Avoid byte[] allocation for IPAddressUtil.IsMulticast #68591
Conversation
ghost
added
the
community-contribution
|
Tagging subscribers to this area: @dotnet/ncl Issue DetailsPing.RawSocket his its own inline implementaion of this, maybe this
|
Bond-009
force-pushed
the
ipaddressutils
branch
from
50afe26 to
7750536
Compare
src/libraries/System.Net.NetworkInformation/src/System/Net/NetworkInformation/IPAddressUtil.cs
Outdated
Show resolved
Hide resolved
Bond-009
force-pushed
the
ipaddressutils
branch
from
7750536 to
2557f87
Compare
Ping.RawSocket his its own inline implementaion of this, maybe this should be consolidated somewhere? src/libraries/System.Net.Ping/src/System/Net/NetworkInformation/Ping.RawSocket.cs L101
Bond-009
force-pushed
the
ipaddressutils
branch
from
2557f87 to
911b20e
Compare
| byte firstByte = address.GetAddressBytes()[0]; | ||
| #pragma warning disable CS0618 // using Obsolete Address API because it's the more efficient option in this case | ||
| byte firstByte = (byte)address.Address; | ||
| #pragma warning restore CS0618 | ||
| return firstByte >= 224 && firstByte <= 239; |
There was a problem hiding this comment.
| return firstByte >= 224 && firstByte <= 239; | |
| return firstByte is >= 224 and <= 239; |
Once roslyn optimizes this (cf. dotnet/roslyn#60534), codegen improvement comes for free.
There was a problem hiding this comment.
I understand these are different constructs as far as Roslyn is concerned, but it's an unfortunate pit of failure if these otherwise identical expressions end up being optimized differently. I'd really like for us to not have to change all such range checks everywhere in order for the right things to happen.
Ping.RawSocket his its own inline implementaion of this, maybe this
should be consolidated somewhere?
src/libraries/System.Net.Ping/src/System/Net/NetworkInformation/Ping.RawSocket.cs
L101