Use Leptonica API to access internals of Box

[stweil]

Signed-off-by: Stefan Weil [email protected]

[stweil]

See also #3673 (comment).

Please review carefully (the modified code is not covered by unittests, and I don't want to introduce new bugs).

[amitdo]

I reviewed this patch. I looks right, but too verbose :(

I approve it.

[egorpugin]

I think we need wrappers here.

auto box_get_geometry(Box *b) {
   struct coords {
      int32_t x;
      int32_t y;
      int32_t w;
      int32_t h;
   } c;
   boxGetGeometry(b, &c.x, &c.y, &c.w, &c.h);
   return c;
}
// and usage:
auto &&next = box_get_geometry(b);
next.x;
auto &&prev = box_get_geometry(b);
prev.w;
auto &&[x,y,w,h] = box_get_geometry(b);
// etc.
auto &&bb = box_get_geometry(b);
bb.x += xshift;
bb.y += yshift;
box_set_geometry(b, bb);

[amitdo]

Another option is to add

struct boxGeometry {
      int32_t x;
      int32_t y;
      int32_t w;
      int32_t h;
   };

and

struct boxGeometry boxGetGeometryAsStruct()

to Leptonica's public interface.

In C++17 code we can use structured binding (like in Egor's example).

[stweil]

So maybe we should postpone this change until a new major release which is using C++17 is made.

[egorpugin]

Next major release can jump into C++20 already.

[amitdo]

@stweil, I don't understand your comment. We already require C++17 since 5.0.0.

[p12tic]

Patch looks good to me.

[zdenop]

Can we merge this or are we waiting for something?