While most Qt types cna use implicit data sharing
pasing our own structs means copies. const& ensure
it's only copied as needed by Qt.
Signed-off-by: Rachel Powers <508861+Ryex@users.noreply.github.com>
feat: add details to mod pack downloading
feat: add logging rule sloading form `ligging.ini at data path root
feat: add `launcher.task` `launcher.task.net` and `launcher.task.net.[down|up]load` logging categories
fix: add new subtask progress to the end of the lay out not the beginning (cuts down on flickering)
Signed-off-by: Rachel Powers <508861+Ryex@users.noreply.github.com>
Oh boy this is big.
> TaskStepProgress struct is now QMetaObject compatabile and can be sent through signals
> Task now has a method to propogates sub task progress it must be signal bound by each task containing a task wishing to report progress of it's children.
> Downloads report speed
> Tasks now have UUIDS to track them
- use when reporting
- use when logging
- use when storeing them or objects related to them
Signed-off-by: Rachel Powers <508861+Ryex@users.noreply.github.com>
This turns issues like creating two shared ptrs from a single raw ptr
from popping up at runtime, instead making them a compile error.
Signed-off-by: flow <flowlnlnln@gmail.com>
By now, it's a recurring pattern of wanting to restrict aborting in
certain situations. This avoids further code duplication, and adds a
signal that external users can hook up to to respond to such change.
Signed-off-by: flow <flowlnlnln@gmail.com>
This allows us to emit all the necessary stuff when we're finished in
the case of starting a task with no subtasks. In particular, this caused
the mod updater to not work properly :)
Signed-off-by: flow <flowlnlnln@gmail.com>
In a way, sequential tasks are just concurrent tasks with only a single
task running concurrently, so we can remove LOTS of duplicated logic :)
Signed-off-by: flow <flowlnlnln@gmail.com>
This fixes an annoying issue where concurrent tasks would try to start
multiple tasks even when there was not that many tasks to run in the
first place, causing some amount of log spam.
Signed-off-by: flow <flowlnlnln@gmail.com>
This makes it possible to run a task in another thread.
I added a variable to toggle debug prints because they seem to trigger
an assertion on Qt internals when the task in on another thread. Of
course, this isn't awesome, but can wait until we improve our logging.
Signed-off-by: flow <flowlnlnln@gmail.com>
Heavy workloads can consume a ton of time doing their stuff, and starve
the event loop out of events. This adds an event processing call after
every concurrent task has been completed, to decrease the event loop
stravation on such loads.
Signed-off-by: flow <flowlnlnln@gmail.com>
This is a variation of a Sequential Task, in which a subtask failing
will prompt the next one to execute, and a subtask being successful will
stop the task.
This way, this can be used for easily managing fallbacks with tasks. :D
Signed-off-by: flow <flowlnlnln@gmail.com>
Previously, it would not update the global counter if the subTask didn't
update its progress, even though progress was being made.
This also prevents a segmentation fault while aborting the task.
Signed-off-by: flow <flowlnlnln@gmail.com>
