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).
Supports server configurations that might prefer wss:// connections
on the default port, then proxies them through the web server to the
VNC server.
This proxy configuration is helpful on servers using self-signed
certificates. Accessing the https://host/vnc_auto.html page and
adding an exception for that host is sufficient to also satisfy the
wss://host/ request, unlike requests to wss://host:port/ which
may require an extra certificate exception.
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.
There is a specific event for when an image has finished loading,
so trigger on that rather than polling. The polling interval of
requestAnimationFrame() can also be very large.
The callers don't need to concern themselves with how images are
rendered, so hide the details behind the API. This also avoids
exposing the render queue.
The states 'loaded', 'failed' and 'fatal' were all variations of the
'disconnected' state. Removing these states allows us to get rid of
many ugly workarounds and special cases. Error messages to the UI can
now instead be delivered via a new onDisconnected callback.
Names such as 'disconnect' and 'disconnected' are inconsistent in the
way that one describes an action and the other a state. The state that
was called 'normal' didn't fit in with the others because the other
names describe a connection state. The new names are: 'disconnecting',
'connecting' and 'connected'
We already have a callback mechanism for this, so let's use that.
Adds an optional parameter 'msg' to the callback.
Fixes vnc_auto.html (#646) which was broken after
4e0c36dda7
The RFB protocol specifies a max version of 3.8, but RealVNC likes to
increment their version when they add new features. RealVNC 5.3 sends
"005.000" as the version string, since they've extended their
implementation upon RFB v3.8. We can detect that, and just tell them we
only speak 3.8, instead of barfing on "005.000" as an invalid version.
This commit moves all the input-related files from `core/`
to `core/input/`, and renames a couple as relevant
(input.js --> input/devices.js, keyboard.js --> input/util.js).
The event-related wrapper functions in Util existed mainly for
backwards-compat. However, all currently supported browsers
support the standard functions, so these wrappers are no longer needed.
This commits prevents Util from modifying the window object.
- `window.requestAnimFrame` was removed (no polyfill is needed anymore)
- the potential redefinition of `console.log` and friends was removed
(all supported browsers have `console.xyz` defined anyway)
This commit switches over to use PhantomJS 2.x, bringing in a whole host
of improvements (including `Function#bind`, so we can remove the
`Function#bind` shim in core/util.js).
This commit removes our modification of the Array prototype.
It wasn't actually used much in the main code, anyway, and it's a
bad practice to modify built-in prototypes.
This commit restructures noVNC, splitting it into the core directory
and the app directory, with the former containing core noVNC parts,
and the latter containing parts specific to the application.
In input.js, a new keyboard handler was added to deal exclusively
with the QEMU key event extension. '_onKeyPress()' signature
was changed to allow the same method to treat both cases.
The extension will only be enabled if the browser has support
for the KeyboardEvent.code property.
Changes in rfb.js:
- added a new extension code, QEMUExtendedKeyEvent, value -258.
- handleKeyPress now receives 'keyevent' instead of 'keysym' and
'down'. Both values are retrieved from keyevent as they were
in the previous signature. This method now can send QEMU RFB
extended key messages if the flag was set to 'true'.
- tests/test.rfb.js were changed folowing the onKeyPress() signature
change.
- added a new function to send the QEMU extended key message.
Signed-off-by: Daniel Henrique Barboza <danielhb@linux.vnet.ibm.com>
This new file contains the XT scancode mapping that
the extension will use in rfb.js file.
Signed-off-by: Daniel Henrique Barboza <danielhb@linux.vnet.ibm.com>
Instead of requesting frame buffer updates we can, if the server
supports it, continuously recieve frame buffer updates at a rate
determined by the server.
The server can use fencing messages and measure response times to
determine how often it will continue to send updates.
Try to avoid using helper functions with complex logic when verifying
results as those helper functions are also something we want to verify.
Also add a test for a mix of clean and dirty areas specifically to make
sure that helper function behaves properly.
Make sure our messages go away right away, rather than having to
remember to call flush from the caller, or causing extra delays by
waiting for the send timer. This should result in a more responsive
system.
We were completely mishandling the length of the data. Make sure
we look at the length of the websocket rather than the websock object,
and also compare with the expected length.
In order to follow the surrounding coding-standards, the
setDesktopSize client message is split from the public function which
now is called requestDesktopSize().
This commit updates vnc_perf.html so that it works with the current
version of noVNC. It also introduces a utility to convert noVNC session
recordings recorded in base64-mode to binary-mode recordings, since noVNC
no longer supports base64-mode.
Fixes#479.
There was a bug caused by 38781d931e
which prevented color map look-ups sent by rfb.js from working properly,
since display.js expected a single-item array, and rfb.js sent just them
item value itself (a number) instead. This fixes that, and tweaks the
corresponding test to match that behavior.
This `displayed` assertion had a bug that was causing it to not
actually check anything (it was using obj.length instead of
data_cl.length). This fixes that.
This commit adds the '--debugger <port>' option, which triggers
the PhantomJS remote debugger. The initial output of the terminal
when running the debugger gives more information on how to use
it.
This commit converts the send queue to use typed arrays, and converts
message creation functions in 'rfb.js' to create messages directly into
the socket's send queue. This commit also removes the separate mouse array,
which is no longer needed.
**This commit removes Base64 (and Flash) support**
This commit converts websock.js to used Typed Arrays for the
receive queue (and tweaks rfb.js to ensure that it continues
to function, since only Firefox implements
`%TypedArray%.prototype.slice`). Base64 support was removed
to simplify code paths, and pave the way for using Typed Arrays
for the send queue as well.
This provides two advantages: first, we allocate a buffer ahead
of time, meaning the browser doesn't have to do any work dynamically
increasing the receive queue size. Secondly, we are now able to pass
around Typed Array Views (e.g. `Uint8Array`), which are lightweight, and
don't involve copying.
The downside is that we initially allocate more memory -- we currently
start out with 4 MiB, and then automatically double when it looks like
the amount unused is getting to small.
The commit also explicitly adds a check to the compacting logic that
avoids calling the copy functions if `_rQlen === _rQi`.
In e543525faa, we switched to creating
a new RFB object on disconnect. This caused issues, however, since
any errors were only displayed briefly before the new "loaded" text
was displayed instead.
Now, we create the RFB object on connect. This essentially removes
the usefulness of the "loaded" state, but prevents the aforementioned
problem.
To facilitate this, the code which does detection of cursor URI support
was moved from this Display constructor (which now calls the new
function) into its own function, `Util.browserSupportsCursorURIs()`.
Fixes#467
* Renamed and reworked fbuClip to clippingDisplay
* Added tests for clippingDisplay
* Use the a noVNC_container which covers the entire page to get the full size
(Fixes#463)
* Added maxWidth and maxHeight to the canvas which can limit the viewport size
* Only show either the canvas or the logo, hide one when the other is shown
* Always center the canvas (previously it was only centered when not clipping)
* Removed iOS specific "position-fixed" fixes and start calling setBarPosition
on every resize
* Removed the noVNC_screen_pad
* Added better error handling in ExtendedDesktopSize
* Added helper function to share code with DesktopSize
* Update test.rfb.js to only check for error handling if we were the
ones requesting the resize
This commit adds two new addition scaling options. Both options do
local scaling. The first "Local Scaling", does both upscaling and
downscaling. The second option, "Local Downscaling", only downscales.
This is based on work by @mightypenguin (with an additional bug
reported by @glazik12).
* Split viewportChange into two functions, one for changing size and the other for changing position.
* Modified viewport code to be capable of changing to a bigger size in the context of a
client-initiated resize.
* Made clearer distinctions between when viewport-clipping or not.
* Added public function for telling when viewport-clipping.
* Updated tests that were using viewportChange.
The HEXTILE test which tested for a background tile followed
by an empty tile was only wide enough to actually test for one
tile, thus not actually testing the functionality. It now actually
uses two tiles, thus actually testing the functionality.
Previously, if a HEXTILE tiles was received with a subencoding
of 0x00, it would draw a rectangle using data from the render
queue, which would result in random colored blocks when using
the HEXTILE encoding. This is the result of a miscopy during
the refactoring. It now has the correct functionality according
to the RFB protocol specification, which is to draw a rectangle
with the last set background color.
Closes#411
Commit 795fca23dc changed the default
size from 640 to 240. This broke a couple tests which depended on
the default size being 640. Those tests have now been fixed.
This patch adds support for using relative paths
with the '-r' or '--relative' methods. This can
be useful if you want to output HTML (with the
'--output-html' option) and use it in a webpage.
Additionally, the '-o' was removed from the documentation
of '--output-html', since it hasn't worked for that in a
while ('-o' means open in browser instead).
Because we use the XOR (`^`) operator, the button mask must be
set before a MouseUp event happens, otherwise we'll send a pointer
event like it was a MouseDown event. The button mask was not set
in one of the tests, so the test was failing.
There was a bug in cursor URI support detection due to the way
set_defaults now works -- the code was checking for `null`, whereas
when not set, options default to `undefined` unless otherwise
specified. The code now checks for either `null` or `undefined`.
Tests have been added to ensure that this works properly.
There was a typo in one of the instances of the _buttonMask field
(it was written as _buttonMaks), causing MouseUp to never be sent.
This has been rectified, and the unit tests for the mouse handler
have been changed to check for explicitly sending mouseup and
mousedown.
Fixes#393
When run via karma, all the tests are loaded into the same page.
This was causing a collision in the 'displayed' assertion dealing
with using viewportLoc.
The assertions are now in their own file, pulled in by tests that
need them. Additionally, several tests which only set fb_width
and fb_height were correct to set viewportLoc as well.
Closes#392
Also-Authored-By: Martin André (github: mandre)
This adds support for Travis CI and SauceLabs
testing. Testing on SauceLabs in done via
the Karma test runner. Note that encrypted
Sauce username and access key values need
to be inserted into .travis.yml as global
environment variables. Additionally, the
local test runner (which is still useful
for debugging tests and code) was updated
to reflect that the 'node_modules' folder
now gets placed in the root directory.
File: display.js
Tests Added: True (preliminary)
Changes:
- De-crockford-ified the file
NOTE: the tests included for display.js cover basic functionality, but
are by no means nearly as comprehensive as the ones presented for
rfb.js.
File: websock.js
Tests Added: True
Changes:
- Cleaned up JSHint errors
- Converted to normal JS constructor pattern with "private" fields and
methods now simply being prepended by underscores
- Added a "bind" polyfill for use in PhantomJS 1.x in util.js
- Added FakeWebSocket to fill in for actual WebSocket objects when
testing
- Made exception handler actually log exception name and message,
to console, in addition to stack trace
File: webutil.js
Tests Added: False
Changes:
- Fixed JSHint Errors (global "use strict", spaces)
- added some newline characters when appropriate for readability
- moved variable declarations to the places they were actually used
for readability
File: util.js
Tests Added: True (partial -- for logging and array push methods)
Changes:
- Fixed JSHint Errors (indentation, semicolons, global "use strict")
- Made browser detection methods more readable
- added some newline characters when appropriate for readability
- throw Errors not strings!
- Removed conf_defaults, and added make_properties and set_defaults
instead (see below)
The removal of conf_defaults and switch to make_properties and
set_defaults is to facilitate the switch over to normal Javascript
constructors instead of Crockford-style constructors. Now, methods
are added to the objects prototype (and thus make properties is called
outside the constructor).
This is the first commit in a series of commits
which improve the readability of some of the code
and add tests.
File: base64.js
Tests Added: True
Changes:
- Improved indentation
- Fixed JSHint errors
- Moved loop variables to be declared in the loop for better readability
(N.B. Javascript does not have block scoping, so the variables are
still technically available outside the loop -- it just makes the code
clearer to place them inside the loop, since they are only used there)
Previously, if you did not specify a tests file,
you had to be in the 'tests' directory for the
"run all tests" functionality to work. Now it
will work in any directory.
Previously, there would be a case where if your tests took
too long to run, the casper test runner would only report
on certain tests. This has been fixed.
Now, 'error' events from the test runner are output to stderr.
Additionally, when debug is enabled, debug output is logged to
stderr instead of stdout (as was the case previously).
Now, the phrase `requires test modules: ` may be place in a comment
in a file to require modules local to the test directory, similarly
to the way the `require local modules: ` line may be used to inject
files in the 'include' directory. This is useful for when common
fakes need to be injected into a test.
When using the '-g' option with run_from_console.js, you can
now pass the '-o' option to automatically open the generated
HTML file in your default browser. This relies on the 'open'
NPM module.
When shortcut modifiers (modifier keys such as CTRL, which do not participate in
composing character input) are pressed, we try to suppress the keypress
event, as browsers do not reliably generate it. This means that
subsequent key events are decoded only based on the keydown event.
Due to a type error (comparing a string to a number), shift was
mistakenly treated as a shortcut modifier, preventing text input which
relied on shift, such as _ and %, from being generated.
If the files passed to the '-t' option are all '.js' files (or
the 'run all tests' option is used) and the '-i' option is not
passed, all tests will be search for the string
'require local modules: '. Only the first instance of this string
will be used. Following the colon should be a list of either local
modules (i.e. files in the '../include/' folder relative to the
test runner's directory, without the '.js' extension) or paths
to other Javascript files. The list of modules and/or files should
be comma-separated. These files will then be included in the generated
HTML file for the appropriate tests as if the '-i' option had been used.
Now, if the '-t' option is passed but no tests are listed,
all tests in the same directory as the launcher will be run.
A file is considered a test if it matches the RegEx
/^test\.(\w|\.|-)+\.js$/ (for those who cannot read PCRE,
that's roughly 'test.*.js').
The test runner now will not break when Mocha skips tests,
and will properly report them. Additionally, several JSHint
warnings were fixed, and a `--debug` option was added to see
output from the provider.
This commit introduces two flags, '-g' and '-o' to
the `run_from_console.js`. Both flags do not run
the tests. Instead, deal with the autogenerated
HTML. The former outputs the paths to the autogenerated
HTML temp files, and then pauses the program until Ctrl-C
is pressed (or SIGINT is sent). The latter outputs the
generated HTML for each files to STDIN with the names
of the tests to which they belong.
Previously, the only way to run the Mocha tests
(in 'test.*.js') is to write a web page to wrap
them (or use a provided one), and then load that
file in a browser.
This commit introduces a series of files to allow
you to run the Mocha tests from the command line
instead.
Normally, Mocha tests can be run from
the command line anyway. However, since this
project was designed to work in web browsers
and not node, the code doesn't contain the
proper `require` calls, nor does it contain the
proper `module.exports` declarations. Additionally,
some of the code is dependent on having a browser
environment.
To overcome these issues, a headless browser environment
is used. The command file introduced in the commit,
`run_from_console.js`, can use one of two environments:
ZombieJS, a pure-javascript headless browser simulator, or
SpookyJS/CasperJS/PhantomJS, an actually WebKit-based
environment.
Because the environment-dependent code is separated
out in to different files ('run_from_console.zombie.js'
and 'run_from_console.casper.js'), the program can be
safely used if only one of the supported environments
is installed.
Additionally, the command will automatically generate
HTML and inject the required tests if there is no
pre-existing HTML file (although you can still use
pre-existing HTML files if you want to).
The required NPM modules for the base program are:
- commander
- ansi
- mocha (must be installed locally for the HTML files to use)
- chai (must be installed locally for the HTML files to use)
- temp
For Zombie, you need:
- zombie
- q
For Casper, you need:
- casperjs (must be installed locally in order to work properly)
- phantomjs
- phantom
- spooky
The command itself can be invoked as
$ node run_from_console.js -t html_files
or
$ node run_from_console.js -t js_test_files -i js_required_files
In both cases, the 'files' options should be a comma-separated list of
files. The first case runs pre-existing HTML files. The second case
generates HTML files to run the specified Mocha tests, and injects
the requirements specified as well.
Additionally, there are extra arguments that apply to both forms:
'-a' can be used to print all test results, not just the failures,
'-c' may be used to force color to be enabled (when outputting to
a pipe, such as when `less -R` is in use), and '-e' is used to
set the environment. Use the '-h' or '--help' options to see
a detailed description of all options, and their long-form versions.
Related to issue/pulls:
https://github.com/kanaka/noVNC/issues/194https://github.com/kanaka/noVNC/pull/201https://github.com/kanaka/noVNC/pull/202
In IE9, the window.onload event can fire before dynamically loaded
scripts have finished loading. This can result in either WebSocket (in
the case of vnc_auto.html) or RFB (in the case of vnc.html) not being
defined at the point when window.onload is called.
- Move the load_scripts routine from vnc.js to util.js (so that
websockify can use it too). Also, refactor to work when load_scripts
is called by a script that itself uses load_scripts. When the whole
chain of dynamically loaded scripts is finished then call
window.onscriptsload. Use this mechanism in all the places that
depend on dynamic loading of scripts: vnc.html, vnc_auto.html,
websock.js, tests/vnc_playback.html, and tests/vnc_perf.html.
- Use the new window.onscriptsload handler instead of window.onload.
- Remove include/start.js and do the script loading and startup event
handling in include/ui.js instead.
The MPL 2.0 license is a "file-level" copyleft license vs the
"project-level" nature of the L/GPL. The intention of noVNC has
always been that it should be easy to incorporate into existing
projects and sites whether free/open or proprietary/commercial. The MPL
2.0 is designed for this sort of combination project but still
requires that any distributed modifications to noVNC source files must
also be published under the same license.
In addition, the MPL 2.0 allows the code to be used in L/GPL projects
(the secondary license clause). This means that any projects that are
already incorporating noVNC should not be impacted by this change and
in fact it should clarify the licensing situation (the exact
application of the L/GPL to web applications and interpreted code is
somewhat ambiguous).
The HTML, CSS, image and font files continue to be under more
permissive licenses (see LICENSE.txt). The included websockify python
code remains under a LGPLv3 license although the include/websock.js
file from the websockify component is now under MPL 2.0 as well.
Permission was received from other noVNC authors to make this change to their
code license on the following dates:
- Chris Gordon (UI): Jun 24, 2012
- Antoine Mercadal (DOM,*util.js): Oct 10, 2012
- William Lightning (UltraVNC repeater): Oct 10, 2012
- Mike Tinglof (tight encoding): Oct 15, 2012
Conflicts:
include/display.js
include/rfb.js
This merges in the fix for https://github.com/kanaka/noVNC/issues/70
This changes noVNC to use the preferred color ordering that most VNC
server prefer and that VMWare VNC requires. It's possible this may
break some VNC servers out there in which case we might have to do
something a bit more subtle such as having alternate render functions
for little and big endian color ordering.
API changes (forward compatible):
- Display: add 'viewport' conf option to turn on and off viewport
mode.
- RFB: add 'viewportDrag' option to enable/disable viewport dragging
mode.
Other:
- Add clip mode setting to default UI. For touch devices, clipping is
forced on.
- Use CSS media queries to adjust visual elements based on screen
size. Especially disconnected logo size/position and button text size.
- Catch page unload while connected and give a confirm dialog.
- Change mouse button selector to a single button that changes between
' ', 'L', 'M', 'R' when clicked (empty means mouse is just being
moved and doesn't send clicks).
- include/ui.js:setViewClip() routine sets the clipping of the
viewport to the current size of the viewport area (if clipping is
enabled).
- include/ui.js:setViewDrag() toggles/enables/disables viewport
dragging mode.
- Add several images for the UI and for Apple devices:
- images/clipboard.png: clipboard menu icon
- images/connect.png: connect menu icon
- images/disconnect.png: disconnect button icon
- images/keyboard.png: show keyboard button
- images/move.png: viewport drag/move toggle button
- images/settings.png: settings menu icon
- images/screen_320x460.png: iOS app/desktop link start image
- images/screen_57x57.png: iOS app icon
- images/screen_700x700.png: full size noVNC image
Samuel Mannehed
Pierre Ossman
Solly Ross
Kyle Russell
Pierre Ossman
Lars Ole Hurlen
Daniel Henrique Barboza
Drew DeVault
Jesper Dam (jalfd)
Jacob Swanner
Ramon de Klein
Solly
Martin André
Jesper Dam
Joel Martin