<hannes>
mato: nice ocaml-freestanding! I just confirmed that it still builds and works on FreeBSD here
jermar has quit [Ping timeout: 246 seconds]
<mato1>
hannes: good stuff! i'm finally working on the PASS/FAIL script for running the solo5 standalone tests
<hannes>
nice
yomimono has joined #mirage
<hannes>
for console unikernels I can imagine how to do that... for network... you need some remote kill mechanism!?
<mato1>
yeah, network is tricky, you also have two processes to deal with, the unikernel and whatever is talking to it. working on it.
mort___ has joined #mirage
mort___ has quit [Ping timeout: 256 seconds]
<hannes>
mato1: I've been working on sth in respect to testing TCP/IP in the FreeBSD kernel... with a remote helper process that emits TCP/IP frames (via system calls or raw)... what I do (in OCaml) is to setup a control connection, do tests (with expected outcome etc., each thing on the remote side is blocking (I send command 'connect to foo' and wait until connect returned)
abeaumont has quit [Ping timeout: 260 seconds]
<hannes>
of course that only works if you're getting to a network connection... same could be played with using syslog and a collector as part of the program running the test (once a test is finished, the control connection is teared down which leads the remote process to die -- or the unikernel to stop)
mort___ has joined #mirage
agarwal1975 has joined #mirage
yomimono has quit [Quit: Leaving]
seangrove has joined #mirage
jermar has joined #mirage
brson has joined #mirage
mort___ has quit [Quit: Leaving.]
mort___ has joined #mirage
<mato1>
hannes: ok, i have a gnarly shell script, but it works! (locally and on Travis). still need to document it properly, tomorrow.
<miragebot>
mirage/master 6c336b6 Liang Wang: add owl to documentation
miragebot has left #mirage [#mirage]
brson has quit [Ping timeout: 256 seconds]
brson has joined #mirage
copy` has joined #mirage
<dave_tucker>
question: looking to make some contributions to mirage-tcpip and having trouble seeing where the initial handshake takes place (in the code)... i'm planning on vetting the congestion control against rfc 2581 - starting with ensuring the mss sizes in use are correct