[doc] Fix Doxygen parameter names in io and the extensions

[karpovantonme]

Follow-up to #792, covering the parts of the library that PR deliberately left out. I said there I would send these if wanted; Marshall Clow also suggested it in boostorg/algorithm#131 after merging the same kind of fix there. If you would rather not have them, close this and nothing is lost.

Nine \param / \tparam names that do not match the declaration below them.

io

read_and_convert_view.hpp — all four Device overloads document \param file, the argument is device. The descriptions already read "It's a device", so only the names were wrong.

read_and_convert_image.hpp:241 — the Device overload documents file_name File name. Must satisfy is_supported_path_spec metafunction., which is the string overload's line copied down. The three other Device overloads in the same file (47, 134, 189) already say device, and I followed the one at 189 word for word.

reader_base.hpp:50 — init_image documents @param info, the argument is settings.

extensions

extension/io/png/tags.hpp:681 — the image_read_settings constructor documents \param gamma, the argument is screen_gamma.

extension/dynamic_image/algorithm.hpp:155 — the two-argument copy_and_convert_pixels documents \tparam Type, the pack is Types. The three-argument overload eleven lines above spells it correctly, so the two overloads describe the same pack differently.

extension/dynamic_image/any_image.hpp:162 — const_view documents \tparam Types, the pack is Images. view() directly above it is correct.

Checked and left alone

read_view.hpp, read_image.hpp and the remaining Device overloads genuinely name the argument file, so their \param file is right. GIL uses both spellings and I did not touch that.

Two I did not fix, and why

io/read_image_info.hpp:46 and :89 document \param tag on overloads whose second parameter is unnamed (FormatTag const&). So the name really is absent, but the fix would mean naming a parameter in the signature rather than editing a comment, and that is your call rather than mine.

My checker also gets these two wrong in a way worth stating: it reports the parameter list as file, const, dummy, reading const in FormatTag const& as an argument name. So treat the two counts above as nine verified by hand, not eleven from a tool.

One more, not a mismatch so not included: the same png constructor never documents apply_screen_gamma at all.

Checks

Comments only, the diff has no non-comment lines. Re-running the same check on this branch reports 0 remaining in io and the extensions, down from 11 on develop at dcfcc69.

[mloskot]

Thank you!

Let's leave the unnamed parameters as they are. I think it was my preference to use the east-const convention, see https://hackingcpp.com/cpp/design/east_vs_west_const