Add Exif Gamma tag

[kmilos]

No description provided.

[clanmills]

Several comments:

1. Do you have a test file and python script?
2. Are there other missing Exif Tags from Table 7 of https://fotomagazin.hu/wp-content/uploads/2020/05/CIPA_DC_008_EXIF_2019.pdf
3. What do you think about adding the tags in Support for DNG 1.5 #1397 to this PR?
4. Is printFloat() the correct prettyPrinter? Do you know how ExifTool presents this data.? I think it's between 0.0 and 1.0 If in doubt ask @alexvanderberkel because he knows about Photography.

[clanmills]

I see this has died a horrible (and well deserved) death in the test suite. You'll have to update that. For example:

/Users/travis/build/Exiv2/exiv2/test/data/exiv2-test.out: 8756 lines
863/Users/travis/build/Exiv2/exiv2/test/tmp/exiv2-test.out: 8756 lines
864The first mismatch is in line 979:
865< 20040329_224245.jpg   Exif.Photo.0xa500                            Rational    1  22/10
866> 20040329_224245.jpg   Exif.Photo.Gamma                             Rational    1  2.2

It's easy to fix. You'll have to edit test/data/exiv2-test.out and replace >0xa500< with >Gamma <. Watch out for the space following Gamma.

[clanmills]

When the test suite has been fixed, we don't need to bother adding anything to the python tests (in tests/bugfixes/github). It's already tested by exiv2-test (which is in tests/bash_tests). It's is called 'bash_tests' for historical reasons. Leo rewrote them in python and that make them much easier to run on Visual Studio builds.

[kmilos]

Sorry, I'll fix up the test data, I wrongly assumed there was nothing there...

For the other questions:

2. Are there other missing Exif Tags from Table 7

Nope, this was the only one. I checked tables 7, 8, and 15.

3. What do you think about adding the tags in Support for DNG 1.5 #1397 to this PR?

We have all the DNG tags in up to 1.6 already in 0.27.4-RC2. What's missing in #1397 is "only" support for separate DCP (DNG Camera Profile) IFDs and standalone ".dcp" TIFF-like files... I think you tested these in your tvisitor and covered it in your book, but we don't support them in exiv2 yet...

4. Is printFloat() the correct prettyPrinter? Do you know how ExifTool presents this data.?

It's not suitable for all rationals for sure. The main question for the ones where it does make sense is whether we want to print units as well or not... I'll check what ExifTool does as well for some of these.

Let's keep that discussion to the other issue report, and handle this PR for Gamma first, and maybe back-port this only missing tag to 0.27.4?

[clanmills]

Thanks. I'm happy to see this merged. The existing way (prior to the change) is:

632 rmills@rmillsm1:~/gnu/github/exiv2/0.27-maintenance $ exiv2 -g a500 --unknown test/data/exiv2-nikon-d70.jpg
Exif.Photo.0xa500                            Rational    1  22/10
633 rmills@rmillsm1:~/gnu/github/exiv2/0.27-maintenance $ 

It's so much nicer with your PR:

641 rmills@rmillsm1:~/gnu/github/exiv2/main/build $ exiv2 -g gamma/i ../test/data/exiv2-nikon-d70.jpg
Exif.Photo.Gamma                             Rational    1  2.2
642 rmills@rmillsm1:~/gnu/github/exiv2/main/build $ 

And that "rhymes" well with ExifTool:

642 rmills@rmillsm1:~/gnu/github/exiv2/main/build $ exiftool ../test/data/exiv2-nikon-d70.jpg | grep -i gamma
Gamma                           : 2.2
643 rmills@rmillsm1:~/gnu/github/exiv2/main/build $ 

I'm not going to back-port it to 0.27-maintenance. After RCfinal, I only want to change the revision number and update documentation. I avoid C++ changes. It's the same reason that I don't want to add Exiv2::enableBMFF() without EXV_ENABLE_BMFF.

If we must have an RC3 (for a security fix or darktable reports a bmff issue), I can back-port this PR to 0.27-maintenance.

[clanmills]

Good Job, @kmilos. Thanks for dealing with this.

[kmilos]

Sure, that release policy makes sense.

Thanks.