| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On some versions of Solaris, we observed a strange effect of close(2)
on a socket: After the server (here winbindd) called close, the client fd
was not marked as readable for select. And a write call to the fd did
not produce an error EPIPE but just returned as if successful.
So while winbindd had called remove_client(), the corresponding smbd
still thought that it was connected, but failed to retrieve answers
for its queries.
This patch works around the problem by forcing the client fd to
the readable state: Just write one byte into the socket before
closing.
Michael
|
|
|
|
| |
Guenther
|
|
|
|
| |
Guenther
|
|
|
|
| |
Guenther
|
|
|
|
|
|
|
|
| |
"test -r"
instead of test -e, which /bin/sh on solaris does not know.
Michael
|
|
|
|
|
|
|
| |
This checks the writeability of a config source specified by the
config source name.
Michael
|
|
|
|
|
|
| |
This allows for per-config-source checking of write support.
Michael
|
|
|
|
| |
Michael
|
|
|
|
| |
Michael
|
|
|
|
|
|
|
|
|
| |
Interpret the source string as a file name when it contains
a ':' sign but the initial part is not a known backend.
This might occur even implicitly when "%T" is used in an
include file name (even though this is not realistic..).
Michael
|
|
|
|
| |
Michael
|
|
|
|
|
|
|
| |
In a clustered environment, the registry backend needs messaging
to be set up since ctdb requires this.
Michael
|
|
|
|
| |
Michael
|
|
|
|
| |
Michael
|
|
|
|
|
|
| |
Use whole argument string in test_one_global_option(), not only first.
Michael
|
|
|
|
|
|
|
|
|
| |
Note: The change of commit r21962 (or fe962d2e48, aka f540c18b2)
that was partly lost in the SVN->git glue commit e5a951325a6
(aka 5c6c8e1fe93f34) is not needed any more since the handler
for option "-s" calls set_dyn_CONFIGFILE().
Michael
|
|
|
|
|
|
|
|
|
| |
and other systems where sh does not support "export FOO=bar"
by separating setting and exporting the variable.
Thanks to Yasuma Takeda <yasuma@osstech.co.jp> for the patch.
Michael
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a central piece of the "merged build" thing: Forward named pipes from
samba3 to samba4. This patch is not finished yet, as we will have to forward
the smb-level authentication information to samba4, but I'm pushing this patch
already to demonstrate the implementation without clutter.
It adds an intermediate parameter
np:proxy = srvsvc samr winreg wkssvc ... and so on
that states which of the pipes should be forwarded to the s4 unix domain socket
DEFAULT. The parameter is intermediate because once we have a proper endpoint
mapper implementation, this information will be retrieved out of a database.
If anybody wants to try this, do the merged build and configure s4 with
server services = samba3_smb, rpc, nbt, wrepl, ldap, cldap, kdc, drepl
samba3:smbd = /data/inst/sbin/smbd
and s3 with
auth methods = guest netlogond
np:proxy = srvsvc samr winreg wkssvc netlogon ntlsa ntsvcs lsass lsarpc netdfs \
rpcecho initshutdown epmapper svcctl eventlog drsuapi
Then run rpcclient against samba4. It will fork s3, which authenticates against
s4, and then forwards the rpc requests to s4.
Volker
|
| |
|
|
|
|
|
|
| |
Jelmer, when I include lib/util/util.h into some s3 file I get errors, this is
why I put those prototypes here as a workaround. Might be fixed differently
later.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
- There were a few places in SMBC_getatr() that returned -1 instead of a
boolean. -1 was intended to mean error, but that's what False is for, and
the usages of this function assume that it returns a boolean with False as
the error condition.
Derrell
|
| |
|
| |
|
|\ |
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
This patch removes the gensize option from krb5pac.idl because it
generates an unused function that has been causing warnings during the
s3 build. I re-ran 'make samba3-idl' to update the generated files
for s3. For s4 I ran a full build with no problems.
|
| |
| |
| |
| | |
Guenther
|
| |
| |
| |
| | |
Guenther
|
| |
| |
| |
| | |
Guenther
|
| |
| |
| |
| | |
Guenther
|
| |
| |
| |
| | |
Guenther
|
| | |
|
|/ |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When libcli-util was moved to the top level
(3be0f6ea56ed8e43cd287ed020e942efb675b87b) it introduced a build
warning. Since ndr_map_error2ntstatus depends on the definition of
enum ndr_err_code and libndr.h can't be easily included in error.h, I
moved the definition of enum ndr_err_code to libndr.h.
This also eliminates the need for the duplicate definition in proto.h
|
|/ |
|
|\ |
|
| |
| |
| |
| | |
Guenther
|
| | |
|
|/
|
|
|
| |
otherwise (to clarify we can also pass in structs smaller than
sockaddr_storage, such as sockaddr_in).
|
|\ |
|
| |
| |
| |
| | |
Guenther
|
|/ |
|
|\ |
|
| | |
|
|/ |
|
|\ |
|
| | |
|