It's a reoccurring issue that users try to build individual components
by pointing cmake at a specific subdirectory, e.g. 'cmake vncviewer'.
CMake, unfortunately, has insufficient protection against this so we'll
need to add a manual check.
This commit only adds it to the most likely places for misuse so we
don't have to pollute every CMakeLists.txt.
This is mainly a copy of XKeysymToString() from libX11. We've also added
a wrapper that still gives a string, even if there is no name for the
requested keysym.
This grows the binaries a bit, but not with any extreme amount so is
hopefully worth it to get better debug logging.
We can get races with clipboard managers in the server that is very
confusing to the user.
When the clipboard changes locally, we tell the server to drop the old
clipboard (as it is now lost). But we don't send over the new clipboard
until we get focus again, in order to not leak more data than necessary.
This causes some clibpoard managers to take over ownership in order to
avoid an empty clipboard. And this takes precedence over the new client
clipboard as it happened later. Effectively reverting the clipboard the
user sees.
Avoid all of this by simply ignoring the server when we don't have
focus. This is likely what users expect anyway as they expect their
currently focused application to control the clipboard, not vncviewer in
the background.
Don't assume a lack of TCP listeners means the server will be
unreachable. There might be other methods of access, so let the higher
levels do that sanity check instead.
FLTK only allows 256 different box types, but it doesn't actually check
this when registering new ones.
Move our custom types to a valid range, and add an assert for good
measure to make sure we don't overflow FLTK's internal structures.
There is something broken with these FLTK draw routines on Windows. They
leave gaps at the start and end of the arc/pie rather than filling the
whole specified span. So we need to nudge the numbers a bit to work
around this.
Inspired by modern Windows appearance, and to some extent macOS. They
have flat boxes and use white, or very light, colors for interactive
elements. Unfortunately we can't directly control the colors of
widgets, so instead we just lighten everything that uses this box type.
GNOME uses a different design, both their older and newer style. But UI
look is less consistent on Linux, so hopefully our new look is decent
enough there as well.
gcc doesn't support -Wformat for the wide format versions of printf()
and friends yet:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=38308
Do what glibc does and have some commented out tags to show future
intent.
gettext replaces all *printf() functions on platforms that don't fully
conform to the POSIX behaviour. Unfortunately, gettext fails to tag
these replacement functions properly so that -Wformat can still do its
thing.
Resolve this by adding a redudant declaration of the relevant functions,
with the attribute tagging in place.
The size of size_t depends on the architecture, so we need to have
different conversion to and from strings. But we don't really need that
range, so avoid the issue by using a standard integer size.
We should handle this in the low-level protocol code as much as possible
to avoid mistakes. This way the rest of the code can assume that strings
are always UTF-8 with \n line endings.
The coordinates we get are relative the root window of each screen, so
we can only trust them if we are on the same screen. So let's explicitly
check that we are still getting events from the expected screen by
checking the root window field of the event.
These files don't use anything from this header, so remove the include.
This exposes some missing includes in other places, though. So add an
explicit include in the files that were relying on an indirect
inclusion.
We mostly use classical C strings, but the memory management around them
can get confusing and error prone. Let's use std::string for the cases
where we need to return a newly allocated string.
This mimics how some system functions (like inet_ntop()) work, and
avoids complexity around ownership of the returned string buffer.
The downside is that the string must be consumed directly as it will be
overwritten on the next call, but that is not an issue with the current
usage.