Test Image.histogram() for all modes

[Yay295]

And add a hash of the histogram to the values that are checked.

[Yay295]

Current Errors:
amazon-2-amd64, mode "LAB": assert (768, 0, 1945) == (768, 0, 1946)
debian-12-bookworm-x86, mode "1": assert -470796945 == -6588993366496339844

So it would seem like the histogram calculation is going wrong somehow.

[Yay295]

I tried changing the histogram C data type from long to UINT32 in case it was an overflow error somehow, but that didn't change anything. https://github.com/Yay295/Pillow/actions/runs/10932277548

[Yay295]

Mode "1" was failing because the basic hash function isn't deterministic, so I've changed it to use a deterministic hash. I also added the sum to the list of values being checked, and the sum is correct for the failing tests. So all of the values are being counted, but they're not all going in the expected buckets. This makes me think there's actually something going on in the conversion from "RGB" to "LAB" and "HSV" that is somehow not the same in every environment.

[radarhere]

Rather than introducing the new idea of hashes to our test suite, would it not be simpler to add individual files to check against?

[wiredfool]

Using hashes seems like a now we have two problems sort of issue. Tests should be clear and obvious, and point to a cause. Comparing to a hash says right or wrong, but doesn’t aid debugging.

[Yay295]

I used hashes because each mode's histogram data has at least 256 values, and there are 20 modes. That's a lot of numbers to put in a test file. Using hashes allows for a much smaller set of expected data to be stored, and it still solves the purpose of showing when something isn't right.

[Yay295]

Correct HSV Histogram: [171, 51, 63, 71, 52, 73, 97, 158, 181, 226, 327, 321, 332, 369, 293, 327, 264, 135, 151, 88, 94, 89, 49, 69, 60, 45, 44, 30, 51, 36, 15, 24, 21, 32, 6, 13, 13, 6, 3, 3, 2, 2, 49, 2, 0, 4, 1, 1, 3, 5, 12, 0, 1, 1, 5, 0, 2, 1, 0, 6, 9, 1, 1, 1, 0, 1, 1, 4, 2, 0, 20, 0, 5, 1, 0, 0, 0, 8, 0, 4, 1, 4, 0, 0, 3, 0, 0, 1, 0, 3, 0, 0, 1, 7, 1, 0, 1, 0, 2, 1, 4, 0, 2, 2, 8, 1, 39, 1, 1, 1, 1, 0, 10, 21, 1, 0, 10, 0, 1, 1, 2, 1, 7, 10, 2, 1, 0, 46, 0, 4, 4, 8, 2, 0, 21, 27, 0, 0, 10, 13, 1, 3, 8, 19, 1, 31, 33, 25, 23, 21, 78, 36, 158, 595, 1696, 1140, 262, 174, 150, 100, 48, 114, 183, 72, 110, 59, 134, 156, 134, 861, 52, 148, 121, 85, 81, 197, 278, 592, 412, 49, 30, 62, 63, 14, 68, 14, 17, 236, 130, 8, 14, 37, 34, 9, 0, 62, 21, 14, 168, 15, 110, 5, 7, 10, 4, 108, 9, 48, 24, 4, 6, 2, 144, 0, 4, 16, 34, 7, 2, 18, 11, 157, 13, 0, 39, 29, 58, 2, 24, 11, 13, 97, 56, 79, 6, 34, 11, 10, 9, 39, 155, 12, 14, 14, 112, 16, 48, 53, 21, 29, 100, 43, 77, 48, 31, 0, 37, 25, 36, 32, 21, 54, 48, 49, 18, 26, 38, 35, 41, 27, 40, 52, 34, 46, 36, 44, 37, 38, 34, 29, 31, 28, 34, 21, 37, 27, 101, 178, 21, 34, 186, 24, 81, 21, 35, 39, 39, 22, 66, 33, 14, 45, 28, 62, 14, 54, 20, 107, 30, 46, 57, 35, 49, 43, 47, 22, 60, 42, 17, 125, 23, 34, 34, 102, 46, 65, 166, 29, 96, 21, 26, 161, 80, 41, 36, 116, 93, 53, 57, 22, 12, 218, 27, 70, 50, 45, 114, 71, 62, 90, 62, 66, 61, 59, 68, 48, 87, 17, 117, 119, 59, 70, 77, 102, 59, 100, 110, 72, 119, 99, 144, 204, 223, 284, 222, 354, 223, 165, 95, 123, 180, 77, 92, 176, 74, 70, 99, 101, 129, 90, 128, 86, 120, 100, 83, 105, 102, 107, 79, 112, 85, 105, 100, 119, 117, 98, 137, 161, 84, 157, 127, 141, 168, 203, 144, 164, 129, 113, 99, 95, 74, 85, 87, 69, 82, 31, 114, 86, 49, 79, 57, 78, 72, 48, 54, 49, 91, 61, 68, 65, 63, 57, 38, 67, 45, 52, 39, 56, 45, 47, 41, 40, 34, 44, 38, 37, 32, 35, 43, 16, 42, 31, 21, 26, 22, 22, 35, 25, 21, 22, 14, 16, 21, 20, 23, 15, 18, 26, 16, 17, 14, 19, 16, 18, 15, 23, 13, 15, 21, 14, 18, 17, 16, 19, 14, 18, 13, 18, 4, 6, 12, 9, 10, 8, 4, 6, 1, 1, 1, 3, 0, 121, 0, 1, 0, 0, 0, 3, 1, 6, 5, 6, 14, 31, 38, 75, 147, 289, 350, 348, 332, 294, 223, 174, 212, 163, 209, 232, 199, 211, 167, 190, 160, 168, 147, 154, 130, 88, 90, 76, 80, 71, 62, 59, 69, 71, 64, 59, 59, 49, 64, 59, 44, 50, 59, 48, 45, 45, 52, 55, 53, 43, 37, 51, 48, 45, 48, 37, 56, 54, 51, 52, 50, 45, 42, 44, 31, 45, 53, 40, 34, 45, 39, 39, 43, 43, 30, 39, 36, 22, 26, 25, 27, 22, 19, 25, 23, 18, 19, 12, 7, 15, 8, 18, 14, 20, 12, 19, 15, 5, 9, 15, 15, 9, 18, 16, 19, 19, 15, 19, 17, 15, 17, 21, 16, 25, 26, 25, 25, 31, 24, 15, 21, 18, 17, 11, 26, 27, 19, 25, 25, 21, 28, 27, 24, 25, 23, 19, 20, 23, 22, 29, 32, 24, 31, 26, 35, 22, 33, 29, 18, 24, 29, 25, 34, 32, 30, 33, 26, 34, 37, 37, 35, 42, 49, 45, 54, 62, 110, 167, 178, 166, 116, 119, 102, 112, 108, 117, 131, 111, 140, 133, 106, 94, 91, 119, 105, 114, 142, 93, 121, 182, 218, 308, 295, 188, 182, 204, 154, 155, 88, 86, 84, 72, 67, 64, 52, 59, 42, 44, 50, 31, 47, 42, 33, 42, 31, 49, 44, 54, 42, 30, 35, 38, 35, 33, 26, 35, 37, 27, 24, 25, 27, 26, 24, 31, 27, 29, 33, 26, 36, 33, 18, 29, 35, 38, 27, 316]

Incorrect HSV Histogram: [171, 51, 63, 71, 48, 77, 97, 158, 181, 232, 321, 321, 332, 370, 292, 332, 250, 144, 151, 88, 94, 89, 49, 69, 60, 45, 44, 30, 51, 36, 15, 24, 21, 32, 6, 13, 13, 6, 3, 3, 2, 2, 49, 2, 0, 4, 1, 1, 3, 5, 0, 12, 1, 1, 5, 0, 2, 1, 0, 6, 9, 1, 1, 1, 0, 1, 1, 4, 2, 0, 20, 0, 5, 1, 0, 0, 0, 8, 0, 4, 1, 4, 0, 0, 0, 3, 0, 1, 0, 3, 0, 0, 1, 7, 1, 0, 1, 0, 2, 1, 4, 0, 2, 2, 8, 1, 39, 1, 1, 1, 1, 0, 10, 21, 1, 0, 10, 0, 0, 2, 2, 1, 7, 10, 2, 1, 0, 46, 0, 4, 4, 8, 2, 0, 15, 33, 0, 0, 10, 13, 1, 3, 8, 19, 1, 31, 33, 25, 23, 21, 78, 36, 158, 595, 1696, 1140, 265, 171, 150, 100, 48, 114, 183, 72, 110, 59, 134, 156, 134, 132, 781, 148, 121, 85, 81, 197, 278, 592, 412, 61, 18, 62, 63, 14, 68, 14, 17, 236, 130, 8, 14, 37, 34, 9, 0, 62, 21, 14, 168, 15, 110, 5, 7, 10, 4, 108, 9, 48, 24, 4, 6, 2, 144, 0, 4, 16, 34, 7, 2, 18, 11, 157, 13, 0, 39, 29, 58, 2, 24, 11, 13, 97, 56, 79, 6, 34, 11, 10, 9, 39, 155, 12, 14, 14, 112, 16, 48, 53, 21, 29, 100, 43, 77, 48, 31, 0, 37, 25, 36, 32, 21, 54, 48, 49, 18, 26, 38, 35, 41, 27, 40, 52, 34, 46, 36, 44, 37, 38, 34, 29, 31, 28, 34, 21, 37, 27, 101, 178, 21, 34, 186, 24, 81, 21, 35, 39, 39, 22, 66, 33, 14, 45, 28, 62, 14, 54, 20, 107, 30, 46, 57, 35, 49, 43, 47, 22, 60, 42, 17, 125, 23, 34, 34, 102, 46, 65, 166, 29, 96, 21, 26, 161, 80, 41, 36, 116, 93, 53, 57, 22, 12, 218, 27, 70, 50, 45, 114, 71, 62, 90, 62, 66, 61, 59, 68, 48, 87, 17, 117, 119, 59, 70, 77, 102, 59, 100, 110, 72, 119, 99, 144, 204, 223, 284, 222, 354, 223, 165, 95, 123, 180, 77, 92, 176, 74, 70, 99, 101, 129, 90, 128, 86, 120, 100, 83, 105, 102, 107, 79, 112, 85, 105, 100, 119, 117, 98, 137, 161, 84, 157, 127, 141, 168, 203, 144, 164, 129, 113, 99, 95, 74, 85, 87, 69, 82, 31, 114, 86, 49, 79, 57, 78, 72, 48, 54, 49, 91, 61, 68, 65, 63, 57, 38, 67, 45, 52, 39, 56, 45, 47, 41, 40, 34, 44, 38, 37, 32, 35, 43, 16, 42, 31, 21, 26, 22, 22, 35, 25, 21, 22, 14, 16, 21, 20, 23, 15, 18, 26, 16, 17, 14, 19, 16, 18, 15, 23, 13, 15, 21, 14, 18, 17, 16, 19, 14, 18, 13, 18, 4, 6, 12, 9, 10, 8, 4, 6, 1, 1, 1, 3, 0, 121, 0, 1, 0, 0, 0, 3, 1, 6, 5, 6, 14, 31, 38, 75, 147, 289, 350, 348, 332, 294, 223, 174, 212, 163, 209, 232, 199, 211, 167, 190, 160, 168, 147, 154, 130, 88, 90, 76, 80, 71, 62, 59, 69, 71, 64, 59, 59, 49, 64, 59, 44, 50, 59, 48, 45, 45, 52, 55, 53, 43, 37, 51, 48, 45, 48, 37, 56, 54, 51, 52, 50, 45, 42, 44, 31, 45, 53, 40, 34, 45, 39, 39, 43, 43, 30, 39, 36, 22, 26, 25, 27, 22, 19, 25, 23, 18, 19, 12, 7, 15, 8, 18, 14, 20, 12, 19, 15, 5, 9, 15, 15, 9, 18, 16, 19, 19, 15, 19, 17, 15, 17, 21, 16, 25, 26, 25, 25, 31, 24, 15, 21, 18, 17, 11, 26, 27, 19, 25, 25, 21, 28, 27, 24, 25, 23, 19, 20, 23, 22, 29, 32, 24, 31, 26, 35, 22, 33, 29, 18, 24, 29, 25, 34, 32, 30, 33, 26, 34, 37, 37, 35, 42, 49, 45, 54, 62, 110, 167, 178, 166, 116, 119, 102, 112, 108, 117, 131, 111, 140, 133, 106, 94, 91, 119, 105, 114, 142, 93, 121, 182, 218, 308, 295, 188, 182, 204, 154, 155, 88, 86, 84, 72, 67, 64, 52, 59, 42, 44, 50, 31, 47, 42, 33, 42, 31, 49, 44, 54, 42, 30, 35, 38, 35, 33, 26, 35, 37, 27, 24, 25, 27, 26, 24, 31, 27, 29, 33, 26, 36, 33, 18, 29, 35, 38, 27, 316]

All of the incorrect values are in the "H" part of the histogram. This makes some sense, as the "H" values have the most complicated calculation. It's still not that complicated that I would expect to see differences though.

[hugovk]

Using hashes seems like a now we have two problems sort of issue. Tests should be clear and obvious, and point to a cause. Comparing to a hash says right or wrong, but doesn’t aid debugging.

I agree with this.

Thanks for the suggestion, but let's close this one.

[Yay295]

Would you prefer it if the entire histogram data was stored in the test file? Or do you just not want to test this code at all?