| Age | Commit message (Collapse) | Author |
|---|
|
Previously, the key-handling logic (keyQueue, etc) was and the backend
whereas passKeys were handled in the content scripts - so they were a
long way apart.
Now that they're in the same place, it makes more sense to integrate
passKey handling into the regular key handling, because they depend upon
the same data structures.
|
|
... and fix two bugs:
- not suppressing keyup event after keyChar matched in keydown.
- we cannot check the passKeys keyChar in keyup because the key state
has changed; so we track what the next keyup response should be.
|
|
The styles guide says not to use standalone `@`. So this changes the
occurrences I could find (with sed) to `this`. Occurrences within files
with major outstanding PRs are omitted.
|
|
|
|
PhantomJS (2.1.1)".
URL: https://github.com/philc/vimium/issues/1944
|
|
|
|
|
|
Certain background-page utilities are actually shared, and are therefore
best placed in place that reflects that.
Here, tabRecency is moved to the new gb_utils.coffee in preparation for
implementing a go-to-previous-tab command. In particular, it is no
longer appropriate that tabRecency be embedded within the completion
code.
logMessage() from main.coffee is also a candidate for moving to
bg_utils.coffee.
|
|
This tweaks @mrmr1993's logging ideas discussed in #1629:
- Do not generate log entries from the logging page itself.
- Use the logger for *all* logging (including from `commands.coffee`, and from `bgLog`).
@mrmr1993's original idea is 91fb337c9d92f6291ef42c55c4d29ca35b710203.
|
|
|
|
Previously, we blocked keyboard events for a fixed 200ms.
With this PR, we continue blocking keyboard events until 150ms after the
last `keydown` or `keypress` event. So, we wait until we think the user
has stopped typing.
Fixes #1842.
|
|
|
|
|
|
Don't decode javascript URIs on Chrome 46.0.2467.2 and later
|
|
JavaScript's sort function is not stable; this PR makes the sort used for filtered link hints stable.
There are two reasons for doing this:
- High-scoring hints are more likely to keep the same hint string as the user continues typing. (Currently, the hints assigned change based on the vaguaries of the non-stable sort.)
- For equal-scoring hints, we retain the visit-child-before-parent ordering (which is used to NOT match a parent's text if we have already matched that text in a child).
And, as a result of all of that, the UX is more predictable and hence better.
|
|
The fix for Chromium issue 483000 landed in this version, so it is no
longer necessary for #1611
|
|
|
|
|
|
This prevents issues like #1731 and is an (better) alternative to #1732.
|
|
(@mrmr1993: This is yet another approach to the Settings problem.)
With the new Settings implemetation, settings which have a non-default
value and which are not in synced storage (that is, they have not been
changed since synced storage was introduced) are not currently
accessible to content scripts. This commit makes such settings
accessible via chrome.storage.local.
Important:
- There's a change to the established settings data model here.
Previously, settings with default values were not stored; here, they
are. This eliminates a considerable amount logic from Settings, but
means that migrations will be required if default values are changed
in future. (Other than type changes, have we ever changed a default
value?)
- There's also a change (bug fix?) to the behaviour when an affected
setting is reset to its default value. Previously, the change would
*not* have been synced (whereas all other changes are). Here, such
changes *are* synced. The previous behaviour was inconsistent with
the syncing behaviour of all other options changes.
Note:
- This isn't particularly well tested. It's being committed mainly just
for consideration of the approach, initially.
|
|
|
|
(Not sure when this crept in.)
We need to ensure that we always generate the same hint strings for the
same filter state. Here, we do this by always using the same mechanism
(@filterLinkHints) to set the hint strings.
|
|
Functionality wise, this is a no-op. However, since we now create a new
LinkHintsMode object on every "f", we don't need to be so careful about
resetting the state on exit.
|
|
|
|
LinkHints.init() isn't doing anything.
|
|
This is a minor re-working of #1705 from @mrmr1993.
The main changes are:
- Simplify initialisation logic.
- Always initialise Settings immediately and automatically (ie. don't
initialise Settings separately and manually in the background,
content scripts, options and tests).
- Get rid of addEventListener (it's only being used for on "load").
- Add Settings.use() in its place.
|
|
|
|
Direct keyboard access to custom-search engines via keyword flag
|
|
For ...
map s Vomnibar.activate keyword=g
... we verify that "g" is indeed a custom search-engine keyword before
setting it. If it is not, we output a console.log message and launch a
vanilla vomnibar. (An alternative would be to bail.)
|
|
The Settings object used by the background page now uses 1 of 3 caches,
depending on the context it is available in:
* localStorage - in the background page
* a copy of localStorage - in non-background extension pages
(options.html, popup.html, etc.)
* an empty object - in all other pages (where localStorage doesn't point
to the extension's localStorage object).
For any extension page which is *not* the background page, a copy of
localStorage is used instead of true localStorage:
* Once localStorage is updated by one background page, the others can
only see the updated copy.
- Pages with an updated cache can't tell which changes are new, and so
don't know which postUpdateHooks to run.
* By copying localStorage's contents into a new object, extension pages
can still access settings synchronously.
- This is especially important to options.html and popup.html; they
will not work without it.
|
|
|
|
|
|
This completely decouples settings.coffee from all other background
source files, so that it can (eventually) also be used in the frontend.
|
|
This stops Sync from being referred to from anywhere except
settings.coffee and settings_test.coffee.
|
|
|
|
completion-on-custom-search-only-merge
Conflicts:
background_scripts/completion.coffee
|
|
Fix for #1554.
|
|
|
|
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.
|
|
Conflicts:
background_scripts/completion.coffee
|
|
|
|
A custom search engine like this...
i: https://www.google.ie/search?q=%s&num=30&newwindow=1&biw=1918&bih=1015&tbm=isch Google image search
Should not match a URL like this...
https://www.google.ie/search?q=bbc+sport
|
|
If the input is "w ", the user is going for something like "w query
terms" (a custom search engine).
So the domain completer should not offer completions if the user has
finished the first word.
|
|
This makes the behaviour consistent between custom and non-custom
searches when there is no selection in the vomnibar (omni-mode only).
Approach:
- When there is no selection in the vomnibar, we *always* send the
query to to background completer (that is, both for default
searches and for custom searches) and ask the completer to provide
only the primary suggestion. The primary suggestion is just what
you get if you append the query terms to the search URL (default or
custom). We then immediately open the first response.
The round trip for default searches isn't strictly necessary. However,
this uniform approach disentangles some nasty logic in the vomnibar when
we're trying to handle several cases (default or custom search, with or
without prompted text, with or without any suggestions at all). The
extra round trip simplifies the logic to such a great extend that it's
worth it.
|
|
|
|
|
|
This revamps how search-engine configuration is handled, and revises
some rather strange legacy code.
|
|
|
|
|
|
|
