Support for 5D3 firmwares 1.2.3 and 1.1.3 in the same source tree

#611 Merged at 1d8c97a
Repository
Branch
5D3-113-and-123
Repository
Branch
unified
Author
  1. Alex
Reviewers
Description

Looks like we should support both firmwares: 1.1.3 for general purpose and 1.2.3 for certain users (those with external monitors and those with f/8 lenses). Rationale: http://www.magiclantern.fm/forum/index.php?topic=14704

I would postpone the upgrade to 1.3.3 a bit. It will replace 1.2.3, but I think 1.1.3 is here to stay (so we'll have to support 1.1.3 and 1.3.3).

This PR attempts to support both firmware versions in the same source tree.

Comments (8)

  1. chris_miller

    For obvious reasons, having support for both firmwares in the same source tree is fantastic. Thanks a1ex! I've just tried building this branch for 1.2.3 and it builds fine and (superficially at least) runs on my camera.

    I agree that 1.1.3 needs to continue to be supported, also agree that 1.2.3 can be discarded in favour of 1.3.3 in due course.

    [Note that I'm not on 1.2.3 for external monitor support and I only use f/8 TCs rarely; the primary reason I moved to 1.2.3 is because I had been bitten by the 1.1.3 AFMA bug a few times]

    1. Alex author

      1.1.3 AFMA bug? Sounds like another good reason to use the newer version.

      Have a link with a good description for this bug?

      1. chris_miller

        It causes your AFMA values to get lost occasionally. The really annoying thing is there's no obvious way to notice it has happened, other than regularly check the Canon menus (or of course if you realise that all your recent photos are somewhat front or back focused!). I don't know if I was affected worse than most people or I have particular lens/settings combinations that trigger it on my camera but the bug was causing me enough frustration that I don't want to use 1.1.3 ever again.

        The bug is mentioned here as point number 2:

        http://cpn.canon-europe.com/content/news/new_firmware_for_eos_5d_mark_iii.do

        IIRC Canon had tried to fix it in 1.2.1 but a further fix was needed in 1.2.3.

        1. Jarno Paananen

          I got this bug a few times as well, though I think I got it only after trying 1.2.1 and 1.2.3 and going back to 1.1.3 before 1.2.3 port was available. Either the AFMA value got reset to zero or set to some arbitrary value. I don't use f/8 that much either (yet, waiting for 100-400 II), but this makes me want to stay away from 1.1.3.

  2. Daniel Fort

    I've been merging this branch to make the raw_video_10bit_12bit test builds and it is working great. It would be wonderful to have 5D3.123 in the unified branch.