Fix build when iterators are raw pointers

[georgthegreat]

Similar to #11879, typename Iterator might be just a raw pointer. The most corrent way would be using std::iterator_traits to get the type properly, but it seems that current static_cast can be simply removed.

[rouault]

There's a typo in this commit (extra parenthesis), but I would be more confortable if this change was submitted upstream to https://github.com/p-ranav/argparse and accepted there prior to here

[georgthegreat]

Thanks!
I have noticed the typo too, but was waiting for two other PRs to be merged.
I will create the PR to upstream, let's keep this one open for now.

[georgthegreat]

I filed an upstream PR here:
p-ranav/argparse#399

[georgthegreat]

@rouault it seems that upstream is not quite active.
Maybe we can accept this change without upstream accepting it?
The code in gdal is different from upstream anyway.

[rouault]

can you precise which environment requires this change ? GDAL tests on quite a large number of platforms and we haven't hit such issue

[coveralls]

coverage: 70.351% (+0.3%) from 70.047%
when pulling 454817d on georgthegreat:patch-2
into 26fc9a3 on OSGeo:master.

[georgthegreat]

We use patched libc++ with iterators replaced with raw pointers.
You can observe our sources e. g. here:
https://github.com/yandex/perforator/blob/c680583aea53e44ebee0189ac5d8b146f09bb5d2/contrib/libs/cxxsupp/libcxx/include/__config_site#L21

gdal is used internally in our monorepo, but libc++ is the same.

[jratike80]

We use patched libc++ with iterators replaced with raw pointers.

Would the change help anybody else than you in the GDAL community?

[georgthegreat]

We use patched libc++ with iterators replaced with raw pointers.

Would the change help anybody else than you in the GDAL community?

Yes, it will make the code simpler by removing unnecessaries.