Attention is currently required from: arehbein, laforge.
9 comments:
File src/stream.c:
I'd go for EINVAL here. […]
Done
I'd usually say that it is "established" after connect succeeds?
Sounds better. I'll also change it in the old code (some of the other changes as well)
Patch Set #1, Line 654: error %d to send
sounds a bit like we are "about to send ... as an error". […]
Ack
Patch Set #1, Line 683: osmo_iofd_setup
osmo_iofd_setup may return NULL. missing error handling.
Ack
Patch Set #1, Line 1715: LOGL_DEBUG
this message might be worth INFO or even NOTICE, I think.
Done
Patch Set #1, Line 1758: OSMO_ASSERT
why is this commented out? Is the idea to permit a srv_link in ofd mode and the stream_srv in osmo_ […]
Wow, I just saw that osmo_stream_srv_link_create_iofd() is a 100% copy of the old variant.
I would say I just remove that version - I'm not sure what we gain handling accept() with osmo_io.
Patch Set #1, Line 1762: LOGP(DLINP, LOGL_ER
if the memory allocator fails, it's always an ENOMEM situation, so it's not realyl worth calling str […]
Ack, removing it in osmo_stream_srv_create() as well.
Patch Set #1, Line 1768: osmo_iofd_setup
osmo_iofd_setup may return NULL. […]
Ack
Patch Set #1, Line 1775: talloc_free
is there nothing we need to call to undo whatever osmo_iofd_setup() has done above?
No, it simply allocates memory and fills in the struct. Everything else is done in osmo_iofd_register()
To view, visit change 33193. To unsubscribe, or for help writing mail filters, visit settings.