It doesn't need to be this general as the issue is mostly about
Windows. Also use the same modifier shuffle that RealVNC and
TigerVNC uses to get macOS working well.
The fields provided cannot tell us if it is the left or right
version of the key that's pressed, so they are inherently unreliable.
It is also not a huge problem in practice as we'll get in sync on
the next press or release of the modifier.
Look up keys that are independent of layout and state first,
followed by keys that are only mild variations in layouts.
This is more robust as there might be multiple physical keys
generating the same symbols, and Keysyms don't map directly to
Unicode in all cases.
At the same time switch over to using the modern, standardised
'code' field for lookup.
Use the more modern 'key' field, and remove some legacy fallbacks
that are no longer required. This also removes the "stall" mechanism
as it is not needed with current browsers.
There is a drag-threshold in the panning code which the tests didn't
account for. This caused the tests to fail when window.devicePixelRatio
was higher than 1.
This commit makes vnc_playback.html functional once more, and completely
refactors tests/playback.js to make it usable in other scenarios.
In order for vnc_playback.js to properly load playback files now, they
must `export` their variables.
This updates the tests to work with the new structure, and removes the
old `utils/run_from_console.js` files in favor of just using Karma
directly. The Karma debug page now displays the normal mocha HTML, so
we can use that instead of the HTML generation functionality of the old
test runner.
Note that PhantomJS does not work at the moment (PhantomJS 1.5 should
make it possible to test on PhantomJS again).
Previously, setting `innerHTML` was used to display the statuses. These
could include content communicated from the remote VNC server, allowing
the remove VNC server to inject HTML into the noVNC page.
This commit switches all uses of `innerHTML` to use `textContent`, which
is not vulnerable to the HTML injection.
Do all rendering to a hidden canvas and then copy over the finished
frame to the visible canvas once everything is done. This simplifies
things and solves some bugs as we can retain a copy of the entire
frame buffer.
setTimeout() is subject to delays, possible massive ones. As such it
is rather useless for performance sensitive code. Use the non-standard
setImmediate() API instead, emulating it on postMessage() when it
isn't available.
The hacks needed to run these tests require proper handling of
properties. Unfortunately IE and old versions of Chrome mess up,
so just skip the tests there.
We have to temporarily replace window.navigator due to the fact that
its property languages is read-only. The tests for the translation
utilities require different values of this property. The failing tests
were added in merge of PR #718 (in commit
f5bf2d84ef).
RFB's _fail function logs the error, disconnects the session and sets
disconnect_reason. The disconnect_reason is upon disconnection sent to
the user interface. It is thus not suitable for including error details
that aren't user friendly. The idea is that you will look in the
browser console for a full log with details of the error.
Node.js doesn't handle characters high up in the unicode range
properly with charCodeAt(). Adding a new test for UCS-4 codepoints
using codePointAt() to cover this.
