We broke handling of keydown/keyup when we moved the focus to the
canvas, as events from our input element would then no longer be
caught when they bubbled up to the document object (where we
previously caught events).
Restore the previous behaviour in a cleaner manner by creating a
second Keyboard object to handle this extra input variant.
Avoid the deprecated keypress event in favour of the keydown event.
It has the benefit of triggering for all keys, not just those that
produce symbols.
The focus can now move to the canvas so it is no longer a source of
confusion. It is also important to indicate that they have focus now
that we actually respect it.
Give the canvas proper focus handling. This avoids messy logic that
needs to disable and enable event handling when we want to interact
with other UI elements.
It also makes sure we can properly inhibit the browser from triggering
local actions on key presses.
During the last round of translations a few fuzzy markings was not
removed even though the string was updated. This can happen when the
translators edit the .po-file manually. Fuzzy markings result in that
the strings do not get included into the generated .json files.
The control bar can be dragged to the other side, this isn't obvious
however. This adds a hint on the opposite side in the form of a subtle
glowing half-ellipse.
If a user disconnects but leaves the browser tab open, the password
input field would still contain the password for future connections.
We now clear the input field after reading the password into memory.
It wasn't working properly anymore because it expected to be triggered
by the "load". But we now register that event listener long after the
"load" event has already fired.
Many browsers, for example Chrome on Android will not allow audio to
play unless it's initiated from a user action. It is not reasonable to
display an error for this. Fixes issue #821.
Setting a style to null does restore it in FF, Chrome, Safari and Edge.
But it does not work in Internet Explorer. The proper way to restore to
default values is to set it to the empty string. This works in all
browsers. Fixes issue #808.
This commit moves the global error handler into a separate file,
so that it can catch module loading errors.
This also adds support for properly displaying error messages with
newlines in them (since the module loader may throw those)
This commit restructures many of the ES6 modules, splitting them
up to actual export multiple functions instead of a single object.
It also splits up Util into multiple sub-modules, to make it easier
to maintain.
Finally, localisation is renamed to localization.
The only remaining user of WebUtil.load_scripts was for loading
localisation. Instead, we now load the localization information
over XHR as a JSON blob.
This removes the special comment part of the ES6 module syntax,
opting to enable ES6 module syntax by default.
It also appends `.js` to all import paths to better support in-browser
loading.
`app/ui.js` had an incorrect import path which caused issues
when using the ES6 and/or CommonJS builds of noVNC.
`core/util.js` had a non-strict-compatible declaration of a variable
without a `let` or `var` (it now uses `let`).
This fixes both issues.
It is not a launcher icon, but rather used for display on Google
Play. As such it is not needed and removing it mitigates the issue
of some buggy browsers downloading each and every listed icon.
Change the look of the "busy" spinner a bit. It's mostly used for
connection stuff, so give it a more data flow feel. Also bling it
up a bit with some fading. Perty sells. :)
The element we want scrolling around is noVNC_screen, not the entire
window. This also allows us to compute the screen size without
fiddling the scrollbars on and off.
The previous attempt could leave parts of the canvas outside the
document, making it impossible to reach. Use a safer method as
recommended by Mozilla.
Retire the old settingsApply. This also allows UI logic to check the
state of things using the settings instead of having to look at the
values of HTML elements (we couldn't be sure if the changes were
applied yet or not).
It stopped working when we switched to textContent as it relies
on being able to add new HTML elements. Do things properly by
adding new elements via createElement().
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.
We have enough layers now that we need to have some system for this.
E.g. make sure that dialogs during connect show up in front of the
blocking transition layer.
Pierre Ossman
Samuel Mannehed
Dominik Csapak
Giannis Kosmas
Mariusz Jamro
David Gnedt
Solly Ross
Alexander Neumann