Make the msglist aware of whether we are still initializing or not.
We never stopped spinning the msglist if we didn't get any Directories back
from types.ListDirectories.
With this change, we can set the init state from the account and display
the spinner only if we don't know whether we have directories or not and else
the "no messages" string from the config.
This patch sets up the trigger config section of aerc.conf.
Each trigger has its own function which is called from the place where
it is triggered. Currently only the new-email trigger is implemented.
The triggers make use of format strings. For instance, in the new-email
trigger this allows the user to select the trigger command and also the
information extracted from the command and placed into their command.
To actually execute the trigger commands the keypresses are simulated.
Further triggers can be implemented in the future.
Formatting of the command is moved to a new package.
Before, we were using several IMAP-specific concepts to represent
information being displayed in the UI. Factor these structures out of
the IMAP package to make it easier for other backends to provide the
required information.
Before, the information needed to display different parts of the UI was
tightly coupled to the specific messages being sent back and forth to
the backend worker. Separating out a models package allows us to be more
specific about exactly what a backend is able to and required to
provide for the UI.
This commit introduces a new Aerc.Tick function that should be called to
refresh the internal state. This in turn makes each AccountView process worker
events.
The UI goroutine repeatedly refreshes the internal state before drawing a new
frame. The reason for this is that many worker messages may need to be
processed for a single frame, and drawing the UI is far slower than refreshing
the internal state. This has been confirmed in my testing (calling Aerc.Tick
only once per frame results in a slower display).
Many synchronization code has been removed. We can now write widgets without
having to care so much about races. The remaining sync users are:
- widgets/spinner: the spinner value is updated from inside an internal
goroutine
- lib/ui/invalidatable: Invalidate may be called from any goroutine
- lib/ui/grid: same
- lib/ui/ui: an internal goroutine needs read access to UI.exit
- worker/types/worker: Worker.callbacks is used for both worker and UI
callbacks
The exact goroutine requirements for Drawable have been documented.
MessageStore has a lot of exported fields that can be read from the outside.
Each read must be protected, because a call from Update could happen at any
time.
Many Drawable implementations have their own Invalidate and OnInvalidate
functions, with an unexported onInvalidate field. However OnInvalidate and
Invalidate are usually not called in the same goroutine. This results in a race
on this field, e.g.:
Read at 0x00c000094748 by goroutine 7:
git.sr.ht/~sircmpwn/aerc2/widgets.NewDirectoryList.func1()
/home/simon/src/aerc2/widgets/dirlist.go:85 +0x56
git.sr.ht/~sircmpwn/aerc2/widgets.(*Spinner).Start.func1()
/home/simon/src/aerc2/widgets/spinner.go:93 +0x1bb
Previous write at 0x00c000094748 by main goroutine:
[failed to restore the stack]
Goroutine 7 (running) created at:
git.sr.ht/~sircmpwn/aerc2/widgets.(*Spinner).Start()
/home/simon/src/aerc2/widgets/spinner.go:46 +0x8f
git.sr.ht/~sircmpwn/aerc2/widgets.NewDirectoryList()
/home/simon/src/aerc2/widgets/dirlist.go:37 +0x286
git.sr.ht/~sircmpwn/aerc2/widgets.NewAccountView()
/home/simon/src/aerc2/widgets/account.go:50 +0x5ca
git.sr.ht/~sircmpwn/aerc2/widgets.NewAerc()
/home/simon/src/aerc2/widgets/aerc.go:60 +0x800
main.main()
/home/simon/src/aerc2/aerc.go:65 +0x33e
To fix this, introduce a new type, Invalidatable, which protects the field.
Unfortunately the Drawable must be passed to the callback function in
Invalidate, so we still need to re-implement this in each Invalidatable user.
When changing to an empty directory, ml.selected is 0, and the length
of ml.store.Uids is 0. The loop condition is always true so we have
an infinite loop causing 100% CPU usage and prevents us to change to
other directories.
Signed-off-by: Tom Lebreux <tomlebreux@cock.li>