fonts/display problem, anyone else ?

[indy arm]

Hi Trammel,

here is the problem I have with display.
Attached is a picture of my screen, and the generated autoexec.bin

I used :
BINUTILS=binutils-2.19.1
GCC=gcc-4.4.2
NEWLIB=newlib-1.17.0
GDB=gdb-7.1

with
http://github.com/esden/summon-arm-toolchain
http://hermann-uwe.de/blog/building-an-arm-cross-toolchain-with-binutils-gcc-newlib-and-gdb-from-source?page=1

Am I alone with this problem ?

compilation details are explained here:
http://groups.google.com/group/ml-devel/browse_thread/thread/ba6a0d550c80f33c

thanks for any help.

Arm.Indy

[Trammell Hudson]

On Mon, Sep 13, 2010 at 08:43:47PM +0200, indy arm wrote:
> here is the problem I have with display.
> Attached is a picture of my screen, and the generated autoexec.bin

Does the problem go away if you remove the info display? There is a
long standing bug that the ML overlay and the Canon overlay will
flicker if both are onscreen at the same time.

--
Trammell

[indy arm]

Hi,

Here is my screen with original autoexec.bin from Trammel = OK.
this is no change with my autoexec when removing the info display.

I noticed that this is displayed (FONT_SMALL) :

bmp_printf( FONT_SMALL, 500, 400, "Gain %d/%d Mgain %d",  dgain_l,  dgain_r,  mgain );


and all display with FONT_MED are not displayed, like :

bmp_printf( FONT_MED, 0, 40,
		"Magic Lantern version %s (%s)\n"
		"Built on %s by %s\n",
		build_version,
		build_id,
		build_date,
		build_user
	);

Arm.Indy

2010/9/13 indy arm <arm.i...@gmail.com>

[Trammell Hudson]

On Tue, Sep 14, 2010 at 09:03:01PM +0200, indy arm wrote:
> [...] all display with FONT_MED are not displayed

What is in font-med.c? Was it correctly generated from font-med.in?
It should have 1571 lines that look like:

#include "font.h"
struct font font_med = {
.width = 12,
.height = 16,
.bitmap = {
[ 'A' + ( 0 << 7) ] = 0x00000000, //
[ 'A' + ( 1 << 7) ] = 0x0e000000, // ###
[ 'A' + ( 2 << 7) ] = 0x0e000000, // ###
[ 'A' + ( 3 << 7) ] = 0x1f000000, // #####
[ 'A' + ( 4 << 7) ] = 0x1b000000, // ## ##
[ 'A' + ( 5 << 7) ] = 0x3b800000, // ### ###
[ 'A' + ( 6 << 7) ] = 0x31800000, // ## ##
[ 'A' + ( 7 << 7) ] = 0x71c00000, // ### ###
[ 'A' + ( 8 << 7) ] = 0x60c00000, // ## ##
[ 'A' + ( 9 << 7) ] = 0x7fc00000, // #########
[ 'A' + (10 << 7) ] = 0xffe00000, // ###########
[ 'A' + (11 << 7) ] = 0xc0600000, // ## ##
[ 'A' + (12 << 7) ] = 0xc0600000, // ## ##
[ 'A' + (13 << 7) ] = 0x00000000, //
[ 'A' + (14 << 7) ] = 0x00000000, //
[ 'A' + (15 << 7) ] = 0x00000000, //

// ....

--
Trammell

[indy arm]

Hi Trammel,

I took time to understand the font build process (thanks for the clues), and I think I have found the problem: all font-*.in are empty but font-small.in, so associated *.c and .o are empty.

Could you please check in the right version of font-*.in in 550D branch ?
The 5dm2 branch (0.1.6, maybe earlier) have correct large, med and small *.in version, but not huge one.

Could you apply the Makefile patch I have proposed in this thread ?
http://groups.google.com/group/ml-devel/browse_thread/thread/ba6a0d550c80f33c

thanks for your help.

Arm.Indy

2010/9/13 indy arm <arm.i...@gmail.com>

Hi Trammel,

[indy arm]

Hi,

I found the problem:
1 - the font*.in rules do not work : font*.in are always regenerated EVEN if target files do already exist. Don't ask me why
2 - when bigtext program is not present (in most cases), the generatefont Perl script generate an empty version of font-*.in that overwrite the good version (that ARE on bitbucket, I was wrong here)

So I deleted the font-*.in rules and retrieved again the good font*.in files from the HG depot.

So I confirm being able to compile autoexec.bin from sources on Ubuntu 10.04, with
ARM toolchain from:
http://www.hermann-uwe.de/blog/building-an-arm-cross-toolchain-with-binutils-gcc-newlib-and-gdb-from-source

configured with:
BINUTILS=binutils-2.19.1
GCC=gcc-4.4.2
NEWLIB=newlib-1.17.0
GDB=gdb-7.1

this other toolchain seems a good candidate too (not tested), for Linux and Windows
http://www.codesourcery.com/sgpp/lite/arm/portal/release1293

Please find the modified Makefile I used.

Arm.Indy

[Alex]

Thanks for the fix. It might be a problem with the date/time of the
virtual machine onto which you tested the compilation.

On my Ubuntu Karmic box, the font*.in rules work well with the
original Makefile. I didn't run yet ML on the camera due to a problem
with the card reader, but I'll try it soon.

A nice fix would be to test whether bigtext is present; if not, and
leave font*.c unchanged.

On Sep 18, 12:44 am, indy arm <arm.indi...@gmail.com> wrote:
> Hi,
>
> I found the problem:
> 1 - the font*.in rules do not work : font*.in are always regenerated EVEN if
> target files do already exist. Don't ask me why
> 2 - when bigtext program is not present (in most cases), the generatefont
> Perl script generate an empty version of font-*.in that overwrite the good
> version (that ARE on bitbucket, I was wrong here)
>
> So I deleted the font-*.in rules and retrieved again the good font*.in files
> from the HG depot.
>
> So I confirm being able to compile autoexec.bin from sources on Ubuntu
> 10.04, with

> ARM toolchain from:http://www.hermann-uwe.de/blog/building-an-arm-cross-toolchain-with-b...

>
> configured with:
> BINUTILS=binutils-2.19.1
> GCC=gcc-4.4.2
> NEWLIB=newlib-1.17.0
> GDB=gdb-7.1
>
> this other toolchain seems a good candidate too (not tested), for Linux and

> Windowshttp://www.codesourcery.com/sgpp/lite/arm/portal/release1293

>
> Please find the modified Makefile I used.
>
> Arm.Indy
>

>  Makefile
> 11KViewDownload
>
>  IMAG0250.jpg
> 1146KViewDownload

[arm.indy]

hi Alex,

thanks for the feedback.
yes, the problem must be specific to my configuration, because nobody
else noticed this.

Indy