| Age | Commit message (Collapse) | Author |
|---|
|
This fixes some UI component initialization issues. It's a long
story...
The problem.
- Go to this page: http://www.thejournal.ie/seanad-election-results-2016-2737768-Apr2016/
- Click one of the links from the "Most Popular" box on the right.
- Navigate back (`H`) and, as the original page is loading, activate the Vomnibar.
In 1.54 this renders Vimium unusable. In `master` this renders the
Vomnibar unsable, but the rest of Vimium usable.
It seems the source of the issue is that we're initializing UI
components too soon. We need to wait until the `readyState` is
"complete".
With this PR:
- The Vomnibar is initialised when the `readyState` is "complete" (in
the top frame only, and only if Vimium is enabled). Requests arriving
prior to then are silently discarded.
- The HUD is also initialized only when the `readyState` is "complete";
however, requests arriving before then are queued. So, if the user
immediately enters insert mode, then the "Insert mode" indicator will
eventually be displayed.
- The help dialog silently discards requests until the `readyState` is
"complete.
I'm posting this as a PR because:
1. It needs some visibility.
2. With this, if the `readyState` *never* reaches "complete", then the
Vomnibar would be unusable. And that's pretty serious.
|
|
Previously, you could blur the help dialog on the options page (to read
and type command names).
With the reinstated help-dialog styling (whereby the help-dialog iframe
extends across the entire window), this doesn't make sense. So it's
removed.
|
|
Previously, we had two different approaches. This seems like a simpler
approach.
We simply cache the Vimium CSS in chrome.storage.local (in the
background page) and fetch it from there (in UI components).
There is also a minor change in the we no longer cache the CSS in
memory. This seems to be the right thing to do. Caching the CSS in
memory consumes a small amount of memory. However, it can only speed up
the fastest loads. It cannot speed up the first load -- which is likely
the one that matters most. So caching the CSS in memory seems to make
little sense.
|
|
This reverts commit 0a49cc45732175c65651b5f054c677d6c42a28c0.
|
|
Previously, we had two different approaches. This seems like a simpler
approach.
We simply cache the Vimium CSS in chrome.storage.local (in the
background page) and fetch it from there (in UI components).
There is also a minor change in the we no longer cache the CSS in
memory. This seems to be the right thing to do. Caching the CSS in
memory consumes a small amount of memory. However, it can only speed up
the fastest loads. It cannot speed up the first load -- which is likely
the one that matters most. So caching the CSS in memory seems to make
little sense.
|
|
Previously, requests (like opening the Vomnibar) would be silently
discarded if they arrive before the document is ready. Here, we queue
them instead.
|
|
This replaces c01d7eea8675f9a7d84999777e8de72244d687b6.
All UI components require the document to be ready. So, here the
constructor waits for DomUtils.documentReady(). The diff looks big, but
mainly it's a result of changes in indentation in the constructor and in
hide().
Also, hide() now uses @postMessage() to post a null message. This
forces hide to use @iframePort.use(), which ensures that hide()s cannot
overtake activate()s.
This continues #2100.
|
|
|
|
|
|
|
|
It interferes with the HUD's timing.
|
|
The code to handle the focus for UI components has been tweaked and
adapted over time, and has become quite complicated (and brittle). This
reworks it from scratch, and co-locates similar code which does related
things.
Fixes #2099.
|
|
In testing global link hints, it seems that UI components can be opened
before they're actually ready, and Vimium hangs with a broken UI
component displayed. This issue seems to be in 1.54 too.
It's not clear what's causing this -- and it's hard to reproduce
systematically. However, it only seems to happen on navigation.
Therefore, it seems likely that there is an issue with the
initialization sequence.
Here, we wait for:
- the DOM content to be ready, and
- the port to be provided
before registering the UI component as ready.
|
|
1. Do not initialize UI components until the DOm is ready.
2. Do not register UI components as ready (specifically the HUD) until
the full initialization sequence is complete.
This goes some way towards fixing the issue described in this comment:
https://github.com/philc/vimium/issues/2081#issuecomment-205758102
It relates to link-hints mode hanging when launched during a navigation.
This problem exists in 1.54, but emerged during testing of global link
hints.
"Some way toeards fixing..." because it is still possible to trigger
issues.
|
|
In the previous commits, I omitted to actually check that the help
dialog wasn't closing on the options page.
I should be good now.
|
|
The help-dialog UI component was ignoring requests to "hide" when other
frames are focused (because previously it covered the whole screen, and
no other frame could get the focus).
With f0911e52f0e71c6d2539bdc74a09ff2dbd5ab125, the help dialog no longer
covers the whole screen, so it must listen for and react to
"frameFocused" events.
However, the help dialog should not "hide" when the frame that is
focused is itself! This required a little extra plumbing. That
plumbing is helpful, though, because it allows individual UI components
to decide what to do when another frame receives the focus (as opposed
to the previious version, which simply unilaterally sent a "hide"
message).
|
|
Problem:
- `?`, `o`, `Esc`, `Esc` leaves the focus in the (invisible) help dialog
frame, rendering Vimium broken.
Since we do
@iframeElement.focus()
when we activate a UI component, here we do
@iframeElement.blur()
when hiding such elements.
Fixes #2038.
|
|
|
|
|
|
|
|
|
|
This fixes #1817, where our stylesheet isn't loaded correctly due to a
Chromium issue and the Vomnibar/HUD <iframe>s are always visible on the
new tab page.
|
|
|
|
This implements @gdh1995's idea from #1796.
|
|
|
|
If the page is an XML document, nothing we do works:
* <style> elements show their contents inline,
* <iframe> elements don't load any content,
* document.createElement generates elements that
- have Element.style == null, and
- ignore CSS.
This commit stops us from injecting anything into the DOM from
UIComponent, fixing #1640.
|
|
|
|
I haven't seen this happen, but...
It could be possible for the iframe's contents to load before this
callback is called (on nextTick), in which case the "load" callback
wouldn't be called. So we delay setting the iframe's source until
nextTick has elapsed.
|
|
|
|
Approach: Re-use the existing AsynDataFetcher class to "fetch" and use
the iframe message port. Messages are queued until the iframe's
contents have loaded and the message port is open.
Fixes #1679.
|
|
See #1671.
The problem was that we were making the vomnibar visible in the frontend,
even if the vomnibar initialisation was not yet complete, and the iframe
port was not yet available.
(Also, @activate() is never being called without options, so we can drop
that test.)
Fixes #1671.
|
|
|
|
|
|
|
|
Shadow DOM doesn't support <link>s, so we have to load the stylesheet
manually and inject it into a <style> element.
|
|
This insulates them from page CSS, so we don't have to compete to style
elements correctly.
|
|
|
|
The UX around this is not quite right yet. It's better to disable it
for now.
|
|
|
|
1. Rework event handling to eliminate frame flicker (a la #1485).
2. Tidy up logic. Which should make this more robust.
|
|
Clean up, and fixes following code review from @mrmr1993.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This changes vomnibar commands to activate not in the current frame, but
in window.top. Consequently, the vomnibar always appears in the same
position, and we don't get odd looking vomnibars in small frames.
Apart from the better UX, this seems to be the right thing to do.
Vomnibar commands apply to tabs (not frames).
Currently incomplete:
- On exit, the focus is not returned to the frame which originally
had the focus. (It's returned to window.top).
- The vomnibar can lose the focus and not hide itself for various
frame/click combinations.
|
|
Fix oversight from #1517.
|
|
Fixes #1506.
This takes the opposite approach to #1511. Instead of hiding the
vomnibar when it blurs, we hide it when it's host frame is focused.
|
