-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Open Port Request on Sending Side of IpSocket #2683
Conversation
Looks like it passed CI. We'll need to correct spelling. Your spelling is right, so we need to add the word to the expect list. |
The list is at: .github/actions/spelling/expect.txt. Just place "spammed" in there is mostly alphabetical order. I can help this afternoon, but the above shows you how to should you get to it sooner. |
Fixed the spelling issue. Will review now. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Well I feel dumb, these comments were left as pending for several weeks.
@LeStarch I added a unit test to TcpClient and TcpServer to show the socket auto opening, so I think this should be okay to merge (if it passes the checks). |
Better way of doing this:
|
aa38100
to
b3d8dbb
Compare
…oved behavior to the component level by placing thread protection behavior to SocketReadTask, now called SocketComponentHelper. Calls to socket functionality from TcpClient/TcpServer/Udp will go through SocketComponentHelper rather than directly to the library. This also implemented the original desired functionality of reopening a client socket on a send call rather than only in receive
b3d8dbb
to
baebba1
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
CodeQL found more than 20 potential problems in the proposed changes. Check the Files changed tab for more details.
…tests whos name includes more than the module name
…buffer equality check would fail because the size of the original buffer (owned by the unit test) gets set to 0, sometimes. Maybe 1/10 runs. I changed it to save the size of that buffer before it does the send and use that saved value in the equality check
@LeStarch other than the RPi known issue, I think this PR is ready for review |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have lots of questions, but my guess is this is ready to merge once the dead code and TODOs are deleted!
Fantastic work!
Change Description
Update IpSocket to check socket state on sending side and open socket if necessary
Rationale
This fixes a bug encountered while using IpSocket. The server would close the TCP connection and then the client would try to send and fail because previously IpSocket would only re-open a TCP connection when it received over the socket.
Testing/Review Recommendations
In my project's version of fprime we diverged at fprime 3.0. I added implementations for the close() port in the TcpClient, TcpServer, and Udp as well as unit tests. This change also fixed the issue we were seeing with sending data, but the socket not being there.
This does not include the unit tests or implementations for close() in the ByteStreamDriverModel because I wanted to at least get this part of the change in. I'll work on adding them.
I also ended up commenting out line 85 of Drv/Ip/SocketReadTask.cpp because when I was running other testing without the TCP server, I was being spammed with that logger message.
Future Work
Adding implementations for close() to TcpClient, TcpServer, and Udp as well as unit tests.
(Pipeline failed on "spammed" but I stand behind my spelling, but perhaps not that comment)