An alternative approach which may be better would be to create the
JavaScript context when the html_content is created, rather than
on demand.
This code checks for the JS context and creates one every time we
add a node to the DOM. So when JS is on, every doc with a single
node in it has a JS context. This seems to make on-demand creation
a redundant overhead.
Now we take the value of the javascript_enabled option when the
content is created. We then use the content's script_enabled
boolean everywhere else.
This prevents us getting inconsistent values for javascript_enabled
if a user toggles the setting while a page is loading.
It was read frequently during box construction, and also the
parser's script enabled setting could change where we handled
a change of encoding.
Now we only care about the setting of the javascript_enabled
option at time of html_content creation.
Default is to display the encoded version as this provides some security making phishing domains more obvious, and a lot of our frontends are unable to display the full range of UTF-8 characters on the status bar.
Displaying the decoded address in the URL bar requires frontends to be updated (GTK and Amiga done already), and the same caveats apply.
When the fetch of asynchronous javascript scripts completed the
completion of html rendering was not processed leaving the state
machine waiting forever.
Fixed point representation couldn't store 0.65 exactly, so avoid doing
the divide by 100 first.
I will look at moving this into libcss's fixed point header and doing
it in a way that avoids arithmetic overflow, but for now this fixes
el reg layout.
This changes the LOG macro to be varadic removing the need for all
callsites to have double bracketing and allows for future improvement
on how we use the logging macros.
The callsites were changed with coccinelle and the changes checked by
hand. Compile tested for several frontends but not all.
A formatting annotation has also been added which allows the compiler
to check the parameters and types passed to the logging.
This makes the box_move_xy function return a value on all code
paths. This was not really necessary as there is an assert in the path
that could have returned without a value. The gcc optimiser seems
unable to reason about this in this case causing a warning.
The ftell call in the html renderer handling of drag and drop was not
checking its return value for errors which could have resulted in
attempting to read -1 bytes.
coverity 1251038
The html content script handling needs to invalidate its JavaScript
context when the browsing context (browser_window) containing it is
either closed or the content fetch is aborted (stopped)
Previously the invalidation was only done on browser_window close
which resulted in use after free crashes because of the now invalid
JavaScript context.
When the browser window is about to be destroyed html_close gets
called indicating the browser_window previously associated with the
html content is about to become invalid.
This makes the javascript context invalid within the html content (it
is held by the browser window) when that context is about to be
destroyed. Previously the javascript children would continue to
attempt to use the context after it had been destroyed causing all
sorts of strange errors and failures.
