Module code V2 - test branch. - page 2 - General Discussion and Assistance

waterwingz

12139

Re: Module code V2 - test branch.

« Reply #10 on: 07 / April / 2013, 00:02:47 »

Advertisements

Quote from: dvip on 06 / April / 2013, 23:37:46

I couldn't find the menu item to the CHDK settings menu to allow shortcut keys to be enabled/disabled... I guess this is not there yet.

Hmmm .. turning out to be a popular option. Its there in the CHDK Settings menu - half way down where it says Enable Half-Press Shortcut Keys. But only for builds from the dev / unstable / 1.2.0 / Gen3 trunk.

Logged

Ported : A1200 SD940 G10 Powershot N G16

dvip

451

Re: Module code V2 - test branch.

« Reply #11 on: 07 / April / 2013, 00:18:03 »

@waterwingz, I don't think Phil has added this yet
to philmoz-modulesV2@2680. But it is in trunk@2680.
Yep, this is nice to have.

Logged

lapser

1093

Re: Module code V2 - test branch.

« Reply #12 on: 07 / April / 2013, 00:33:13 »

Quote from: dvip on 07 / April / 2013, 00:18:03

@waterwingz, I don't think Phil has added this yet
to philmoz-modulesV2@2680. But it is in trunk@2680.
Yep, this is nice to have.

I just compiled 2679 and the new menu item is there. I think Phil already merged philmoz-modulesV2 into the trunk now.

I was able to get my time lapse mods patched into trunk 2679 without too much trouble. Does anyone know how to tell what Lua version is installed now? I'd like to confirm that it's been updated so I can see if the Lua yield bug I've been bypassing has been fixed.

Logged

EOS-M3_120f / SX50_100b / SX260_101a / G1X_100g / D20_100b
https://www.youtube.com/user/DrLapser/videos

reyalp

14080

Re: Module code V2 - test branch.

« Reply #13 on: 07 / April / 2013, 00:40:50 »

In trunk lib/lua/lua.h

Code: [Select]

#define LUA_RELEASE	"Lua 5.1.5"

Logged

Don't forget what the H stands for.

lapser

1093

Re: Module code V2 - test branch.

« Reply #14 on: 07 / April / 2013, 14:16:17 »

Quote from: reyalp on 07 / April / 2013, 00:40:50

Code: [Select]
#define LUA_RELEASE "Lua 5.1.5"

OK, thanks. I just took 13,000 pictures with my time lapse script using wait_click(10), (which does yield resume), without any problem. I'm not sure the yield bug has been fixed yet, but this is a good sign.

I wonder if I can sell 13,000 pictures of fog on ebay?

I'll try a foggy sunset tonight with all cameras. If I have the energy, I'll go up the butte again so at least I'll have pictures of a foggy city instead of just fog.

Logged

EOS-M3_120f / SX50_100b / SX260_101a / G1X_100g / D20_100b
https://www.youtube.com/user/DrLapser/videos

rudi

129
A590IS_101B, SX260HS_100B

Re: Module code V2 - test branch.

« Reply #15 on: 08 / April / 2013, 15:57:37 »

I need your support again.
64-bit linux will not compile Module code V2. In this case for german autobuild server. The last complete compiled revision is 2674. Current Revision is 2685.
Here the reduced compiler output with errors:

Code: [Select]

elfflt.c:725: warning: cast from pointer to integer of different size
fltdump.c:156: warning: cast from pointer to integer of different size

In file fselect.elf:
Wrong _module_info->sizeof_struct value. Please check correct filling of this structure
make[1]: *** [fselect.flt] Error 7
make: *** [allmodules] Error 2

In /modules/flt.h I found this:

Code: [Select]

#if __SIZEOF_POINTER__ == 8 // For elfflt.c on 64 bit Linux
    uint32_t            lib;
#else
    base_interface_t*   lib;                    // Pointer to interface library
#endif

Thanks a lot.

full_compiler_output_rev2685.txt (4.84 kB - downloaded 50 times.)

Logged

philmoz

3450

Re: Module code V2 - test branch.

« Reply #16 on: 08 / April / 2013, 18:19:01 »

Quote from: rudi on 08 / April / 2013, 15:57:37

I need your support again.
64-bit linux will not compile Module code V2. In this case for german autobuild server. The last complete compiled revision is 2674. Current Revision is 2685.
Here the reduced compiler output with errors:
Code: [Select]
elfflt.c:725: warning: cast from pointer to integer of different size fltdump.c:156: warning: cast from pointer to integer of different size In file fselect.elf: Wrong _module_info->sizeof_struct value. Please check correct filling of this structure make[1]: *** [fselect.flt] Error 7 make: *** [allmodules] Error 2In /modules/flt.h I found this:
Code: [Select]
#if __SIZEOF_POINTER__ == 8 // For elfflt.c on 64 bit Linux uint32_t lib; #else base_interface_t* lib; // Pointer to interface library #endif
Thanks a lot.

Sounds like your compiler is not defining __SIZEOF_POINTER__ correctly.
If you are using GCC this should be set.

The structures in flt.h were changed to have pointer variables to make the CHDK code clearer.
The tools are built as native code on the build machine so if you use 64 bit Linux these pointer variables are the wrong size - the #if is supposed to fix this by substituting a 32 bit int in place of the pointers for the tools.

Phil.

Logged

CHDK ports:
sx30is (1.00c, 1.00h, 1.00l, 1.00n & 1.00p)
g12 (1.00c, 1.00e, 1.00f & 1.00g)
sx130is (1.01d & 1.01f)
ixus310hs (1.00a & 1.01a)
sx40hs (1.00d, 1.00g & 1.00i)
g1x (1.00e, 1.00f & 1.00g)
g5x (1.00c, 1.01a, 1.01b)
g7x2 (1.01a, 1.01b, 1.10b)

waterwingz

12139

Re: Module code V2 - test branch.

« Reply #17 on: 08 / April / 2013, 19:09:09 »

Quote from: philmoz on 08 / April / 2013, 18:19:01

The tools are built as native code on the build machine so if you use 64 bit Linux these pointer variables are the wrong size - the #if is supposed to fix this by substituting a 32 bit int in place of the pointers for the tools.

I'm running Fedora 64 bit and everything continues to build correctly using the zipped up autobuild environment I got from hacki a year or so ago. Is it possible its actually a 32bit environment though? The various options like ./arm-elf-gcc-4.4.3 -dumpspecs don't tell me much useful.

« Last Edit: 08 / April / 2013, 19:14:55 by waterwingz »

Logged

Ported : A1200 SD940 G10 Powershot N G16

philmoz

3450

Re: Module code V2 - test branch.

« Reply #18 on: 08 / April / 2013, 19:17:44 »

Quote from: waterwingz on 08 / April / 2013, 19:09:09

Quote from: philmoz on 08 / April / 2013, 18:19:01
The tools are built as native code on the build machine so if you use 64 bit Linux these pointer variables are the wrong size - the #if is supposed to fix this by substituting a 32 bit int in place of the pointers for the tools.
I'm running Fedora 64 bit and everything continues to build correctly using the zipped up autobuild environment I got from hacki a year or so ago. Is it possible its actually a 32bit environment though? The various options like ./arm-elf-gcc-4.4.3 -dumpspecs don't tell me much useful.

Hacki's environment is 64 bit. I use it for testing batch builds which is how I discovered the problem when I first made the changes.

Phil.

Logged

CHDK ports:
sx30is (1.00c, 1.00h, 1.00l, 1.00n & 1.00p)
g12 (1.00c, 1.00e, 1.00f & 1.00g)
sx130is (1.01d & 1.01f)
ixus310hs (1.00a & 1.01a)
sx40hs (1.00d, 1.00g & 1.00i)
g1x (1.00e, 1.00f & 1.00g)
g5x (1.00c, 1.01a, 1.01b)
g7x2 (1.01a, 1.01b, 1.10b)

rudi

129
A590IS_101B, SX260HS_100B

Re: Module code V2 - test branch.

« Reply #19 on: 12 / April / 2013, 03:40:32 »

Quote from: philmoz on 08 / April / 2013, 18:19:01

Sounds like your compiler is not defining __SIZEOF_POINTER__ correctly.

Thank you, Phil. There was the error.

The shared server use gcc, but there are missing many precompiler definitions.
My first way was a patch for __SIZEOF_POINTER__

Code: [Select]

#if __SIZEOF_POINTER__ == 4
    base_interface_t*   lib;                    // Pointer to interface library
#else
    uint32_t            lib;                    // all other __SIZEOF_POINTER__ sizes
#endif

Now I use my own x64_64-gcc crosscompiler since rev. 2694. It should be good.

Logged