More migration to size_t

[piponazo]

Continuation of the work done to avoid some of the static_casts we have around the code base. See other recent PRs: #2118 #2109 #2062

[lgtm-com]

This pull request introduces 6 alerts when merging b20e073 into 7aae68e - view on LGTM.com

new alerts:

• 4 for Wrong type of arguments to formatting function
• 1 for Unsigned comparison to zero
• 1 for Comparison result is always the same

[neheb]

Seems good at first glance. I do wonder whether or not static_cast is needlessly verbose. In my code I prefer to use functional casts, eg. size_t(x).

[piponazo]

Seems good at first glance. I do wonder whether or not static_cast is needlessly verbose. In my code I prefer to use functional casts, eg. size_t(x).

Whenever you use C++ it is better to use the C++ casts because they just do what you specify. The usage of C-style casts can be actually dangerous in some cases, because they try to do different kind of casts in a determined order. You can find more information here:
https://stackoverflow.com/questions/332030/when-should-static-cast-dynamic-cast-const-cast-and-reinterpret-cast-be-used

Anyways, as I mentioned in this PR and in others, the intention is to remove as many static_casts as possible from the codebase. They are normally a code smell indicating that a better data model could reduce the need of them.

[clanmills]

I like Rosen's idea to use function casts. The static cast syntax is ugly and the C cast is hackery.

[neheb]

Well, the function cast is the same as a C cast. Just looks less ugly and avoids the C cast clang-tidy warning.

I like using it for simple types like size_t where the compiler's not really going to do anything strange.

[clanmills]

The equivalent of a cast in Pascal was size_t(...). When I started to work in C, I was shocked by the cast. I couldn't believe that a language designer would use anything so hideous. My first program did something with argv, and I had a function with the signature something(char*** args). It was my first program for "The IBM PC". After being used the multi-window Apollo workstation, it was quite a shock. The C tradition of vulgar syntax continued with the static_cast<voodoo> .... C with all its *'s and ->'s looks like Custer at the Little Bighorn.

I read an interview with Stroudstup and he said something like "C++20 is the language I set out the make in the first place". I wonder if it will be as nice as Pascal.

[codecov]

Codecov Report

Merging #2145 (4355d63) into main (173ebb2) will decrease coverage by 0.04%.
The diff coverage is 68.00%.

@@            Coverage Diff             @@
##             main    #2145      +/-   ##
==========================================
- Coverage   63.31%   63.27%   -0.05%     
==========================================
  Files          97       97              
  Lines       19152    19159       +7     
  Branches     9724     9721       -3     
==========================================
- Hits        12127    12123       -4     
- Misses       4762     4778      +16     
+ Partials     2263     2258       -5     

Impacted Files	Coverage Δ
app/actions.hpp	100.00% <ø> (ø)
include/exiv2/exif.hpp	100.00% <ø> (ø)
include/exiv2/image.hpp	100.00% <ø> (ø)
include/exiv2/metadatum.hpp	72.72% <ø> (ø)
include/exiv2/pgfimage.hpp	100.00% <ø> (ø)
include/exiv2/xmp_exiv2.hpp	92.30% <ø> (ø)
src/cr2header_int.hpp	50.00% <ø> (ø)
src/cr2image.cpp	18.98% <0.00%> (ø)
src/makernote_int.hpp	91.66% <ø> (ø)
src/olympusmn_int.cpp	43.63% <0.00%> (ø)
... and 38 more

---

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 173ebb2...4355d63. Read the comment docs.

[lgtm-com]

This pull request introduces 4 alerts when merging 4355d63 into 173ebb2 - view on LGTM.com

new alerts:

• 4 for Wrong type of arguments to formatting function

[piponazo]

The PR is now ready for review after I fixed problems reported by the CI jobs.

@neheb & @clanmills , IMHO I think that the discussion about the casting style is offtopic (This PR is probably not the best place to discuss such things. I am actually mostly removing casts here). Feel free to open another issue or trigger this conversation in the team chat.

[clanmills]

You're quite right, @piponazo. Apologies for the noise.

[piponazo]

You're quite right, @piponazo. Apologies for the noise.

No worries, I appreciate interesting discussions in the right channels 😁