IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
get stuck waiting on no file descriptors, with no timeout, so it sits
forever. I need to fix that bug separately, but to prevent build farm
machines being totally stuck, this timeout will be used.
This makes more clear where the functions belong to. Also
the rule will be that lowercased function names are not
part of "official" libnet API (though it doesn't mean one
absolutely cannot use them).
rafal
- use loopback interface only, otherwise nbtd will answer with a real
ip, which the socket wrapper code will fail to connect to
- enable eadb for posix testing
- sleeping to wait for smbd to start is pointless, as smbd won't start
till its fifo is opened due to the fifo redirection on stdin
- sleep for a few seconds to allow for nbt server to register names after opening fifo
- use bcast name resolution, so we don't rely on /etc/hosts being right
just do the write. This is needed because the write could return
-1/EAGAIN for dgram sockets, if the socket buffer is nearly full. The
epoll loop then goes on forever.
This was causing some failures in 'make test'
chewing far too much cpu, which is unfair on our volunteer systems
andrew, sorry about this, I think we need RPC-SAMLOGON-FULL, and make
RPC-SAMLOGON only do a basic test
written code, and it doesn't work or even compile yet. I am committing
it to make it easier to discuss the approach with jelmer and tpot.
The intention is that this code will eventually end up being mostly
auto-generated (with the utility functions split out, just like
librpc/ndr/*.c)
to change this later but that will be pretty easy.
We can use this type to pass around pointers to handles in C. Talloc
allows us to do type checking too.
I still think this URI syntax for a 'protocol' which has no network component
is stupid; it should just be
sqlite:local_file
or
sqlite:/full/path/to/file
but there is enough precedent to warrant the behavior that Simo wants that it
is not worth arguing.
