Extract 7z archive with UTF-8 filename

[hexabyte23]

Hi all,

For win10, Is there a parameter to set to extract an 7z archive with UTF8 filenames ?

short ExtractorPrivate::writeData(struct archive_entry* entry) {
...
// UTF-8 in archive entry messes up when extracting under Windows
   // when UTF-8 is not set, to fix this we first get the archive
   // entry pathname in raw bytes then convert it to wide characters
   // and set it has the new pathname which should make libarchive
   // handle it better.
...
}

UTF-8 is not set where ?

When debugging

Not sure that at line 902

auto wstr = ptname.toStdWString();

fix the problem, as wstr still showing bad characters

As always in macOS, its working like a charm.

Thanks

[hexabyte23]

on Master branch, this should fix the problem. Its working for me on macOS and Win10 :

      // UTF-8 in archive entry messes up when extracting under Windows
      // when UTF-8 is not set, to fix this we first get the archive
      // entry pathname in raw bytes then convert it to wide characters
      // and set it has the new pathname which should make libarchive
      // handle it better.

      // Get current pathname
      auto ptname_cstr = archive_entry_pathname(entry);

      // Check if UTF-8
      if (isUTF8(ptname_cstr)) {
         auto w = archive_entry_pathname_w(entry);
         archive_entry_copy_pathname_w(entry, w);
/*
        auto ptname = QString::fromUtf8(ptname_cstr);

        auto wstr = ptname.toStdWString();
        archive_entry_copy_pathname_w(entry, wstr.c_str());*/
      }

      ret = archive_write_header(m_ArchiveWrite.data(), entry);
    } else {

[antony-jr]

This is really weird @hexabyte23. What standard of C++ are you using. Anyhow I need to look into this.

archive_entry_copy_pathname_w(entry, wstr.c_str());*/

I don't know how this works with this line commented out.

[antony-jr]

btw this is definitely again related to libarchive/libarchive#587

[hexabyte23]

Sorry , to clarify, I just comment out your fix of 2023

auto ptname = QString::fromUtf8(ptname_cstr);
auto wstr = ptname.toStdWString();
archive_entry_copy_pathname_w(entry, wstr.c_str());

that seems not working properly with VC2019 compiler, by mine:

auto w = archive_entry_pathname_w(entry);
archive_entry_copy_pathname_w(entry, w);

[antony-jr]

Okay, this makes sense. Let me test this on my end.