colinux-devel — Cooperative Linux development mailing list

[coLinux-devel] 20040208 kernel load works, CPU state

[Ballard J. <sac...@ho...>]

I tried the 20040208 kernel on my system where the 20040131 previously
crashed, and it boots fine.

Something to think about, after I reloaded the kernel on my other system, I
only had 1gb left on the drive with 34% fragmentation.  I started the kernel
and ran top.  I started the defrag under winnt.  The top program reports
over 95% CPU idle.   Obviously, the CPU state is an incorrect statement, as
it is only the state of linux processes which where 95% idle.  I know the
defrag would make the system more than 50% busy.  I doubt that is a coLinux
issue.

Re: [coLinux-devel] coLinux 0.5.3-pre1

[Richard G. <ric...@ri...>]

Something appeared to be odd with the networking (at least the first few
times I tried).  If I started with -d, I could ping the cL session, but if I
let the console autostart, I can't.

AFter some more playing around, it seems like when I close the console
sometimes networking dies. (killing console with the X).

Anyone else?

----- Original Message ----- 
From: "Dan Aloni" <da...@gm...>
To: "Cooperative Linux Development" <col...@li...>
Sent: Saturday, February 07, 2004 6:19 PM
Subject: [coLinux-devel] coLinux 0.5.3-pre1


> Hello,
>
> After two days of intensive coding, I'm releasing this pre-release. The
> reason for is that although it appears to be working, I'm not finished
> cleaning it up, so I might missed some real issues.
>
> The files for this release on sourceforge were added to the 0.5.2
> release in order to avoid attention from stray users :)
>
> Here are the list of changes:
>
> * Version 0.5.3-pre1 AKA "flamed chicken"
>   * Fixed a BSOD bug (thanks Richard Goodwin for the memory dump stacks)
>   * No longer depending on the Microsoft DDK for the build (review
>     doc/building), (thanks Nir Perry for the encourgement)
>   * The daemon launches the console by default when it is launched (can be
>     turned off with '-d', see '-h').
>   * The daemon is killable just like a normal Windows process. Doing so
stops
>     the Linux machines cleanly (from Windows' standpoint), so feel free
>     to close the window, end the task, CTRL-C, etc.
>   * It no longer depends on the Microsoft DDK for the build.
>   * An almost complete rewrite - we now use message passing instead of
>     hardwired ioctls(). Only the daemon communicates with the driver. The
>     console and the network daemon communicate with the daemon using a
>     named pipe.
>   * Fixed the block device bug that fsck triggered (thanks Daniel
Glockner)
>   * Hopefully addressed the network corruption issue (thanks Daniel
Glockner
>      again)
>   * Boot parameters are now passed to the kernel. Make sure you update
your
>     XMLs with 'root=/dev/cobd0' if you still want to boot from the first
>     block device.
>   * The conet daemon is a separate process (shouldn't affect you).
>
>
> -- 
> Dan Aloni
> da...@gm...
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel
>
>

Re: [coLinux-devel] coLinux 0.5.3-pre1

[Thomas F. <tf...@no...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Dan,

cool (pre)release. Sounds like all major problem are solved?! I compiled
your *falmed chicken* with success with linux. But sorry... I have
currently no windows here, so I could not test the binary so far.

Just one question to the colinux-community: This is about the tab-win32
"net"-driver. I have this tap-win32 driver in the colinux tree but it
looks like this driver is not build in colinux? Also in the binary-distro
this files come with other date/time. Is this the driver from the openvpn
projekt?
Did you build this drivers from windows or from linux?

So long and thanks for the chicken...
  Thomas



> Hello,
>
> After two days of intensive coding, I'm releasing this pre-release. The
> reason for is that although it appears to be working, I'm not finished
> cleaning it up, so I might missed some real issues.
>
> The files for this release on sourceforge were added to the 0.5.2
> release in order to avoid attention from stray users :)
>
> Here are the list of changes:
>
> * Version 0.5.3-pre1 AKA "flamed chicken"
>   * Fixed a BSOD bug (thanks Richard Goodwin for the memory dump stacks)
>   * No longer depending on the Microsoft DDK for the build (review
>     doc/building), (thanks Nir Perry for the encourgement)
>   * The daemon launches the console by default when it is launched (can be
>     turned off with '-d', see '-h').
>   * The daemon is killable just like a normal Windows process. Doing so
> stops
>     the Linux machines cleanly (from Windows' standpoint), so feel free
>     to close the window, end the task, CTRL-C, etc.
>   * It no longer depends on the Microsoft DDK for the build.
>   * An almost complete rewrite - we now use message passing instead of
>     hardwired ioctls(). Only the daemon communicates with the driver. The
>     console and the network daemon communicate with the daemon using a
>     named pipe.
>   * Fixed the block device bug that fsck triggered (thanks Daniel
> Glockner)
>   * Hopefully addressed the network corruption issue (thanks Daniel
> Glockner
>      again)
>   * Boot parameters are now passed to the kernel. Make sure you update
> your
>     XMLs with 'root=/dev/cobd0' if you still want to boot from the first
>     block device.
>   * The conet daemon is a separate process (shouldn't affect you).
>
>
> --
> Dan Aloni
> da...@gm...
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel
>


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAJYwinwJBIFTVIqwRAl83AJwOpNZhX80cI0sdwBMJ8jS/L7rkYQCfXwuK
TvcrjAZhTCKT4hWcCvJeUjM=
=RZLJ
-----END PGP SIGNATURE-----

[coLinux-devel] coLinux 0.5.3-pre1

[Dan A. <da...@gm...>]

Hello,

After two days of intensive coding, I'm releasing this pre-release. The 
reason for is that although it appears to be working, I'm not finished 
cleaning it up, so I might missed some real issues. 

The files for this release on sourceforge were added to the 0.5.2 
release in order to avoid attention from stray users :)

Here are the list of changes:

* Version 0.5.3-pre1 AKA "flamed chicken"
  * Fixed a BSOD bug (thanks Richard Goodwin for the memory dump stacks)
  * No longer depending on the Microsoft DDK for the build (review 
    doc/building), (thanks Nir Perry for the encourgement)
  * The daemon launches the console by default when it is launched (can be 
    turned off with '-d', see '-h').
  * The daemon is killable just like a normal Windows process. Doing so stops 
    the Linux machines cleanly (from Windows' standpoint), so feel free 
    to close the window, end the task, CTRL-C, etc.
  * It no longer depends on the Microsoft DDK for the build.
  * An almost complete rewrite - we now use message passing instead of 
    hardwired ioctls(). Only the daemon communicates with the driver. The 
    console and the network daemon communicate with the daemon using a 
    named pipe.
  * Fixed the block device bug that fsck triggered (thanks Daniel Glockner)
  * Hopefully addressed the network corruption issue (thanks Daniel Glockner
     again)
  * Boot parameters are now passed to the kernel. Make sure you update your 
    XMLs with 'root=/dev/cobd0' if you still want to boot from the first 
    block device. 
  * The conet daemon is a separate process (shouldn't affect you).

	
-- 
Dan Aloni
da...@gm...

[coLinux-devel] Legacy UNIX layout vs a coLinux potential

[JONATHAN B. <sac...@ho...>]

A lot of software on UNIX is designed with the expectation of the legacy 
layout of files, and we could change that.  What if device driver writers 
could write once and have the driver work for both Linux and Windows.  That 
would solve many issues.  That is where I know coLinux will succeed.

A new base of driver support could mean a change in legacy file system 
layout.  The system would still be unix/linux based, but some of it would 
look like extras or just legacy support.  I don't think of it as a utopian 
format for such layout.  I look at it more to fit in the realm of the design 
of the project name "cooperative."  If the designers of Windows and the 
designers of Linux can not cooperate, then what good is their business to 
me?

I thought about how much this would affect business.  To run Windows and 
Linux in parallel mode, rather than in serial mode, is the sole reason why I 
decided to join in and cooperate myself.  I do not think I am going to be 
the only one with such attitude.  Why fight about if Linux or Windows is 
initially installed on a purchased PC when they could both be available.  No 
doubt, Microsoft has a larger domain of device driver writers than Linux, so 
coLinux may take direct advantage of that.  What is available to coLinux is 
larger than what is available to a standalone linux machine.

I do not have to write my software, Atomatrix, in a seperate fashion to make 
it available for both Linux or Windows.  I can write it once and make it 
available for both Linux and Windows.  With a schedule like mine, that saves 
me lots of time.  I also have much potential to advance the design of 
Atomatrix and coLinux with the tools that I already have, and that saves me 
money.

A newer layout of the the filesystem may be the challenge to a greater 
cooperation.  I am not going to settle for linux emulators that require a 
reboot of my windows system or that require me to repartition the drive.  
And, I am sure I am not the only one.

Good work.

_________________________________________________________________
Check out the great features of the new MSN 9 Dial-up, with the MSN Dial-up 
Accelerator. http://click.atdmt.com/AVE/go/onm00200361ave/direct/01/

Re: [coLinux-devel] Memory size and killing console by Esc

[Dan A. <da...@gm...>]

On Sat, Feb 07, 2004 at 09:26:14PM -0000, Thomas Fritzsche wrote:

> if I understand you right you can'n allocate the memory from within the
> kernel space, but what abount a kind of ramdisk (in userspace). Such a ram
> disc could  then be used as swap-partition in the linux world. Of course
> there is much overhead to transfer memory always from userspace but maybe
> this is nevertheless faster then reading/writing to hd?

While it is already possible to use a Windows file as swap space for 
Linux, future versions of coLinux will use a Windows process to allocate 
virtual userspace memory, without using a file.

-- 
Dan Aloni
da...@gm...

Re: [coLinux-devel] Memory size and killing console by Esc

[Thomas F. <tf...@no...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Dan,

if I understand you right you can'n allocate the memory from within the
kernel space, but what abount a kind of ramdisk (in userspace). Such a ram
disc could  then be used as swap-partition in the linux world. Of course
there is much overhead to transfer memory always from userspace but maybe
this is nevertheless faster then reading/writing to hd?

Thanks and regards,
 Thomas

> On Sat, Feb 07, 2004 at 12:52:31AM +0200, Paul Sokolovsky wrote:
>
>> 1. Could memory size for coLinux be configurable? Current 24Mb or so
>> is a joke. Yes, swap can be mounted, but why experience perfomance of
>> vmware level? ;-) It would be nice to have memory size configurable in
>> XML config, but if it's some time to do, could next release ship with
>> at least slightly decent amount, like 64Mb (128Mb better ;-).
>
> The major problem with what you are suggesting, is the limitations
> enforced by Windows' non cached memory allocator. This pool is
> hardcoded to a limit of 256MB and even 128MB on some systems. Allocating
> too much memory from that pool can destabilze Windows.


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAJVf1nwJBIFTVIqwRAk4xAKCOSFT0pMlHjXRUWwe2LC37N1TNewCgxdFo
wjcCXjLZ5HpvQGbZmoIxn0s=
=X7kG
-----END PGP SIGNATURE-----

Re: [coLinux-devel] Memory size and killing console by Esc

[andre <avb...@gm...>]

On Saturday 07 February 2004 19:46, you wrote:
> There are a few "lin4win" titles on the net.  Do you mean to use lin4win,
> the emulator in a window, to emulate linux first to do the initial install
> of the disk image?
>
> Could you send the url of the one you meant?
>

lin4win A.K.A. lnx4win 


It is when you install linux in a loopback partition on a windows fat 
partition. If you install Mandrake 9.2 as expert it is the bottom choice of 
which partition you want to use for your installation 

Re: [coLinux-devel] Memory size and killing console by Esc

[andre <avb...@gm...>]

On Saturday 07 February 2004 17:54, you wrote:
> > > We will distribue further images in the future.
> >
> > Would rather know what has changed between a normal install of deb and
>
> this
>
> > image.
>
> Perhaps, we could include a first time startup configuration.  It would be
> something like the initial Debian config but geared for coLinux.  It could
> proportionalize the size of the disk.  Then, it could put the user into
> dselect to choose how to get more files.
>
> Jonathan

To make an image on a vfat partition is trivial when you use lin4win. I think
Xandros still supports lin4win and i know Mandrake does. So then you only
have to install the colinux kernel and edit some settings

Re: [coLinux-devel] cross probs while building fltk

[Thomas F. <tf...@no...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Dan Aloni + colinux-hackers,

thanks a lot for this patch. Meanwhile I was just on the same way, but you
saved a lot of time with it. With this patch I could build fltk and that
was the final step for the colinux-system (without DDK ;-))!
The only thing that didn't work was the installation of the fltk
header-files I copied them by hand in my cross-building directory and that
was it.

But what a brummer! I don't have a windows-box here this weekend to test
this builds :-(. Hmmm but maybe it's better to wait for the next
release... (today? tomorrow?)

One other thing: I'm thinking about to write a shell script to automate
this building proccess. Is already someone working on this issue?

Thanks and kind regards,
 Thomas




> On Sat, Feb 07, 2004 at 01:52:59PM -0000, Thomas Fritzsche wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Hi colinux-hackers,
>>
>> I build cygwin tools as described in cygwin-cross-build without
>> problems.
>> I also compiled mxlm with this cygwin-tools, but now I have some
>> problems
>> with building fltk. This is because fltk has a lot of depends on other
>> libs for example jpeglib or stdc++ and it looks like the system would
>> like
>> to build a version for X11.
>>
>> I can disable a few libs in the configuration but I could still not
>> build
>> fltk.
>>
>> What parameters do you use to cross build fltk?
>
> You need to use the mingw makefiles with a bit of modifications.
> Applying the patch below does the trick.
>
> diff -urN fltk-1.1.4/Makefile fltk-1.1.4-patched/Makefile
> --- fltk-1.1.4/Makefile	2003-06-15 08:39:12.000000000 +0300
> +++ fltk-1.1.4-patched/Makefile	2001-01-22 17:13:39.000000000 +0200
> @@ -1,9 +1,9 @@
>  #
> -# "$Id: Makefile,v 1.12.2.6.2.16 2003/06/15 05:39:12 easysw Exp $"
> +# "$Id: Makefile.mingw,v 1.1.2.2 2001/01/22 15:13:39 easysw Exp $"
>  #
>  # Top-level makefile for the Fast Light Tool Kit (FLTK).
>  #
> -# Copyright 1998-2003 by Bill Spitzak and others.
> +# Copyright 1998-2001 by Bill Spitzak and others.
>  #
>  # This library is free software; you can redistribute it and/or
>  # modify it under the terms of the GNU Library General Public
> @@ -23,79 +23,74 @@
>  # Please report all bugs and problems to "flt...@fl...".
>  #
>
> -include makeinclude
> -
> -DIRS	=	src fluid test documentation
> -
> -all: makeinclude
> -	for dir in $(DIRS); do\
> -		echo "=== making $$dir ===";\
> -		(cd $$dir; $(MAKE) $(MFLAGS)) || break;\
> -	done
> -
> -install: makeinclude
> -	-mkdir -p $(bindir)
> -	$(RM) $(bindir)/fltk-config
> -	-cp fltk-config $(bindir)
> -	-chmod 755 $(bindir)/fltk-config
> -	for dir in FL $(DIRS); do\
> -		echo "=== installing $$dir ===";\
> -		(cd $$dir; $(MAKE) $(MFLAGS) install) || break;\
> -	done
> -
> -uninstall: makeinclude
> -	$(RM) $(bindir)/fltk-config
> -	for dir in FL $(DIRS); do\
> -		echo "=== uninstalling $$dir ===";\
> -		(cd $$dir; $(MAKE) $(MFLAGS) uninstall) || break;\
> -	done
> -
> -depend: makeinclude
> -	for dir in $(DIRS); do\
> -		echo "=== making dependencies in $$dir ===";\
> -		(cd $$dir; $(MAKE) $(MFLAGS) depend) || break;\
> -	done
> +# By default use a Un*x-style shell
> +#
> +# uncomment next line to use NT command shell
> +# SHELL=cmd.exe
> +#
> +# or uncomment next line to explicitly use a unix-compatible shell
> +# SHELL=sh.exe
> +#
> +# If we are using a Un*x-based make, make sure the environmental var is
> set
> +# set MAKE_MODE=unix
> +#
> +# By default, a Un*x style make is assumed, so towards the bottom of this
> +# file filenames are copied with 'cp' and use forward slashes throughout
> +# in filepaths.
> +#
> +# Changing to a Windows-based command line would require 'cp' replaced
> with
> +# a local equivalent ('copy' is OK) and potentially for forward-slashes
> in
> +# filepaths to be replaced by back-slashes.
> +
> +
> +all: makeinclude config.h
> +	echo "=== making src ==="
> +	touch src/makedepend
> +	cd src ; $(MAKE)
> +	echo "=== making fluid ==="
> +	touch fluid/makedepend
> +	cd fluid ; $(MAKE)
> +	echo "=== making test ==="
> +	touch test/makedepend
> +	cd test ; $(MAKE)
> +
> +install:
> +	echo "=== installing src ==="
> +	touch src/makedepend
> +	cd src ; $(MAKE) install
> +	echo "=== installing fluid ==="
> +	touch fluid/makedepend
> +	cd fluid ; $(MAKE) install
> +
> +depend:
> +	echo "=== making src dependencies ==="
> +	touch src/makedepend
> +	cd src ; $(MAKE) depend
> +	echo "=== making fluid dependencies ==="
> +	touch fluid/makedepend
> +	cd fluid ; $(MAKE) depend
> +	echo "=== making test dependencies ==="
> +	touch test/makedepend
> +	cd test ; $(MAKE) depend
>
>  clean:
> -	-$(RM) core *.o
> -	for dir in $(DIRS); do\
> -		echo "=== cleaning $$dir ===";\
> -		(cd $$dir; $(MAKE) $(MFLAGS) clean) || break;\
> -	done
> -
> -distclean: clean
> -	$(RM) config.*
> -	$(RM) fltk-config fltk.list makeinclude
> -	$(RM) FL/Makefile
> -	$(RM) documentation/*.$(CAT1EXT)
> -	$(RM) documentation/*.$(CAT3EXT)
> -	$(RM) documentation/fltk.pdf
> -	$(RM) documentation/fltk.ps
> -	$(RM) -r documentation/fltk.d
> -	for file in test/*.fl; do\
> -		$(RM) test/`basename $file .fl`.cxx; \
> -		$(RM) test/`basename $file .fl`.h; \
> -	done
> -
> -makeinclude: configure configh.in makeinclude.in
> -	if test -f config.status; then \
> -		./config.status --recheck; \
> -		./config.status; \
> -	else \
> -		./configure; \
> -	fi
> -	touch config.h
> -
> -configure: configure.in
> -	autoconf
> -
> -portable-dist:
> -	epm -v -s fltk.xpm fltk
> +	-@ rm -f core config.cache *.o *.bck
> +	echo "=== cleaning src ==="
> +	touch src/makedepend
> +	cd src ; $(MAKE) clean
> +	echo "=== cleaning fluid ==="
> +	touch fluid/makedepend
> +	cd fluid ; $(MAKE) clean
> +	echo "=== cleaning test ==="
> +	touch test/makedepend
> +	cd test ; $(MAKE) clean
>
> -native-dist:
> -	epm -v -f native fltk
> +config.h: makefiles/config.mingw
> +	cp $< $@
>
> +makeinclude: makefiles/makeinclude.mingw
> +	cp $< $@
>
>  #
> -# End of "$Id: Makefile,v 1.12.2.6.2.16 2003/06/15 05:39:12 easysw Exp
> $".
> +# End of "$Id: Makefile.mingw,v 1.1.2.2 2001/01/22 15:13:39 easysw Exp
> $".
>  #
> diff -urN fltk-1.1.4/config.h fltk-1.1.4-patched/config.h
> --- fltk-1.1.4/config.h	1970-01-01 02:00:00.000000000 +0200
> +++ fltk-1.1.4-patched/config.h	2003-01-30 23:41:07.000000000 +0200
> @@ -0,0 +1,219 @@
> +/*
> + * "$Id: config.mingw,v 1.1.2.3.2.10 2003/01/30 21:41:07 easysw Exp $"
> + *
> + * Configuration file for the Fast Light Tool Kit (FLTK).
> + *
> + * Copyright 1998-2003 by Bill Spitzak and others.
> + *
> + * This library is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Library General Public
> + * License as published by the Free Software Foundation; either
> + * version 2 of the License, or (at your option) any later version.
> + *
> + * This library is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + * Library General Public License for more details.
> + *
> + * You should have received a copy of the GNU Library General Public
> + * License along with this library; if not, write to the Free Software
> + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
> + * USA.
> + *
> + * Please report all bugs and problems to "FLT...@fl...".
> + */
> +
> +/*
> + * Where to find files...
> + */
> +
> +#define FLTK_DATADIR	"C:/FLTK"
> +#define FLTK_DOCDIR	"C:/FLTK/DOC"
> +
> +/*
> + * BORDER_WIDTH:
> + *
> + * Thickness of FL_UP_BOX and FL_DOWN_BOX.  Current 1,2, and 3 are
> + * supported.
> + *
> + * 3 is the historic FLTK look.
> + * 2 is the default and looks like Microsoft Windows, KDE, and Qt.
> + * 1 is a plausible future evolution...
> + *
> + * Note that this may be simulated at runtime by redefining the boxtypes
> + * using Fl::set_boxtype().
> + */
> +
> +#define BORDER_WIDTH 2
> +
> +/*
> + * HAVE_GL:
> + *
> + * Do you have OpenGL? Set this to 0 if you don't have or plan to use
> + * OpenGL, and FLTK will be smaller.
> + *
> + * In order to set to 1 you will need to obtain the OpenGL header and
> + * library files.  Consult the README.win32 file for further details.
> + */
> +
> +#define HAVE_GL 1
> +
> +/*
> + * HAVE_GL_GLU_H:
> + *
> + * Do you have the OpenGL Utility Library header file?
> + * (many broken Mesa RPMs do not...)
> + */
> +
> +#define HAVE_GL_GLU_H 1
> +
> +/*
> + * USE_COLORMAP:
> + *
> + * Setting this to zero will save a good deal of code (especially for
> + * fl_draw_image), but FLTK will only work on TrueColor visuals.
> + */
> +
> +#define USE_COLORMAP 1
> +
> +/*
> + * USE_XFT
> + *
> + * Use the new Xft library to draw anti-aliased text.
> + */
> +
> +#define USE_XFT 0
> +
> +/*
> + * HAVE_XDBE:
> + *
> + * Do we have the X double-buffer extension?
> + */
> +
> +#define HAVE_XDBE 0
> +
> +/*
> + * USE_XDBE:
> + *
> + * Actually try to use the double-buffer extension?  Set this to zero
> + * disable use of XDBE without breaking the list_visuals program.
> + */
> +
> +#define USE_XDBE HAVE_XDBE
> +
> +/*
> + * HAVE_OVERLAY:
> + *
> + * Use the X overlay extension?  FLTK will try to use an overlay
> + * visual for Fl_Overlay_Window, the Gl_Window overlay, and for the
> + * menus.  Setting this to zero will remove a substantial amount of
> + * code from FLTK.  Overlays have only been tested on SGI servers!
> + */
> +
> +#define HAVE_OVERLAY 0
> +
> +/*
> + * HAVE_GL_OVERLAY:
> + *
> + * It is possible your GL has an overlay even if X does not.  If so,
> + * set this to 1.
> + */
> +
> +#define HAVE_GL_OVERLAY HAVE_OVERLAY
> +
> +/*
> + * WORDS_BIGENDIAN:
> + *
> + * Byte order of your machine: 1 = big-endian, 0 = little-endian.
> + */
> +
> +#define WORDS_BIGENDIAN 0
> +
> +/*
> + * U16, U32, U64:
> + *
> + * Types used by fl_draw_image.  One of U32 or U64 must be defined.
> + * U16 is optional but FLTK will work better with it!
> + */
> +
> +#define U16 unsigned short
> +#define U32 unsigned
> +/* #undef U64 */
> +
> +/*
> + * HAVE_DIRENT_H, HAVE_SYS_NDIR_H, HAVE_SYS_DIR_H, HAVE_NDIR_H,
> HAVE_SCANDIR:
> + *
> + * Where is <dirent.h> (used only by fl_file_chooser and scandir).
> + */
> +
> +#define HAVE_DIRENT_H 1
> +/* #undef HAVE_SYS_NDIR_H */
> +/* #undef HAVE_SYS_DIR_H */
> +/* #undef HAVE_NDIR_H */
> +/* #undef HAVE_SCANDIR */
> +
> +/*
> + * Possibly missing sprintf-style functions:
> + */
> +
> +/* #undef HAVE_VSNPRINTF */
> +/* #undef HAVE_SNPRINTF */
> +
> +/*
> + * String functions...
> + */
> +
> +/* #undef HAVE_STRINGS_H */
> +#define HAVE_STRCASECMP 1
> +/* #undef HAVE_STRLCAT */
> +/* #undef HAVE_STRLCPY */
> +
> +/*
> + * HAVE_SYS_SELECT_H:
> + *
> + * Whether or not select() call has its own header file.
> + */
> +
> +/* #undef HAVE_SYS_SELECT_H */
> +
> +/*
> + * USE_POLL:
> + *
> + * Use poll() if we don't have select().
> + */
> +
> +#define USE_POLL 0
> +
> +/*
> + * Do we have various image libraries?
> + */
> +
> +/* #undef HAVE_LIBPNG */
> +/* #undef HAVE_LIBZ */
> +/* #undef HAVE_LIBJPEG */
> +
> +/*
> + * Which header file do we include for libpng?
> + */
> +
> +/* #undef HAVE_PNG_H */
> +/* #undef HAVE_LIBPNG_PNG_H */
> +
> +/*
> + * Do we have the png_xyz() functions?
> + */
> +
> +/* #undef HAVE_PNG_GET_VALID */
> +/* #undef HAVE_PNG_SET_TRNS_TO_ALPHA */
> +
> +/*
> + * Do we have POSIX threading?
> + */
> +
> +/* #undef HAVE_PTHREAD */
> +/* #undef HAVE_PTHREAD_H */
> +
> +
> +/*
> + * End of "$Id: config.mingw,v 1.1.2.3.2.10 2003/01/30 21:41:07 easysw
> Exp $".
> + */
> diff -urN fltk-1.1.4/makeinclude fltk-1.1.4-patched/makeinclude
> --- fltk-1.1.4/makeinclude	1970-01-01 02:00:00.000000000 +0200
> +++ fltk-1.1.4-patched/makeinclude	2004-01-30 23:50:47.000000000 +0200
> @@ -0,0 +1,126 @@
> +#
> +# "$Id: makeinclude.mingw,v 1.1.2.3.2.11 2002/07/14 17:03:31 easysw Exp
> $"
> +#
> +# Make include file for the Fast Light Tool Kit (FLTK).
> +# (this file for MingW using GCC 2.95.x)
> +#
> +# Copyright 1998-2001 by Bill Spitzak and others.
> +#
> +# This library is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU Library General Public
> +# License as published by the Free Software Foundation; either
> +# version 2 of the License, or (at your option) any later version.
> +#
> +# This library is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +# Library General Public License for more details.
> +#
> +# You should have received a copy of the GNU Library General Public
> +# License along with this library; if not, write to the Free Software
> +# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
> +# USA.
> +#
> +# Please report all bugs and problems to "flt...@fl...".
> +#
> +
> +# Notes:
> +#
> +# 1) Console windows appearing
> +#
> +#    Consoles will not appear when run from a command line prompt
> +#    but will cause an 'unnecessary' console to appear when run directly
> from
> +#    windows (say explorer). It provides i/o for stdin/stdout for
> debugging
> +#
> +#    To avoid this add "-mwindows" to LDLIBS and GLDLIBS below
> +#
> +# 2) Stripping the executable to reduce size
> +#
> +#    An external program may be run to strip out any debug symbols:
> +#    strip filename.exe
> +#
> +#    Alternatively add -s option to LDLIBS and GLDLIBS below to strip all
> +#    symbols. (or -S for stripping of debug symbols only). See 'ld' man
> page
> +#    for other useful link-time options.
> +#
> +# The default is to leave -mwindows and the stripping off, so as to
> facilitate
> +# debugging. For release it may be appropriate to add these additional
> switches
> +
> +
> +prefix          = /usr/local
> +exec_prefix     = ${prefix}
> +bindir          = ${exec_prefix}/bin
> +includedir      = ${prefix}/include
> +libdir          = ${exec_prefix}/lib
> +srcdir          = .
> +
> +# programs we use...
> +CHMOD		= :
> +CP		= cp
> +HTMLDOC		= htmldoc.exe
> +LN		= :
> +MKDIR		= mkdir
> +NROFF		= nroff
> +RM		= rm
> +RMDIR		= rmdir
> +STRIP		= :
> +
> +# compiler names:
> +CXX		= i686-pc-cygwin-g++
> +CC		= i686-pc-cygwin-gcc
> +MAKEDEPEND	= $(CXX) -M
> +POSTBUILD	= :
> +
> +# flags for C++ compiler:
> +CFLAGS		= -Wall -O2 -DWIN32
> +CXXFLAGS	= -Wall -O2 -DWIN32
> +
> +# program to make the archive:
> +LIBNAME		= ../lib/libfltk.a
> +FLLIBNAME	= ../lib/libfltk_forms.a
> +GLLIBNAME	= ../lib/libfltk_gl.a
> +IMGLIBNAME	= ../lib/libfltk_images.a
> +LIBCOMMAND	= i686-pc-cygwin-ar -ruv
> +RANLIB		= i686-pc-cygwin-ranlib
> +DSONAME		=
> +FLDSONAME	=
> +GLDSONAME	=
> +IMGDSONAME	=
> +DSOCOMMAND	= echo
> +
> +# libraries to link with (in addition to default libs):
> +LDLIBS		=-mwindows -lgdi32 -lcomctl32 -lwsock32 -luuid -lole32
> +GLDLIBS		=-mwindows -lglu32 -lopengl32 -lgdi32 -lcomctl32 -lwsock32
> -luuid -lole32
> +LINKFLTK	=-L../lib -lfltk
> +LINKFLTKGL	=-L../lib -lfltk_gl
> +LINKFLTKFORMS	=-L../lib -lfltk_forms -lfltk
> +LINKFLTKIMG	=-L../lib -lfltk_images -lfltk
> +IMAGELIBS	=
> +
> +# Do we build the OpenGL demos?
> +GLDEMOS		=gldemos
> +
> +# Do we build the threads demo?
> +THREADS		=threads
> +
> +# Be quiet when building...
> +.SILENT:
> +
> +# Build commands and filename extensions...
> +.SUFFIXES:	.c .cxx .h .fl .o
> +
> +.cxx:
> +	echo Compiling and linking $@...
> +	$(CXX) -I.. $(CXXFLAGS) $< $(LINKFLTK) $(LDLIBS) -o $@
> +
> +.c.o:
> +	echo Compiling $@...
> +	$(CC) -I.. $(CXXFLAGS) $< -c
> +
> +.cxx.o:
> +	echo Compiling $@...
> +	$(CXX) -I.. $(CXXFLAGS) $< -c
> +
> +#
> +# End of "$Id: makeinclude.mingw,v 1.1.2.3.2.11 2002/07/14 17:03:31
> easysw Exp $".
> +#
> diff -urN fltk-1.1.4/src/Fl_win32.cxx fltk-1.1.4-patched/src/Fl_win32.cxx
> --- fltk-1.1.4/src/Fl_win32.cxx	2003-06-15 07:27:35.000000000 +0300
> +++ fltk-1.1.4-patched/src/Fl_win32.cxx	2004-01-30 23:40:23.000000000
> +0200
> @@ -47,7 +47,6 @@
>  // The following include files require GCC 3.x or a non-GNU compiler...
>  #if !defined(__GNUC__) || __GNUC__ >= 3
>  #  include <ole2.h>
> -#  include <ShellApi.h>
>  #endif // !__GNUC__ || __GNUC__ >= 3
>
>
> diff -urN fltk-1.1.4/src/filename_list.cxx
> fltk-1.1.4-patched/src/filename_list.cxx
> --- fltk-1.1.4/src/filename_list.cxx	2003-01-30 23:43:21.000000000 +0200
> +++ fltk-1.1.4-patched/src/filename_list.cxx	2004-01-30 23:40:59.000000000
> +0200
> @@ -51,7 +51,7 @@
>                       Fl_File_Sort_F *sort) {
>  #if defined(__hpux) || defined(__CYGWIN__)
>    // HP-UX, Cygwin define the comparison function like this:
> -  return scandir(d, list, 0, (int(*)(const dirent **, const dirent
> **))sort);
> +  return scandir(d, list, 0, (int(*)(dirent **, dirent **))sort);
>  #elif defined(__osf__)
>    // OSF, DU 4.0x
>    return scandir(d, list, 0, (int(*)(dirent **, dirent **))sort);
>
>
> --
> Dan Aloni
> da...@gm...
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel
>


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAJSVXnwJBIFTVIqwRAhaJAKCIos2ZQ2LhQtQOgOlvb3TzSvkqewCglTXj
7dw26Y7RoAhGDrRptf+FSnM=
=UHdn
-----END PGP SIGNATURE-----

Re: [coLinux-devel] Memory size and killing console by Esc

[andre <avb...@gm...>]

On Saturday 07 February 2004 00:42, Dan Aloni wrote:
> On Sat, Feb 07, 2004 at 12:52:31AM +0200, Paul Sokolovsky wrote:
> >   Ah, and yes - default image provided is pretty bare. It would be
> > nice to have at least ftpd and telnetd, and less included - it's not
> > always suitable to do apt-get (for example, for people with networking
> > problems). I guess, that's request for some kind sole who can
> > contribute such an image, coLinux author has other stuff to do ;-)
>
> We will distribue further images in the future.

Would rather know what has changed between a normal install of deb and this 
image.

Re: [coLinux-devel] cross probs while building fltk

[Dan A. <da...@gm...>]

On Sat, Feb 07, 2004 at 01:52:59PM -0000, Thomas Fritzsche wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi colinux-hackers,
> 
> I build cygwin tools as described in cygwin-cross-build without problems.
> I also compiled mxlm with this cygwin-tools, but now I have some problems
> with building fltk. This is because fltk has a lot of depends on other
> libs for example jpeglib or stdc++ and it looks like the system would like
> to build a version for X11.
> 
> I can disable a few libs in the configuration but I could still not build
> fltk.
> 
> What parameters do you use to cross build fltk?

You need to use the mingw makefiles with a bit of modifications. 
Applying the patch below does the trick.

diff -urN fltk-1.1.4/Makefile fltk-1.1.4-patched/Makefile
--- fltk-1.1.4/Makefile	2003-06-15 08:39:12.000000000 +0300
+++ fltk-1.1.4-patched/Makefile	2001-01-22 17:13:39.000000000 +0200
@@ -1,9 +1,9 @@
 #
-# "$Id: Makefile,v 1.12.2.6.2.16 2003/06/15 05:39:12 easysw Exp $"
+# "$Id: Makefile.mingw,v 1.1.2.2 2001/01/22 15:13:39 easysw Exp $"
 #
 # Top-level makefile for the Fast Light Tool Kit (FLTK).
 #
-# Copyright 1998-2003 by Bill Spitzak and others.
+# Copyright 1998-2001 by Bill Spitzak and others.
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Library General Public
@@ -23,79 +23,74 @@
 # Please report all bugs and problems to "flt...@fl...".
 #
 
-include makeinclude
-
-DIRS	=	src fluid test documentation
-
-all: makeinclude
-	for dir in $(DIRS); do\
-		echo "=== making $$dir ===";\
-		(cd $$dir; $(MAKE) $(MFLAGS)) || break;\
-	done
-
-install: makeinclude
-	-mkdir -p $(bindir)
-	$(RM) $(bindir)/fltk-config
-	-cp fltk-config $(bindir)
-	-chmod 755 $(bindir)/fltk-config
-	for dir in FL $(DIRS); do\
-		echo "=== installing $$dir ===";\
-		(cd $$dir; $(MAKE) $(MFLAGS) install) || break;\
-	done
-
-uninstall: makeinclude
-	$(RM) $(bindir)/fltk-config
-	for dir in FL $(DIRS); do\
-		echo "=== uninstalling $$dir ===";\
-		(cd $$dir; $(MAKE) $(MFLAGS) uninstall) || break;\
-	done
-
-depend: makeinclude
-	for dir in $(DIRS); do\
-		echo "=== making dependencies in $$dir ===";\
-		(cd $$dir; $(MAKE) $(MFLAGS) depend) || break;\
-	done
+# By default use a Un*x-style shell
+#
+# uncomment next line to use NT command shell 
+# SHELL=cmd.exe
+#
+# or uncomment next line to explicitly use a unix-compatible shell
+# SHELL=sh.exe
+#
+# If we are using a Un*x-based make, make sure the environmental var is set
+# set MAKE_MODE=unix
+#
+# By default, a Un*x style make is assumed, so towards the bottom of this
+# file filenames are copied with 'cp' and use forward slashes throughout
+# in filepaths.
+#
+# Changing to a Windows-based command line would require 'cp' replaced with
+# a local equivalent ('copy' is OK) and potentially for forward-slashes in
+# filepaths to be replaced by back-slashes.
+
+
+all: makeinclude config.h
+	echo "=== making src ==="
+	touch src/makedepend
+	cd src ; $(MAKE)
+	echo "=== making fluid ==="
+	touch fluid/makedepend
+	cd fluid ; $(MAKE)
+	echo "=== making test ==="
+	touch test/makedepend
+	cd test ; $(MAKE)
+
+install:
+	echo "=== installing src ==="
+	touch src/makedepend
+	cd src ; $(MAKE) install
+	echo "=== installing fluid ==="
+	touch fluid/makedepend
+	cd fluid ; $(MAKE) install
+
+depend:
+	echo "=== making src dependencies ==="
+	touch src/makedepend
+	cd src ; $(MAKE) depend
+	echo "=== making fluid dependencies ==="
+	touch fluid/makedepend
+	cd fluid ; $(MAKE) depend
+	echo "=== making test dependencies ==="
+	touch test/makedepend
+	cd test ; $(MAKE) depend
 
 clean:
-	-$(RM) core *.o
-	for dir in $(DIRS); do\
-		echo "=== cleaning $$dir ===";\
-		(cd $$dir; $(MAKE) $(MFLAGS) clean) || break;\
-	done
-
-distclean: clean
-	$(RM) config.*
-	$(RM) fltk-config fltk.list makeinclude
-	$(RM) FL/Makefile
-	$(RM) documentation/*.$(CAT1EXT)
-	$(RM) documentation/*.$(CAT3EXT)
-	$(RM) documentation/fltk.pdf
-	$(RM) documentation/fltk.ps
-	$(RM) -r documentation/fltk.d
-	for file in test/*.fl; do\
-		$(RM) test/`basename $file .fl`.cxx; \
-		$(RM) test/`basename $file .fl`.h; \
-	done
-
-makeinclude: configure configh.in makeinclude.in
-	if test -f config.status; then \
-		./config.status --recheck; \
-		./config.status; \
-	else \
-		./configure; \
-	fi
-	touch config.h
-
-configure: configure.in
-	autoconf
-
-portable-dist:
-	epm -v -s fltk.xpm fltk
+	-@ rm -f core config.cache *.o *.bck
+	echo "=== cleaning src ==="
+	touch src/makedepend
+	cd src ; $(MAKE) clean
+	echo "=== cleaning fluid ==="
+	touch fluid/makedepend
+	cd fluid ; $(MAKE) clean
+	echo "=== cleaning test ==="
+	touch test/makedepend
+	cd test ; $(MAKE) clean
 
-native-dist:
-	epm -v -f native fltk
+config.h: makefiles/config.mingw
+	cp $< $@
 
+makeinclude: makefiles/makeinclude.mingw
+	cp $< $@
 
 #
-# End of "$Id: Makefile,v 1.12.2.6.2.16 2003/06/15 05:39:12 easysw Exp $".
+# End of "$Id: Makefile.mingw,v 1.1.2.2 2001/01/22 15:13:39 easysw Exp $".
 #
diff -urN fltk-1.1.4/config.h fltk-1.1.4-patched/config.h
--- fltk-1.1.4/config.h	1970-01-01 02:00:00.000000000 +0200
+++ fltk-1.1.4-patched/config.h	2003-01-30 23:41:07.000000000 +0200
@@ -0,0 +1,219 @@
+/*
+ * "$Id: config.mingw,v 1.1.2.3.2.10 2003/01/30 21:41:07 easysw Exp $"
+ *
+ * Configuration file for the Fast Light Tool Kit (FLTK).
+ *
+ * Copyright 1998-2003 by Bill Spitzak and others.
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Library General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Library General Public License for more details.
+ *
+ * You should have received a copy of the GNU Library General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
+ * USA.
+ *
+ * Please report all bugs and problems to "FLT...@fl...".
+ */
+
+/*
+ * Where to find files...
+ */
+
+#define FLTK_DATADIR	"C:/FLTK"
+#define FLTK_DOCDIR	"C:/FLTK/DOC"
+
+/*
+ * BORDER_WIDTH:
+ *
+ * Thickness of FL_UP_BOX and FL_DOWN_BOX.  Current 1,2, and 3 are
+ * supported.
+ *
+ * 3 is the historic FLTK look.
+ * 2 is the default and looks like Microsoft Windows, KDE, and Qt.
+ * 1 is a plausible future evolution...
+ *
+ * Note that this may be simulated at runtime by redefining the boxtypes
+ * using Fl::set_boxtype().
+ */
+
+#define BORDER_WIDTH 2
+
+/*
+ * HAVE_GL:
+ *
+ * Do you have OpenGL? Set this to 0 if you don't have or plan to use
+ * OpenGL, and FLTK will be smaller.
+ *
+ * In order to set to 1 you will need to obtain the OpenGL header and
+ * library files.  Consult the README.win32 file for further details.
+ */
+
+#define HAVE_GL 1
+
+/*
+ * HAVE_GL_GLU_H:
+ *
+ * Do you have the OpenGL Utility Library header file?
+ * (many broken Mesa RPMs do not...)
+ */
+
+#define HAVE_GL_GLU_H 1
+
+/*
+ * USE_COLORMAP:
+ *
+ * Setting this to zero will save a good deal of code (especially for
+ * fl_draw_image), but FLTK will only work on TrueColor visuals.
+ */
+
+#define USE_COLORMAP 1
+
+/*
+ * USE_XFT
+ *
+ * Use the new Xft library to draw anti-aliased text.
+ */
+
+#define USE_XFT 0
+
+/*
+ * HAVE_XDBE:
+ *
+ * Do we have the X double-buffer extension?
+ */
+
+#define HAVE_XDBE 0
+
+/*
+ * USE_XDBE:
+ *
+ * Actually try to use the double-buffer extension?  Set this to zero
+ * disable use of XDBE without breaking the list_visuals program.
+ */
+
+#define USE_XDBE HAVE_XDBE
+
+/*
+ * HAVE_OVERLAY:
+ *
+ * Use the X overlay extension?  FLTK will try to use an overlay
+ * visual for Fl_Overlay_Window, the Gl_Window overlay, and for the
+ * menus.  Setting this to zero will remove a substantial amount of
+ * code from FLTK.  Overlays have only been tested on SGI servers!
+ */
+
+#define HAVE_OVERLAY 0
+
+/*
+ * HAVE_GL_OVERLAY:
+ *
+ * It is possible your GL has an overlay even if X does not.  If so,
+ * set this to 1.
+ */
+
+#define HAVE_GL_OVERLAY HAVE_OVERLAY
+
+/*
+ * WORDS_BIGENDIAN:
+ *
+ * Byte order of your machine: 1 = big-endian, 0 = little-endian.
+ */
+
+#define WORDS_BIGENDIAN 0
+
+/*
+ * U16, U32, U64:
+ *
+ * Types used by fl_draw_image.  One of U32 or U64 must be defined.
+ * U16 is optional but FLTK will work better with it!
+ */
+
+#define U16 unsigned short
+#define U32 unsigned
+/* #undef U64 */
+
+/*
+ * HAVE_DIRENT_H, HAVE_SYS_NDIR_H, HAVE_SYS_DIR_H, HAVE_NDIR_H, HAVE_SCANDIR:
+ *
+ * Where is <dirent.h> (used only by fl_file_chooser and scandir).
+ */
+
+#define HAVE_DIRENT_H 1
+/* #undef HAVE_SYS_NDIR_H */
+/* #undef HAVE_SYS_DIR_H */
+/* #undef HAVE_NDIR_H */
+/* #undef HAVE_SCANDIR */
+
+/*
+ * Possibly missing sprintf-style functions:
+ */
+
+/* #undef HAVE_VSNPRINTF */
+/* #undef HAVE_SNPRINTF */
+
+/*
+ * String functions...
+ */
+
+/* #undef HAVE_STRINGS_H */
+#define HAVE_STRCASECMP 1
+/* #undef HAVE_STRLCAT */
+/* #undef HAVE_STRLCPY */
+
+/*
+ * HAVE_SYS_SELECT_H:
+ *
+ * Whether or not select() call has its own header file.
+ */
+
+/* #undef HAVE_SYS_SELECT_H */
+
+/*
+ * USE_POLL:
+ *
+ * Use poll() if we don't have select().
+ */
+
+#define USE_POLL 0
+
+/*
+ * Do we have various image libraries?
+ */
+
+/* #undef HAVE_LIBPNG */
+/* #undef HAVE_LIBZ */
+/* #undef HAVE_LIBJPEG */
+
+/*
+ * Which header file do we include for libpng?
+ */
+
+/* #undef HAVE_PNG_H */
+/* #undef HAVE_LIBPNG_PNG_H */
+
+/*
+ * Do we have the png_xyz() functions?
+ */
+
+/* #undef HAVE_PNG_GET_VALID */
+/* #undef HAVE_PNG_SET_TRNS_TO_ALPHA */
+
+/*
+ * Do we have POSIX threading?
+ */
+
+/* #undef HAVE_PTHREAD */
+/* #undef HAVE_PTHREAD_H */
+
+
+/*
+ * End of "$Id: config.mingw,v 1.1.2.3.2.10 2003/01/30 21:41:07 easysw Exp $".
+ */
diff -urN fltk-1.1.4/makeinclude fltk-1.1.4-patched/makeinclude
--- fltk-1.1.4/makeinclude	1970-01-01 02:00:00.000000000 +0200
+++ fltk-1.1.4-patched/makeinclude	2004-01-30 23:50:47.000000000 +0200
@@ -0,0 +1,126 @@
+#
+# "$Id: makeinclude.mingw,v 1.1.2.3.2.11 2002/07/14 17:03:31 easysw Exp $"
+#
+# Make include file for the Fast Light Tool Kit (FLTK).
+# (this file for MingW using GCC 2.95.x)
+#
+# Copyright 1998-2001 by Bill Spitzak and others.
+#
+# This library is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Library General Public
+# License as published by the Free Software Foundation; either
+# version 2 of the License, or (at your option) any later version.
+#
+# This library is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# Library General Public License for more details.
+#
+# You should have received a copy of the GNU Library General Public
+# License along with this library; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
+# USA.
+#
+# Please report all bugs and problems to "flt...@fl...".
+#
+
+# Notes:
+#
+# 1) Console windows appearing
+#
+#    Consoles will not appear when run from a command line prompt
+#    but will cause an 'unnecessary' console to appear when run directly from
+#    windows (say explorer). It provides i/o for stdin/stdout for debugging
+#
+#    To avoid this add "-mwindows" to LDLIBS and GLDLIBS below
+#
+# 2) Stripping the executable to reduce size
+#
+#    An external program may be run to strip out any debug symbols:
+#    strip filename.exe
+# 
+#    Alternatively add -s option to LDLIBS and GLDLIBS below to strip all
+#    symbols. (or -S for stripping of debug symbols only). See 'ld' man page
+#    for other useful link-time options.
+#
+# The default is to leave -mwindows and the stripping off, so as to facilitate
+# debugging. For release it may be appropriate to add these additional switches
+
+
+prefix          = /usr/local
+exec_prefix     = ${prefix}
+bindir          = ${exec_prefix}/bin
+includedir      = ${prefix}/include
+libdir          = ${exec_prefix}/lib
+srcdir          = .
+
+# programs we use...
+CHMOD		= :
+CP		= cp
+HTMLDOC		= htmldoc.exe
+LN		= :
+MKDIR		= mkdir
+NROFF		= nroff
+RM		= rm
+RMDIR		= rmdir
+STRIP		= :
+
+# compiler names:
+CXX		= i686-pc-cygwin-g++
+CC		= i686-pc-cygwin-gcc
+MAKEDEPEND	= $(CXX) -M
+POSTBUILD	= :
+
+# flags for C++ compiler:
+CFLAGS		= -Wall -O2 -DWIN32
+CXXFLAGS	= -Wall -O2 -DWIN32
+
+# program to make the archive:
+LIBNAME		= ../lib/libfltk.a
+FLLIBNAME	= ../lib/libfltk_forms.a
+GLLIBNAME	= ../lib/libfltk_gl.a
+IMGLIBNAME	= ../lib/libfltk_images.a
+LIBCOMMAND	= i686-pc-cygwin-ar -ruv
+RANLIB		= i686-pc-cygwin-ranlib
+DSONAME		=
+FLDSONAME	=
+GLDSONAME	=
+IMGDSONAME	=
+DSOCOMMAND	= echo
+
+# libraries to link with (in addition to default libs):
+LDLIBS		=-mwindows -lgdi32 -lcomctl32 -lwsock32 -luuid -lole32
+GLDLIBS		=-mwindows -lglu32 -lopengl32 -lgdi32 -lcomctl32 -lwsock32 -luuid -lole32
+LINKFLTK	=-L../lib -lfltk
+LINKFLTKGL	=-L../lib -lfltk_gl
+LINKFLTKFORMS	=-L../lib -lfltk_forms -lfltk
+LINKFLTKIMG	=-L../lib -lfltk_images -lfltk
+IMAGELIBS	=
+
+# Do we build the OpenGL demos?
+GLDEMOS		=gldemos
+
+# Do we build the threads demo?
+THREADS		=threads
+
+# Be quiet when building...
+.SILENT:
+
+# Build commands and filename extensions...
+.SUFFIXES:	.c .cxx .h .fl .o
+
+.cxx:
+	echo Compiling and linking $@...
+	$(CXX) -I.. $(CXXFLAGS) $< $(LINKFLTK) $(LDLIBS) -o $@
+
+.c.o:
+	echo Compiling $@...
+	$(CC) -I.. $(CXXFLAGS) $< -c
+
+.cxx.o:
+	echo Compiling $@...
+	$(CXX) -I.. $(CXXFLAGS) $< -c
+
+#
+# End of "$Id: makeinclude.mingw,v 1.1.2.3.2.11 2002/07/14 17:03:31 easysw Exp $".
+#
diff -urN fltk-1.1.4/src/Fl_win32.cxx fltk-1.1.4-patched/src/Fl_win32.cxx
--- fltk-1.1.4/src/Fl_win32.cxx	2003-06-15 07:27:35.000000000 +0300
+++ fltk-1.1.4-patched/src/Fl_win32.cxx	2004-01-30 23:40:23.000000000 +0200
@@ -47,7 +47,6 @@
 // The following include files require GCC 3.x or a non-GNU compiler...
 #if !defined(__GNUC__) || __GNUC__ >= 3
 #  include <ole2.h>
-#  include <ShellApi.h>
 #endif // !__GNUC__ || __GNUC__ >= 3
 
 
diff -urN fltk-1.1.4/src/filename_list.cxx fltk-1.1.4-patched/src/filename_list.cxx
--- fltk-1.1.4/src/filename_list.cxx	2003-01-30 23:43:21.000000000 +0200
+++ fltk-1.1.4-patched/src/filename_list.cxx	2004-01-30 23:40:59.000000000 +0200
@@ -51,7 +51,7 @@
                      Fl_File_Sort_F *sort) {
 #if defined(__hpux) || defined(__CYGWIN__)
   // HP-UX, Cygwin define the comparison function like this:
-  return scandir(d, list, 0, (int(*)(const dirent **, const dirent **))sort);
+  return scandir(d, list, 0, (int(*)(dirent **, dirent **))sort);
 #elif defined(__osf__)
   // OSF, DU 4.0x
   return scandir(d, list, 0, (int(*)(dirent **, dirent **))sort);


-- 
Dan Aloni
da...@gm...

[coLinux-devel] cross probs while building fltk

[Thomas F. <tf...@no...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi colinux-hackers,

I build cygwin tools as described in cygwin-cross-build without problems.
I also compiled mxlm with this cygwin-tools, but now I have some problems
with building fltk. This is because fltk has a lot of depends on other
libs for example jpeglib or stdc++ and it looks like the system would like
to build a version for X11.

I can disable a few libs in the configuration but I could still not build
fltk.

What parameters do you use to cross build fltk?

Thanks and regards,
 Thomas



-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAJO27nwJBIFTVIqwRApx4AJ9HoVPoK8fSoPtQbTtdL957/fJ2ggCfTFR9
jFij8C8h1jxsjFZT+1FQS8E=
=TeTg
-----END PGP SIGNATURE-----

Re: [coLinux-devel] colinux/os/winnt/user/poll.c

[Dan A. <da...@gm...>]

On Fri, Feb 06, 2004 at 10:18:41PM -0800, Ballard, Jonathan wrote:
> Hello,
> 
> A few suggestions that help portability:

That code is going away in the next release :) 
 
> int co_os_poll_chain_wait(co_os_poll_chain_t chain)
> {
> //--    HANDLE wevents[chain->count];
> ++	HANDLE *wevents ;
> 	DWORD result;
> 	co_os_poll_t poll;
> 	int i;
> 
> ++	wevents = co_os_malloc( sizeof(HANDLE) * chain->count ) ;
> 
> 	i=0;
> 	co_list_each_entry(poll, &chain->list, node) {
> 		wevents[i] = poll->event;
> 		i++;
> 		}
> 
> 	result = MsgWaitForMultipleObjects(chain->count, (void * const 
> 	*)&wevents,
> 			FALSE, INFINITE, QS_ALLINPUT);
> 
> ++	co_os_free( wevents ) ;
> 
> 	if (result == WAIT_OBJECT_0 + chain->count)
> 		return 0;
> .
> 
> .
> 
> .
> 
> 
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel

-- 
Dan Aloni
da...@gm...

Re: [coLinux-devel] 0.5.2 status: uptime 4 days, yet corrupted files

[John L. <jo...@su...>]

Ballard Jonathan wrote:

> I installed ftpd and transfered the linux-2.4.24 sources to coLinux 
> via ftp.  After I unpacked the files, I installed them in the /usr/src 
> directory, and I ran the make {mrproper,oldconfig,dep}.  The "make 
> dep" errored on the mm.h file.    I looked at the file and compared 
> with the sources on my other system.  The two files were different: 
> obvious corruption.  Since I could not run fsck, I do not know if it 
> the corruption was from the filesystem or the network.
>  
>  

Daniel Glockner already found the source of this bug(in the network), 
and Dan has said that it will be fixed in the next release.

John

Re: [coLinux-devel] list.h

[Dan A. <da...@gm...>]

On Fri, Feb 06, 2004 at 07:03:25PM -0800, Ballard Jonathan wrote:
> In list.h line 65:
> 
> #define co_list_entry_assign(ptr, var, member) \
> 
>     var = co_list_entry(ptr, typeof(*var), member)
> 
> 
> 
> The typeof() is not standard C.  I presumed it works fine with GNU C.  My compiler complained.  I can workaround it.  Perhaps, somebody knows of an easy alternative?

What's your compiler version?

-- 
Dan Aloni
da...@gm...

Re: [coLinux-devel] colinux/os/winnt/user/poll.c

[Dan A. <da...@gm...>]

On Sat, Feb 07, 2004 at 09:02:38AM +0200, Nir Perry wrote:
> Hi y'all,
> 
> Just a small comment - not really coLinux related, but it is a general 
> fault-tolerance issue:
> 
> When using function calls that can fail (assuming co_os_malloc(), like 
> malloc(), can run out of memory),
> one should always handle the failures.
> Of course, just returning an error here would not be enough - the higher 
> levels must handle it too.
> 
> Before you shoot me, yes - I know such code appears in SO MANY places, 
> and in so many projects, but that doesn't mean we want more...
> 
> Some people tend to say - "So what? if malloc() fails, I don't care, 
> because nothing can work after it. Just buy more memory". But 
> fault-tolerance is important to software. Especially to OS kernels.

Yes, I agree. The general solution is to allow the program (or module / 
branch of execution) to exit *cleanly* in an out-of-memory situations, 
keeping the problem as local as possible.

-- 
Dan Aloni
da...@gm...

Re: [coLinux-devel] colinux/os/winnt/user/poll.c

[Nir P. <ni...@em...>]

Hi y'all,

Just a small comment - not really coLinux related, but it is a general 
fault-tolerance issue:

When using function calls that can fail (assuming co_os_malloc(), like 
malloc(), can run out of memory),
one should always handle the failures.
Of course, just returning an error here would not be enough - the higher 
levels must handle it too.

Before you shoot me, yes - I know such code appears in SO MANY places, 
and in so many projects, but that doesn't mean we want more...

Some people tend to say - "So what? if malloc() fails, I don't care, 
because nothing can work after it. Just buy more memory". But 
fault-tolerance is important to software. Especially to OS kernels.

Nir


Ballard, Jonathan wrote:

> Hello,
>
> A few suggestions that help portability:
>
>
> int co_os_poll_chain_wait(co_os_poll_chain_t chain)
> {
> //--    HANDLE wevents[chain->count];
> ++    HANDLE *wevents ;
>     DWORD result;
>     co_os_poll_t poll;
>     int i;
>
> ++    wevents = co_os_malloc( sizeof(HANDLE) * chain->count ) ;
>
>     i=0;
>     co_list_each_entry(poll, &chain->list, node) {
>         wevents[i] = poll->event;
>         i++;
>         }
>
>     result = MsgWaitForMultipleObjects(chain->count, (void * const 
> *)&wevents,
>             FALSE, INFINITE, QS_ALLINPUT);
>
> ++    co_os_free( wevents ) ;
>
>     if (result == WAIT_OBJECT_0 + chain->count)
>         return 0;
> .
>
> .
>
> .
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel

[coLinux-devel] colinux/os/winnt/user/poll.c

[Ballard, J. <sac...@ho...>]

Hello,

A few suggestions that help portability:


int co_os_poll_chain_wait(co_os_poll_chain_t chain)
{
//--    HANDLE wevents[chain->count];
++	HANDLE *wevents ;
	DWORD result;
	co_os_poll_t poll;
	int i;

++	wevents = co_os_malloc( sizeof(HANDLE) * chain->count ) ;

	i=0;
	co_list_each_entry(poll, &chain->list, node) {
		wevents[i] = poll->event;
		i++;
		}

	result = MsgWaitForMultipleObjects(chain->count, (void * const *)&wevents,
			FALSE, INFINITE, QS_ALLINPUT);

++	co_os_free( wevents ) ;

	if (result == WAIT_OBJECT_0 + chain->count)
		return 0;
.

.

.

[coLinux-devel] 0.5.2 status: uptime 4 days, yet corrupted files

[Ballard J. <sac...@ho...>]

I installed ftpd and transfered the linux-2.4.24 sources to coLinux via =
ftp.  After I unpacked the files, I installed them in the /usr/src =
directory, and I ran the make {mrproper,oldconfig,dep}.  The "make dep" =
errored on the mm.h file.    I looked at the file and compared with the =
sources on my other system.  The two files were different: obvious =
corruption.  Since I could not run fsck, I do not know if it the =
corruption was from the filesystem or the network.

Re: [coLinux-devel] Memory size and killing console by Esc

[Dan A. <da...@gm...>]

On Sat, Feb 07, 2004 at 12:52:31AM +0200, Paul Sokolovsky wrote:

> 1. Could memory size for coLinux be configurable? Current 24Mb or so
> is a joke. Yes, swap can be mounted, but why experience perfomance of
> vmware level? ;-) It would be nice to have memory size configurable in
> XML config, but if it's some time to do, could next release ship with
> at least slightly decent amount, like 64Mb (128Mb better ;-).

The major problem with what you are suggesting, is the limitations 
enforced by Windows' non cached memory allocator. This pool is 
hardcoded to a limit of 256MB and even 128MB on some systems. Allocating 
too much memory from that pool can destabilze Windows.
 
> 2. Could Esc key stop killing console? I currently work that around by
> Alt+27 on numeric keyboard, but it's just pain in the eye.

An FLTK fixable issue, I'll take care of that.
 
>   Ah, and yes - default image provided is pretty bare. It would be
> nice to have at least ftpd and telnetd, and less included - it's not
> always suitable to do apt-get (for example, for people with networking
> problems). I guess, that's request for some kind sole who can
> contribute such an image, coLinux author has other stuff to do ;-)

We will distribue further images in the future.

-- 
Dan Aloni
da...@gm...

[coLinux-devel] list.h

[Ballard J. <sac...@ho...>]

In list.h line 65:

#define co_list_entry_assign(ptr, var, member) \

    var =3D co_list_entry(ptr, typeof(*var), member)



The typeof() is not standard C.  I presumed it works fine with GNU C.  =
My compiler complained.  I can workaround it.  Perhaps, somebody knows =
of an easy alternative?

Re: [coLinux-devel] Memory size and killing console by Esc

[Richard G. <ric...@ri...>]

I don't know about 1 or 2, but I highly second the request for a different
image. I'd actually like to use Arch Linux; would being 686 targeted be a
problem? The package management sounds much like Debian and it sounds like
you could start with a fairly small image as well.

I tried to make one at work earlier but something happened with the
networking and I couldn't dd off the image :-)

Richard

----- Original Message ----- 
From: "Paul Sokolovsky" <coo...@ua...>
To: <col...@li...>
Sent: Friday, February 06, 2004 4:52 PM
Subject: [coLinux-devel] Memory size and killing console by Esc


> Hello colinux-devel,
>
>   I would like to report success running coLinux on
> P4/2.66GHz(no-hyperthreading)/1Gb/SiS chipset (forget which one ;-).
> I had that image size problem but apart from that everything seems to
> run smooth.
>
>   There're two feature requests I'd like to sound which somehow were
> touched on the list, but doesn't seem to draw much attention:
>
> 1. Could memory size for coLinux be configurable? Current 24Mb or so
> is a joke. Yes, swap can be mounted, but why experience perfomance of
> vmware level? ;-) It would be nice to have memory size configurable in
> XML config, but if it's some time to do, could next release ship with
> at least slightly decent amount, like 64Mb (128Mb better ;-).
>
> 2. Could Esc key stop killing console? I currently work that around by
> Alt+27 on numeric keyboard, but it's just pain in the eye.
>
>   Ah, and yes - default image provided is pretty bare. It would be
> nice to have at least ftpd and telnetd, and less included - it's not
> always suitable to do apt-get (for example, for people with networking
> problems). I guess, that's request for some kind sole who can
> contribute such an image, coLinux author has other stuff to do ;-)
>
>
> Thanks and have a great weekend!
>
> -- 
> Best regards,
>  Paul                          mailto:coo...@ua...
>
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel
>
>

[coLinux-devel] Memory size and killing console by Esc

[Paul S. <coo...@ua...>]

Hello colinux-devel,

  I would like to report success running coLinux on
P4/2.66GHz(no-hyperthreading)/1Gb/SiS chipset (forget which one ;-).
I had that image size problem but apart from that everything seems to
run smooth.

  There're two feature requests I'd like to sound which somehow were
touched on the list, but doesn't seem to draw much attention:

1. Could memory size for coLinux be configurable? Current 24Mb or so
is a joke. Yes, swap can be mounted, but why experience perfomance of
vmware level? ;-) It would be nice to have memory size configurable in
XML config, but if it's some time to do, could next release ship with
at least slightly decent amount, like 64Mb (128Mb better ;-).

2. Could Esc key stop killing console? I currently work that around by
Alt+27 on numeric keyboard, but it's just pain in the eye.

  Ah, and yes - default image provided is pretty bare. It would be
nice to have at least ftpd and telnetd, and less included - it's not
always suitable to do apt-get (for example, for people with networking
problems). I guess, that's request for some kind sole who can
contribute such an image, coLinux author has other stuff to do ;-)


Thanks and have a great weekend!

-- 
Best regards,
 Paul                          mailto:coo...@ua...

Re: [coLinux-devel] ext3 filesystem

[Dan A. <da...@gm...>]

On Fri, Feb 06, 2004 at 08:03:28PM +0100, Daniel Gl?ckner wrote:
> The source of this bug is a typo in fileblock.c.
> Where it says
> 
> 		if (CO_RC(OK))
> 
> it should be
> 
> 		if (CO_OK(rc))

Fix applied, thanks.

-- 
Dan Aloni
da...@gm...