If --from-stdin was used non-interactive (e.g. in a script
echoing connection parameters) xfreerdp looped endlessly
when a certificate needed to be trusted because
fgets returns eof.
Now a hint is printed that xfreerdp should be run without
--from-stdin to set the certificate trust.
(cherry picked from commit 4d870b9421)
- Discard new lines with scanf.
- If domain is left empty (STRG+d pressed) clear EOF flag - this fixes
an enless loop in xfreerdp on OS X when input is required after a certificate
warning.
(cherry picked from commit 7a002270d8)
When --from-stdin is given and parameters are passed via pipe/file
don't use passphrase function for the password.
This fixes#698.
(cherry picked from commit d846024e85)
This switch will prompt username, password, domain and connection
host if not specified at the command line.
Signed-off-by: Rex Tsai <chihchun@kalug.linux.org.tw>
(cherry picked from commit a31f0c62e6)
Conflicts:
client/X11/xfreerdp.1.xml
client/X11/xfreerdp.c
include/freerdp/settings.h
libfreerdp-core/settings.c
libfreerdp-utils/args.c
Due to the client is parsing the args in pre_connect,
we should start loading the extensions after client parsing the args.
Signed-off-by: Ying-Chun Liu (PaulLiu) <paul.liu@canonical.com>
(cherry picked from commit 48ad5feb0a)
Conflicts:
libfreerdp-core/rdp.c
These server flags tell the SVC layer to temporarily stop sending
data, but are currently not implemented. This patch purposefully
ignores them so they don't cause a crash.
(cherry picked from commit 4bda6b5cbc)
This fixes --sec rdp --salted-checksum in combination with farm redirect.
The counters are implicitly set to 0 in rdp_new, but they also has to be reset
after redirects. Resetting the counters when the keys are reset seems like the
cleanest solution.
EBX (or RBX on x86_64) is used for the PIC base address.
The current cpuid inline assembly only saved ebx which is fine
under i386 but only half of the fun on x86_64.
fastpath_input_pdu_init already reserved space for fastpath_get_sec_bytes()
which thus already was included in stream_get_length() in
fastpath_send_input_pdu(). Adding sec_bytes again just added extra invalid (but
correctly hashed/encrypted) bytes to the PDU.
Writing to the pipe may block if the pipe is full. Holding the mutex while writing prevents listeners from
aquiring the mutex and unblocking the write. This leads to a deadlock situation.
