boost/sort/common/util/insert.hpp #includes itself

[urnathan]

working on clang modules (as a proxy for C++20 header-units), I see that boost/sort/common/util/insert.hpp #includes itself. That seems unnatural.

boost-includes.txt

[nigels-com]

Resolved via PR #78, can be closed now.

[fjtapia]

The insert sort algorithm is in the file insert_sort.hpp. This file has an include insert.hpp, which contains quick insert functions in an ordered vector. DOES NOT INCLUDE HIMSELF These functions were initially part of the insert_sort algorithm. But they were later implemented in the insert sort.hpp file, so right now both the include and the insert.hpp files are unnecessary. The file test_insert_sort.cpp also tested those functions, and they part were deleted from the file They are remains of an old implementation, which I forgot to clean when I uploaded the files to the repository, as well as a function placed between comments in the file. I had modified insert_sort.hpp, delete insert.hpp, and modified, too, test_insert_sort By the way, what program did you use to detect this? Thanks Francisco El lun, 11 sept 2023 a las 23:30, Nigel Stewart ***@***.***>) escribió:

…

Resolved via PR #78 <#78>, can be closed now. — Reply to this email directly, view it on GitHub <#64 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5O5GGBXEA6OWELBP4KDXLXZ57I7ANCNFSM5763OKJA> . You are receiving this because you were assigned.Message ID: ***@***.***>

[urnathan]

By the way, what program did you use to detect this?

I probably hacked the compiler to avoid its header-macro optimization and then ran some python of the preprocessed output. FWIW I've been wondering about the usefulness of a compiler warning for this kind of thing.