Before, we used the 'changed' signal for the SaveManager - however, that also
was emitted when only the internal structure changed. Now we add a new signal
for that.
KeyConfig needs this feature, because it can fix some deprecated commands
during __init__ and emit its dirty-signal, but that happens before the saveable
is added.
Revert "Fix maxsplit-splitting with empty args (""/'')."
This reverts commit 46396cce1e.
Revert "Remove quotes with split=False commands."
This reverts commit 81bc5dae94.
See #564 and #453.
Before, the first item was unconditionally selected when none was selected
before. With :completion-item-prev (e.g. Shift-Tab), it makes more sense to
select the *last* one.
We want to ignore some positional arguments without ignoring flags/values - and
since there's no easy way to "unparse" an argparse namespace, we instead pass
it as json.
Also note we can't pass it as a file easily, as args have to be available very
early. Passing it as an argument shouldn't be an issue though.
Before c5a2039da4 (standarddir refactoring), we
only checked the commandline arguments for the config file, but not when
getting the quickmarks location (as the 'args' argument was None). This means
quickmarks were saved to the default config dir even with -c ''.
With that commit, this was "fixed" accidentally, but quickmarks couldn't handle
the filename being None.
Before, we always loaded the default session (if it existed) and then deleted
it. This was surprising as the default session was deleted even when another
session was loaded.
Now we don't delete it at all, and save the session to load in the state file.
See #523.
If an user e.g. has a download-directory of ~/föö, but has LC_ALL=C set, we'll
get an UnicodeEncodeError when trying to validate it. This is now handled
properly by raising a ValidationError.
Fixes#562.
A better solution is to use QSocketNotifier and os.wakeup_fd to get notified
about new signals.
Thanks to Yuya Nishihara / TortoiseHG for the hint!
Fixes#555.
Before we limited the history items we could simply call WebHistory's
historyContains before iterating through all items in the history completion.
Now however it's possible an item is in the real WebHistory, but not actually
in the completion - so we always have to check the whole completion.
Before, if an URL was present early in the history and then again later, we
didn't move it to the end of the OrderedDict. This means it won't be loaded in
the completion.
Before, the item_added signal was emitted *after* an item was added, which
means the on_history_item_added slot always assumed the item already is in the
history.
Before, we initialized the completions once for every window spawned, which was
a waste of CPU-time and RAM.
Now we only initialize them once, when the user uses the completion for the
first time.
This makes qutebrowser.config.websettings much easier to understand, and saves
all defaults so it can restore them properly when a setting is set to an empty
string.
Before, when we set the fonts to empty strings instead of the true default, in
some cases anti-aliasing was broken.
Fixes#549.
Otherwise we would construct a QStandardItem with the
QStandardItem(int rows, int columns = 1) constructor, which will most likely
not do what we want.
- HistoryItem.atime now always should be an int/float.
- The data for the sort role should also be an int, not a string.
A float would also work, but maybe be slower for no real benefit.
Two things here. One is to use `WebHistory._new_history` only as a to-save
queue, so we now add entries to `_old_urls` when they are first created and
can now no longer iterate of `_new_history` in `__iter__()`.
Second is to stop blindly tacking new history entries on the end of the
history completion model. It does involve iterating over the model to find the
existing entry but we only do that if we know the duplicate is there, which is
fast to check.
This also ads another point of mutation to the history completion model which
may prove problematic if it leads to more segfaults.
Each new HistoryEntry is emitted after being added to the global history
store. Current members of the HistoryEntry are `url` and `atime`. `title`
should be coming soon.
Adds a basic completion model implementation around the global browser
history and registers that for the open command.
Modifies WebHistory to add an __iter__ method and to use a dict instead of a
set to store an entire HistoryEntry for each archived item instead of just the
URL. Brief tests showed that the lookup time for set and dict are very
similar. They are at least on the same order of magnitude. Testing membership
of a list on the other hand, as was the case before a set was used, was four
orders of magnitude slower on my machine.
This makes it possible to use Qt's QSortFilterProxyModel::lessThan option for
completions where it doesn't make sense to priorize matches starting with the
entered string, e.g. for URLs. In return, we get a *much* better performance
(several seconds when opening the completion).
See #531.
This is enabled by default to keep the same default behaviour which is like
Vimium - mixing e.g. single-char letters and double-char letters, and
scattering/shuffling the labels to have an uniform hint key distribution.
If disabled, the behaviour is more similiar to dwb, which has a fixed hint
string length and simply fills the string starting with the first possible hint
char.
PyQt uses qHash() for __hash__, and qHash for QSslError was added with Qt 5.4.
This means 2da45e98ca raised TypeError there as
QSslError is unhashable.
For those older Qt versions, we implement __hash__ ourselves which does about
the same thing as Qt does, combining the DER (binary) representation of the
certificate and the error() (which is just a QEnum, hashable as int).
For every (scheme, host, port) tuple, we save all SSL errors we asked the user
about, and if everything matches (scheme, host, port, error, certificate), we
don't ask the user again.
Fixes#422.
