The global counter approach is easy to understand but it's brittle and depends on implicit behaviour that is not very discoverable.
With a global counter, if any goroutine accidentally decrements the counter twice, we'll think lazygit is idle when it's actually busy.
Likewise if a goroutine accidentally increments the counter twice we'll think lazygit is busy when it's actually idle.
With the new approach we have a map of tasks where each task can either be busy or not. We create a new task and add it to the map
when we spawn a worker goroutine (among other things) and we remove it once the task is done.
The task can also be paused and continued for situations where we switch back and forth between running a program and asking for user
input.
In order for this to work with `git push` (and other commands that require credentials) we need to obtain the task from gocui when
we create the worker goroutine, and then pass it along to the commands package to pause/continue the task as required. This is
MUCH more discoverable than the old approach which just decremented and incremented the global counter from within the commands package,
but it's at the cost of expanding some function signatures (arguably a good thing).
Likewise, whenever you want to call WithWaitingStatus or WithLoaderPanel the callback will now have access to the task for pausing/
continuing. We only need to actually make use of this functionality in a couple of places so it's a high price to pay, but I don't
know if I want to introduce a WithWaitingStatusTask and WithLoaderPanelTask function (open to suggestions).
We had some test flakiness involving the index.lock file which is fixed by this commit.
We shouldn't be accessing newTaskID without the mutex, although I'm surprised that this
actually fixes the issue. Surely we don't have tasks (which typically render to the main
view) which use index.lock?
We refresh the view after reading just enough to fill it, so that we see the
initial content as quickly as possible, but then we continue reading enough
lines so that we can tell how long the scrollbar needs to be, and then we
refresh again. This can result in slight flicker of the scrollbar when it is
first drawn with a bigger size and then jumps to a smaller size; however, that's
a good tradeoff for a solution that provides both good speed and accuracy.
Up till now our approach to rendering things like file diffs, branch logs, and
commit patches, has been to run a command on the command line, wait for it to
complete, take its output as a string, and then write that string to the main
view (or secondary view e.g. when showing both staged and unstaged changes of a
file).
This has caused various issues. For once, if you are flicking through a list of
files and an untracked file is particularly large, not only will this require
lazygit to load that whole file into memory (or more accurately it's equally
large diff), it also will slow down the UI thread while loading that file, and
if the user continued down the list, the original command might eventually
resolve and replace whatever the diff is for the newly selected file.
Following what we've done in lazydocker, I've added a tasks package for when you
need something done but you want it to cancel as soon as something newer comes
up. Given this typically involves running a command to display to a view, I've
added a viewBufferManagerMap struct to the Gui struct which allows you to define
these tasks on a per-view basis.
viewBufferManagers can run files and directly write the output to their view,
meaning we no longer need to use so much memory.
In the tasks package there is a helper method called NewCmdTask which takes a
command, an initial amount of lines to read, and then runs that command, reads
that number of lines, and allows for a readLines channel to tell it to read more
lines. We read more lines when we scroll or resize the window.
There is an adapter for the tasks package in a file called tasks_adapter which
wraps the functions from the tasks package in gui-specific stuff like clearing
the main view before starting the next task that wants to write to the main
view.
I've removed some small features as part of this work, namely the little headers
that were at the top of the main view for some situations. For example, we no
longer show the upstream of a selected branch. I want to re-introduce this in
the future, but I didn't want to make this tasks system too complicated, and in
order to facilitate a header section in the main view we'd need to have a task
that gets the upstream for the current branch, writes it to the header, then
tells another task to write the branch log to the main view, but without
clearing inbetween. So it would get messy. I'm thinking instead of having a
separate 'header' view atop the main view to render that kind of thing (which
can happen in another PR)
I've also simplified the 'git show' to just call 'git show' and not do anything
fancy when it comes to merge commits.
I considered using this tasks approach whenever we write to a view. The only
thing is that the renderString method currently resets the origin of a view and
I don't want to lose that. So I've left some in there that I consider harmless,
but we should probably be just using tasks now for all rendering, even if it's
just strings we can instantly make.
