`IPGlobalProperties.DomainName` returns `"(none)"` on Linux

[LukasGelke]

Description

as per Documentation, the property should either return the actual domain name or string.Empty.

Reproduction Steps

• have a Linux install, where the domain is missing(?)
• use eg. pwsh
• run [System.Net.NetworkInformation.IPGlobalProperties]::GetIPGlobalProperties().DomainName

Expected behavior

either return the actual domain name or string.Empty

Actual behavior

returns "(none)"

Regression?

(we're using C# on Linux for the first time here, so no idea)

Known Workarounds

• pwsh
• $h = [System.Net.DNS]::GetHostname()
• [System.Net.DNS]::GetHostEntry($h).HostName
• returns equivalent to BUILD-SERVER-LX.company.local

But this approach might be unreliable if the DNS is not reachable, or perhaps misconfigured(?)

Configuration

Local:

• WSL-Ubuntu
• Linux MY-WORK-PC-123 4.4.0-19041-Microsoft #4355-Microsoft Thu Apr 12 17:37:00 PST 2024 x86_64 x86_64 x86_64 GNU/Linux
• /etc/hosts contains a line a la 127.0.1.1 MY-WORK-PC-123.company.local MY-WORK-PC-123 (actual names replaced)
• tests (via VS-WSL) are passing

Build Server:

• Linux BUILD-SERVER-LX 5.15.0-113-generic #123-Ubuntu SMP Mon Jun 10 08:16:17 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
• no such hosts line
• tests fail, as the FQDN determined is "BUILD-SERVER-LX.(none)", which was (with additional strings) passed to new Uri(...)

This issue was caught by our tests (net6.0), but [Environment]::Version (pwsh) shows 8.0.6 (ob both machines) and it experiences the same behavior (build server)

Not sure if C# gets the information from the hosts file, just something i noticed. WSL re-creates the file upon restart, and on the build server I currently do not have sufficient permissions to change the host file.

Other information

No response

[dotnet-policy-service[bot]]

Tagging subscribers to this area: @dotnet/ncl See info in area-owners.md if you want to be subscribed.

[wfurt]

This calls getdomainname under the cover e.g. this is not DNS domain name. However the documentation is clear about empty string and we return null on Linux. So we should ether update the code to fix the nullability annotation and documentation. I feel former would be easier and perhaps more appropriate.