setGlobalDownloadLimit and setGlobalUploadLimit will now modify the
alternative speed limits if they are currently enabled and the regular
speed limits otherwise.
Add also two new commands to toggle the state of the alternative speed
limits and get their current state.
Closes#2203.
This should prevent situations in which the properties panel completely
covers the transferlist on page load.
Save also the height whenever the window is resized since that changes
the height of the panels without saving the height.
A different DPI setting might make the elements in the dialogue big
enough to make the scrollbars appear, but not big enough to make them
useful. So, never show them.
Now that the properties are loaded immediately when the selected torrent
or the property tab is changed, the frequency of the updates can be
reduced without affecting the user experience.
New update intervals:
* General: 5 seconds (10 if qBT is not reachable)
* Tracker list: 10 seconds (20 if qBT is not reachable)
* File list: 5 seconds (10 if qBT is not reachable)
Don't wait for new data to be fetched, request them as soon as
they are required, i.e. when a property tab is loaded or when a
different torrent is selected.
Don't reload HTML on tab change. Load the entire structure once and
keep the unused bits hidden.
Keep also the JS code separately, one file per tab.
NOTE: tabs content is not loaded. This will be fixed in the next commit.
Each column has a fixed width that allows to read its content.
The column for the name of the torrents has a variable width with an
upper and a lower constraint.
If the width of the page is not sufficient to fit the table, a scrollbar
will appear.
ajaxfn() is called every time the torrent list is filtered. This function
calls itself after a predefined amount of time by setting a timer. If
the function is called manually without destroying the timer, the
frequency of the XHR requests will increase.
Fix this by saving a reference to each timer that is created so that
they can be later destroyed.
Add also two helper functions to request updated data safely.
The input field to change password does not contain the actual
password, but its hash. When the preferences are saved, the hashed
password is considered as a new password.
Prevent this by comparing the new password with the hash of the
previous password.
Closes#2241.
Since e907306b41 ("Right align header text that also has right-aligned
row text"), right-aligned elements are no longer centered vertically.
This can be noticed with some Qt styles (e.g. Breeze).
Although RSS rules were correctly parsed, the wrong separator for the
"OR" condition was used to re-generate the string of the rule displayed
in the preferences.
Closes#2205.
New GET methods:
* version/api: current API version
* version/api_min: min API version supported
The API version is an integer and it's returned as plain text.
Whenever the API is changed, the api version is increased.
Whenever the API is changed in a way that is not backward compatible,
api_min is increased.
Only few file managers are able to select files when started from the
command line.
xdg-mime is required to retrieve the default file manager. If it's not
available or the file manager is not able to automatically select files,
the containing directory will be opened as usual.
Definitions:
Selection in QTreeView consist of two things:
currentIndex -- is a (dotted) cell where user clicked last time. Note
that it is a cell
selectedIndexes -- is a set of cells (blue) of current selection.
Checkboxes in torrent content lists are belong to COL_NAME column.
Problem:
The problem is that spacebar toggled checkbox only in currentIndex
index. This has two consequences:
1. It is impossible to toggle checkboxes on multiple rows
simultaneously.
2. If currentIndex is not in COL_NAME column a space key doesn't work
at all. This problem is amplifyed by the fact that SelectionBehavior
is set to SelectRows. So visually it is impossible to tell which
column does it belong to. For end user it looks like "space doesn't
work sometimes".
This patch addresses the problem by implementing TorrentContentTreeView
derived from QTreeView and overridding keyPressEvent(QKeyEvent*).
The code of TorrentContentTreeView::keyPressEvent is written under
inspiration from QAbstractItemView::keyPressEvent and
QItemDelegate::editorEvent.
It was reported (#2161) that enabling/disabling a downloading of a file
is considerably slow on libtorrent 1.0.3, but not on 0.16.x. The problem
is that a function torrent_file() in libttorrent 1.0.3 does a deep copy
of torrent_info, while get_torrent_info() in libtorrent 0.16.x only
returns a reference.
Sizes are now given in bytes.
Dates are Unix timestamps and converted to ISO 8601 in the web UI.
Numbers are not converted to strings.
-1 is returned for undefined values.
Some keys have been splitted:
Torrent list (json/torrents)
* num_seeds: Torrent seeds connected to
* num_complete: Torrent seeds in the swarm
* num_leechs: Torrent leechers connected to
* num_incomplete: Torrent leechers in the swarm
Torrent generic properties (propertiesGeneral/hash)
* total_uploaded: Total data uploaded
* total_uploaded_session: Total data uploaded this session
* total_downloaded: Total data dowloaded
* total_downloaded_session: Total data downloaded this session
* time_elapsed: Torrent elapsed time
* seeding_time: Torrent elapsed time while complete
* nb_connections: Torrent connection count
* nb_connections_limit: Torrent connection count limit
Global transfer info (json/transferInfo)
* dl_info_speed: Global downalod rate
* dl_info_data: Data downloaded this session
* up_info_speed: Global upload rate
* up_info_data: Data uploaded this session
Closes#1524.
This commit should improve performance when user navigating through
torrent list using up/down keys. A scrolling through all the list
(276 torrents) took:
Total wall time: 18.813s
Total CPU time: 3.210s
misc::parseHtmlLinks(): 0.096s
misc::parseHtmlLinks() is 8th most hottest function on
this use case.
This commit caches icons in TorrentContentModel in the same way they are
cached in TorrentModel.
This commit should improve performance when user navigating through
torrent list using up/down keys. A scrolling through all the list
(276 torrents) took:
Total wall time: 18.813s
Total CPU time: 3.210s
IconProvider::generateDifferentSizes(): 0.170s
IconProvider::generateDifferentSizes is 5th most hottest function on
this use case.
I thought merging all 3 sort-filter proxy models into one should
speedup dataChanged() signal. As turned out this is not the case. The
time difference is within inaccuracy range, so this doesn't affect
performance.
But I still think it is good to merge them into one proxy model as it
simplifies code.
Torrent numbers were recalculated on every dataChanged() signal. The previous commit
greatly increases the number of dataChanged() signals.
HEAD^^:
Total wall clock: 97.069s
updateTorrentNumbers() time: 0.033s
HEAD^:
Total wall clock: 96.132s
updateTorrentNumbers() time: 0.179s
HEAD:
Total wall clock: 95.535s
updateTorrentNumbers() time: 0.047s
After this commit the time of updateTorrentNumbers() is (almost) back to
the level that it was in HEAD^^.
In commit b50d733 TorrentModel moved from a periodic refresh, to using
postStatusUpdate(). In this transition I forgot to remove emition of
dataChanged() signal for the entire table.
According to my measurements this commit reduce CPU usage of qbittorrent
by a factor of 3:
Before:
Total wall clock: 97.07s
CPU time: 21.77s
- Time spent in TransferListDelegate::paint(): 14.60s
- Time spent in TorrentModel::forceModelRefresh(): 1.44s
- Time spent in TorrentModel::stateUpdated(): 0.02s
After:
Total wall clock: 96.13s
CPU time: 6.68s
- Time spent in TransferListDelegate::paint(): 2.63s
- Time spent in TorrentModel::forceModelRefresh(): <0.01s
- Time spent in TorrentModel::stateUpdated(): 1.73s
As it is seen the time spent in painting is reduced by a factor of 6 (14.60->2.63) at
the cost of slightly increased time of notifications that model is
changed (1.44->1.73). The next commits attempt to address this issue.
The problem is that torrentRow() does linear search over the list of all
available torrents. So it doesn't scale well for large number of
torrents. Removing the copying of QString from linear search
inner loop, speed up it considerably.
The proper solution should be using hash table instead of linear search.
This require more radical changes in TorrentModel and may be done in a
separate commit.
