The 2nd loop is supposed to free the privates and also clear the pointer
to them. But it instead of clearing the walk'ed pointer, it only clears
the one in the last screen left from the upper loop.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
The client index (in the client table) can never be negative, thus
make it an unsigned short.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
If we've already got the ScreenPtr, there's no need to retrieve it again
via the screen number, which we're getting via the ScreenPtr that we've
already got.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Use the payload size from the rpcbuf, but also compare that with our
computation as sanity check and log error on mismatch.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Instead doing an exrtra loop for our own calculation of the
payload size, just the wpos from rpcbuf.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
In preparation for subsequent refactoring, assembling the reply header
at once, instead of doing it piece by pice.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
The payload length is already known after writing everything to the
buffer, so no need for extra size computation anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Preparation for subsequent changes: instead of returning early when no
payload to send, move that into a conditional block.
(the WriteRpcbufToClient() call on potentially empty buffer is intentional)
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
As a prepration for upcoming changes, move assembly of the the reply
header further downwards where all values are already known, so we'll
have to touch it only once. Also eliminate the separate write path for
the case where Xv is disabled.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
For easier understanding the code, declare the variables right where
they're assigned first.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
* use x_rpcbuf_t for reply payload assembly and byte-swap
* do byte-swap of header fields explicitly instead of cryptic functions
* drop extra length computation
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This function was badly broken, among other things doing a
use after free, a NULL-pointer dereferrence, and not doing
what it was supposed to do.
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
Flushclient implicitly cast the result from a write-like call
to size_t, which turned it's error return value, -1,
into a large positive value, breaking the surrounding error-checking code.
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
Since FlushClient only writes the data in the client's buffer,
with no extra data or padding, we only write at most as many bytes
as there are in the client buffer.
we just have to set oco->count to notWritten, the amount we still
have to write.
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
Instead of using our own Use x_rpcbuf_t for assembling and
byte-swapping the reply payload.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Since nothing on WIN32 (mingw) targets is calling uname() anymore,
this compat code is now obsolete.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Simplify DefineSelf() by using xhostname(). Neither need to
care about OS specifics here, nor take care of zero-terminating the
hostname string - xhostname() is already doing this.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Now that we have the OS layer xhostname() wrapper, we don't need the
_XSERVTransGetHostname() function anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Simplify set_font_authorizations() by using xhostname(). Neither need to
care about OS specifics here, nor take care of zero-terminating the
hostname string - xhostname() is already doing this.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Simplify DoSubstitution() by using xhostname(). Neither need to
care about OS specifics here, nor take care of zero-terminating the
hostname string - xhostname() is already doing this.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Simplify vfbWriteXWDFileHeader() by using xhostname(). Neither need to
care about OS specifics here, nor take care of zero-terminating the
hostname string - xhostname() is already doing this.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This little helper does the OS specific part for gethostname()
calls. It's putting the result into a fixed-length struct and making
sure it's properly filled and the string is always zero-terminated.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Move down reply struct declaration/assignment and byte-swapping to
right before the write-out.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
* declare and assign in one shot
* move down reply struct declaration to where it's needed
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
* declare and assign them right where they're needed, in one shot
* drop unnecessary assignment and byte-swap of zero values
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
FlushClient() does drops the output buffer, when it becomes empty.
This previously wasn't any problem and actually intented, because we've
returned from the write path directly after FlushClient() was called.
But now we've changed the call order, so it's called from within
OutputBufferMakeRoom(), so we need to make sure we always got valid
output buffer before trying to access it.
For the future, we should think about whether dropping output buffers
really makes sense at all. Instead we could leave them as they are
(over the client's lifetime) and maybe just trim when they've went
too big.
Fixes: 1c13cfa6ca
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
