Enable warnings for both. Favour casting to signed rather than casting to
unsigend for comparisons. Per isocpp's core guidelines ES.102, signed
arithmetic gives more expected results. If we need one extra bit of range,
using longer signed types achives that.
Fix#6010Fix#6012
Message caching is handled by SessionChatLog in memory even when history is
disabled. ChatLog doesn't need to worry about how the messages its rendering
are being stored. Dynamic loading up and down in chatlog is sitll functional.
This commit fixes the behavior when a message is received while the
chatlog is scrolled to the bottom. With this change, the chatlog will
stick to the bottom when it is scrolled all the way down. If it is
somewhere in the middle (e.g. for search) the chatlog will not change
its position.
As of 2019-10-09, toxme.io was taken offline permanently. Remove UI and code
in qTox relating to it. Revert this commit if it comes back online in the
future.
Fix#5897
The previous implementation of hiding the date line would cause 100% cpu
usage. When the date line was shown it would hide the top line, causing
the date line to be hidden again due to a state change in which dates
was visible.
This is a minimal patch to work around the issue by pretending the line
covered by the date line is the first visible line when the dateline is
shown
Fixes#5620
"Sent" to "Delivered", since that's what we're tracking also rename "isOffline"
to "isPending", which is boolean the same, but "isOffline" was opposite of its
correct value before, so boolean reverse the assignment.
following https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#Rh-override
virtual means exactly and only "this is a new virtual function."
override means exactly and only "this is a non-final overrider."
final means exactly and only "this is a final overrider."
Nothing was changed from e.g. override to final, just reduced duplication of
these labels.
ChatLogIdx is a strong type where the underlying data is only supposed
to be used in very rare circumstances. The ChatLog providing the indexes
provides no guarantees about what the first ChatLogIdx or last
ChatLogIdx will be. This commit removes unnecessary casts to underlying
data and fixes assumptions made about the underlying data
Functional changes
* Offline messages are still sent when the chat log is cleared
* Spinner now does not wait for history to be complete, just a receipt
from our friend
* Export chat and load chat history are now available in group chats
* Merged save chat log and export chat log
* Note that we lost the info messages in the process
NonFunctional Changes
* FileTransferWidget slots only called for correct file
* Settings::getEnableGroupChatsColor now embedded in
GenericChatForm::colorizeNames
* Settings::setEnableGroupChatscolor now emits signal connected to
GenericChatForm::setColorizedNames to keep state in sync
* Chatlog history not reloaded on setPassword()
* I am pretty sure this had no purpose
* Removed a lot of responsibility from ChatForm
* History moved to ChatHistory implementation of IChatLog
* OfflineMsgEngine moved to FriendMessageDispatcher
* Export chat and load chat history moved to GenericChatLog
* Backed by IChatLog so can be used generically
* Message processing moved to FriendMessageDispatcher
* The action of sending files to coreFile is still handled by
ChatForm, but displaying of the sent messages is done through IChatLog
-> GenericChatForm
* Search moved to ChatHistory/SessionChatLog
* All insertion of chat log elements should be handled by
GenericChatForm now
* Removed overlapping responsibilities from GroupChatForm
* Search and message sending goes through ichatlog/messagedispatcher
too
* Lots of search functionality pushed down into IChatLog
* Some of the file logic was moved into Widget. This is mostly to avoid
scope increase of this PR even further.
* History APIs removed that were no longer used
Add one more rergular expression that should match magnet links.
Unfortunately, magnet links may contain a lot of spaces, so
the message after the link may also be treated as a part of link.
Make Text::idealSize() a virtual function and override it in the derived
class Timestamp. The idealSize() function for Timestamps now returns a
tighter size to enable better right-alignment of the timestamps in
chatlog.
Fixes#3957. Note that this change assumes that timestamps do not
contain RTL text.
Toxcore has a 3 state pause, us, them, or both. Currently our UI
messes up if both parties pause. This changeset changes our UI behavior
to show whether we're paused, or if we are waiting on the remote to
unpause.
Rational here is that the current FileTransferWidget is quite
entangled with core logic. If we are going to instantiate the
FileTransferWidget without an active file transfer the widget needs to
behave sanely without getting messages from toxcore. This changeset is
an attempt to allow us to move from any FileTransferWidget state to any
other state without having to go through the appropriate state
transitions.
Original work by tox-user.
The default theme is now placed in a different path:
- C:\users\%username%\AppData\roaming\qtox\themes - for Windows
- ~/Library/Application Support/qtox/themes - for MacOS
- ~/.config/qtox/themes - for other Unix systems including GNU/Linux
Thanks to that it's no longer required to recompile the program to modify
the theme.
In addition to that the default theme is also included as a resource. If
the theme folder or some file inside it is missing, qTox will load it
from resource.
This makes the usage of QFileDialogs consistent by using he native
file picker in all cases. Also makes the usage of the last location
consistent. Furthermore this removes default parameter values and uses
Q_NULLPTR.
Fix#3124Fix#3004
Instead of loading a set 7 days of history. Better performance when there are lots of messages, and better context when friends haven't talked in over a week.
Removed historyBaselineDate, introduced in deb8440c6a to fix duplicate messages, but duplicate messages were very likely fixed by https://github.com/qTox/qTox/pull/4607.
Also refactored history loading.
Fix#4853Fix#4295
*Instead of searching strictly valid URIs, allow anything characters following scheme. This allows for UTF-8 characters used in other languages, as well as parentheses and other ASCII characters. This will over-match some invalid URLs.
*Ignored surrounding characters of URIs and ending punctuation
*Fix www-only links by adding http scheme to href
Brief list of changes:
- changed way of getting count of formatting symbols for captured
string: patternSignsCount() function replaced with constant index
which points to regex group that captures formatting symbols.
Regexes updated accordingly;
- updated docs for applyMarkdown() function;
- applyMarkdown() function refactored:
* changed var names;
* logic is more strict: get captured string -> crop formatting symbols
if necessary -> wrap captured string into HTML and replace in result.
Brief list of changes:
- removed several headers;
- added single array of pairs for direct mapping regexes to HTML
wrapper strings;
- COMMON_PATTERN splitted into two regexes;
- wrapper for hrefs is a single constant.
Brief list of changes:
- removed workarounds while applying markdown;
- URL highlighting now is independent for markdown support;
- URL patterns reverted to the state they were before.
The original reason that the Qt picker was used instead of the native
picker was that the native option would cause Nautilus/GNOME-based
pickers to hang.
This turned out to be due with a Qt bug with parenting Gtk windows. As a
result the parent of each file dialog window has been set to NULL,
eliminating the crash. As far as tests have shown, this produces no
adverse effects on either floating or tiling wms.
Fixes#3494
Fix#4275
I did not consider that replacing one substring with another will point
to shifting position of next URL found with regexp. That's the behavior
of Qt's "QRegularExpression" class - it takes a string into constructor
and seems to make its copy inside so changing source string does not
affect this regex object
