,7 +15,11 @@
 
 CPPFLAGS += -I. -I$(SRC_PATH) -MMD -MP
 CPPFLAGS += -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
+ifdef CONFIG_WIN32
+LIBS= -lmingw32 -mno-cygwin
+else
 LIBS=
+endif
 ifdef CONFIG_STATIC
 BASE_LDFLAGS += -static
 endif
Index: cocoa.m
===================================================================
RCS file: /sources/qemu/qemu/cocoa.m,v
retrieving revision 1.14
diff -u -r1.14 cocoa.m
--- cocoa.m	17 Nov 2007 17:14:37 -0000	1.14
+++ cocoa.m	30 Nov 2007 15:29:09 -0000
@@ -751,7 +751,7 @@
 
         [op setMessage:@"Select the disk image you want to boot.\n\nHit the \"Cancel\" button to quit"];
 
-        [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"cow",@"cloop",@"vmdk",nil]
+        [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"qcow2",@"cow",@"cloop",@"vmdk",nil]
               modalForWindow:window modalDelegate:self
               didEndSelector:@selector(openPanelDidEnd:returnCode:contextInfo:) contextInfo:NULL];
     }
Index: monitor.c
===================================================================
RCS file: /sources/qemu/qemu/monitor.c,v
retrieving revision 1.87
diff -u -r1.87 monitor.c
--- monitor.c	18 Nov 2007 01:44:35 -0000	1.87
+++ monitor.c	30 Nov 2007 15:29:10 -0000
@@ -29,6 +29,7 @@
 #include "gdbstub.h"
 #include "net.h"
 #include "qemu-char.h"
+#include "qemu-timer.h"
 #include "sysemu.h"
 #include "console.h"
 #include "block.h"
Index: hw/adlib.c
===================================================================
RCS file: /sources/qemu/qemu/hw/adlib.c,v
retrieving revision 1.8
diff -u -r1.8 adlib.c
--- hw/adlib.c	17 Nov 2007 17:14:40 -0000	1.8
+++ hw/adlib.c	30 Nov 2007 15:29:10 -0000
@@ -23,7 +23,9 @@
  */
 #include <assert.h>
 #include "hw.h"
+#include "audio/audio.h"
 #include "audiodev.h"
+#include "qemu-timer.h"
 
 #define ADLIB_KILL_TIMERS 1
 

Re: [Qemu-devel] another patch against head

[Thiemo Seufer]

Hotmail wrote:
[snip]
> Index: Makefile
> ===================================================================
> RCS file: /sources/qemu/qemu/Makefile,v
> retrieving revision 1.136
> diff -u -r1.136 Makefile
> --- Makefile	24 Nov 2007 23:35:07 -0000	1.136
> +++ Makefile	30 Nov 2007 15:29:08 -0000
> @@ -15,7 +15,11 @@
>  
>  CPPFLAGS += -I. -I$(SRC_PATH) -MMD -MP
>  CPPFLAGS += -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
> +ifdef CONFIG_WIN32
> +LIBS= -lmingw32 -mno-cygwin

This should be done in the configure script via OS_LDFLAGS.

> +else
>  LIBS=
> +endif
>  ifdef CONFIG_STATIC
>  BASE_LDFLAGS += -static
>  endif
> Index: cocoa.m
> ===================================================================
> RCS file: /sources/qemu/qemu/cocoa.m,v
> retrieving revision 1.14
> diff -u -r1.14 cocoa.m
> --- cocoa.m	17 Nov 2007 17:14:37 -0000	1.14
> +++ cocoa.m	30 Nov 2007 15:29:09 -0000
> @@ -751,7 +751,7 @@
>  
>          [op setMessage:@"Select the disk image you want to boot.\n\nHit the \"Cancel\" button to quit"];
>  
> -        [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"cow",@"cloop",@"vmdk",nil]
> +        [op beginSheetForDirectory:nil file:nil types:[NSArray arrayWithObjects:@"img",@"iso",@"dmg",@"qcow",@"qcow2",@"cow",@"cloop",@"vmdk",nil]

That's still incomplete WT to object formats supported, I leave that out
until somebody actually completed and tested it.

>                modalForWindow:window modalDelegate:self
>                didEndSelector:@selector(openPanelDidEnd:returnCode:contextInfo:) contextInfo:NULL];
>      }
> Index: monitor.c
> ===================================================================
> RCS file: /sources/qemu/qemu/monitor.c,v
> retrieving revision 1.87
> diff -u -r1.87 monitor.c
> --- monitor.c	18 Nov 2007 01:44:35 -0000	1.87
> +++ monitor.c	30 Nov 2007 15:29:10 -0000
> @@ -29,6 +29,7 @@
>  #include "gdbstub.h"
>  #include "net.h"
>  #include "qemu-char.h"
> +#include "qemu-timer.h"

I don't see where the monitor would use a timer.
Can you post the compiler error message which is solved by this include?

>  #include "sysemu.h"
>  #include "console.h"
>  #include "block.h"
> Index: hw/adlib.c
> ===================================================================
> RCS file: /sources/qemu/qemu/hw/adlib.c,v
> retrieving revision 1.8
> diff -u -r1.8 adlib.c
> --- hw/adlib.c	17 Nov 2007 17:14:40 -0000	1.8
> +++ hw/adlib.c	30 Nov 2007 15:29:10 -0000
> @@ -23,7 +23,9 @@
>   */
>  #include <assert.h>
>  #include "hw.h"
> +#include "audio/audio.h"

Added.

>  #include "audiodev.h"
> +#include "qemu-timer.h"

Only needed for #define DEBUG, I added a variant which takes this
into account.


Thiemo