WIP: Fix display of images with non-1 indices

[timholy]

Until I finish my fixes to restrict, we don't know if there are more changes needed. So don't merge yet.

[codecov-io]

Codecov Report

Merging #9 into master will not change coverage.
The diff coverage is 100%.

@@          Coverage Diff          @@
##           master     #9   +/-   ##
=====================================
  Coverage     100%   100%           
=====================================
  Files           4      4           
  Lines         117    120    +3     
=====================================
+ Hits          117    120    +3

Impacted Files	Coverage Δ
src/imshow.jl	100% <100%> (ø)	⬆️
src/encodeimg.jl	100% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9d134d5...7900671. Read the comment docs.

[Evizero]

This will need a REQUIRE bump for ImageTransformations 0.2.0 after the merge of JuliaLang/METADATA.jl#8789

(just making a note before I forget about it)

[timholy]

I have a more complete PR coming, just waiting for that METADATA PR to merge. I'm not sure the tests cover enough cases of non-1 indices, though, so I'll be curious to see what you think.

[timholy]

OK, this sort of works. Two issues:

• are there important cases that need testing but aren't?
• one oddity is that if you run the tests in the REPL, they pass, but if you run the tests a second time in the same session, they don't. That doesn't seem to happen on master.

[Evizero]

I'll take a look after a wake-up coffee. Thanks for doing this!

[Evizero]

The test oddity that you ran into was a consequence of the last tests overwriting the default color-depth to 24 bit https://github.com/JuliaImages/ImageInTerminal.jl/blob/master/test/tst_baseshow.jl#L20. Thus running the tests again caused the tests to fail that build on the assumption that the default depth is 256.

I pushed a fix to your branch

[Evizero]

Concerning tests. I think the current tests look great. I like how you reuse other reference files. This tests output consistency, and keeps the memory footprint a bit in check.

[Evizero]

Test checklist for the changed functions:

• imshow
• encodeimg(::SmallBlocks, ::TermColorDepth, ::Matrix) (lighthouse)
• encodeimg(::BigBlocks, ::TermColorDepth, ::Matrix) (lena)
• encodeimg(::SmallBlocks, ::TermColorDepth, ::Vector)
• encodeimg(::BigBlocks, ::TermColorDepth, ::Vector)

To cover the two remaining Vector tests, an adaption of those two testsets should suffice.

• https://github.com/JuliaImages/ImageInTerminal.jl/blob/master/test/tst_encodeimg.jl#L235-L242
• https://github.com/JuliaImages/ImageInTerminal.jl/blob/master/test/tst_encodeimg.jl#L276-L289

[Evizero]

A quick Vector test in the REPL revealed the following remaining OffsetArray obstacle in ImageTransformations: https://github.com/JuliaImages/ImageTransformations.jl/blob/master/src/resizing.jl#L180

[Evizero]

If you are too busy for these additional changes in this PR, let me know. I can't implement them right this instance, but I'd be happy to take care of them soonish.

[timholy]

An error I first saw yesterday with ImageCore cropped up on AppVeyor. I don't know why that happened, but hopefully the fix is easy.

[Evizero]

needs a bump to 0.2.1

[Evizero]

Very cool. I am excited about this change!

[Evizero]

lgtm. go for merge after CI passes?

[Evizero]

Thanks! As primarily a terminal user this change will make hacking on the WarpedView a lot nicer