S. B. Tam

<format>: incorrect output for some floating-point values

1

# Describe the bug Revealed by libc++ test `std/utilities/format/format.functions/locale-specific_form.pass.cpp`. When neither *precision* nor *type* is specified, a floating-point value should be formatted as if by `to_chars(first, last, value)` ([\[format.string.std\]](http://eel.is/c++draft/format.string.std#tab:format.type.float-row-9)). But...

# Describe the bug Revealed by libc++ test `std/utilities/format/format.functions/locale-specific_form.pass.cpp`. When formatting a floating-point number, digit separators are not taken into account when computing the width of output, which may cause...

# Describe the bug `` uses a loop to find the beginning of the exponent part in order to insert trailing zeroes. But when `_Specs._Alt || _Specs._Localized` is true, the...

Enable `__cpp_lib_concepts` for EDG, part 1

1

This PR makes `__cpp_lib_concepts` defined in C++20 mode for all supported compilers, after adding workarounds for various EDG concepts bugs. I have opened two follow-up PRs for non-essential and largely...

`<ranges>`: `lazy_split_view` should use `_Non_propagating_cache` instead of `_Defaultabox`

2

Revealed by libc++ tests `std/ranges/range.adaptors/range.lazy.split/ctor.copy_move.pass.cpp` and `std/ranges/range.adaptors/range.lazy.split/range.lazy.split.inner/iter_swap.pass.cpp`. [\[range.lazy.split.view\]](http://eel.is/c++draft/range.lazy.split.view) says `lazy_split_view` stores a *`non-propagating-cache`*`

This PR mechanically changes preprocessing conditions involving `__cpp_lib_concepts` to say `_HAS_CXX20`, and then removes some redundant conditions. Most of the replacement is done in the first commit (5e05034a12d2a527038e3245ede2a7dccfd1a954), which is...

After #4296, `concepts_20_matrix.lst` and `concepts_latest_matrix.lst` will be near-identical to their usual counterparts. This PR changes the tests to include the usual matrices instead of the concepts ones, and then removes...

Use `if constexpr` in locale facets

1

Towards #189

This PR attempts to build STL targeting ARM64EC as part of the CI process. Note that neither benchmarks nor tests are built. I am completely novice to Azure Pipelines, and...