Fix/monitor disconnect #1255
Open
Fix/monitor disconnect #1255
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
alex-ds13
force-pushed
the
fix/monitor-disconnect
branch
2 times, most recently
from
c4cb73f to
97245c8
Compare
|
I've been running this for a few days and I'm pretty happy with it - do you have any lingering concerns before we start getting ready to merge? |
alex-ds13
force-pushed
the
fix/monitor-disconnect
branch
from
9c98faf to
5f5b83f
Compare
Let me fix the bar issue tonight and then it should be ready to merge... |
Previously, when caching a workspace config for a monitor it would simply store the `DefaultLayout` on `layout` even if the original workspace config had the `layout` as `None` which makes komorebi create a workspace with the `layout` as default `BSP` and the `tile` set at `false`. However this wasn't being taken into consideration, so floating workspaces would becoming tiling `BSP` workspaces after a monitor disconnect and reconnect. This commit fixes this by turning the `layout` to `None` when `tile` is `false`.
Store the `WorkspaceConfig` on the `Workspace` itself so that when we want to cache the workspace as `WorkspaceConfig` on the monitor cache it properly saves things like the workspace rules and the custom layout and custom layout rules.
This commit reworks the way the `postload` and the `reload` functions apply the monitor configs to the monitors. Previously it was looping through the monitor configs and applying them to the monitor with the index corresponding to the config's index. However this isn't correct, since the user might set the preferred indices for 3 monitors (like monitor A, B and C), with the preferred index set to 0 for A, 1 for B and 2 for C, but if only monitors A and C are connected then komorebi would apply config 0 to A and config 1 to C, which is wrong it should be 2 for C. This commit changes the way the configs are applied on those functions. Now it loops through the existing monitors (already in order), then checks if the monitor has a preferred config index, if it does it uses that one, if it doesn't then it uses the first monitor config that isn't a preferred index for some other monitor and that hasn't been used yet. For the situation above it means that it would still apply config 2 to monitor C. And in case there aren't any display_index_preferences set it will still apply the configs in order.
If we have display_index_preferences that set a specific config index for a specific display device, but that device isn't loaded yet, now we store that config with the corresponding `device_id` on the monitor cache so that when the display is connected it can load the correct config from the cache.
When a monitor was disconnected the containers from the removed monitor were being moved to the primary monitor. However they weren't restored so containers that were on an unfocused workspace of the removed monitor would have been cloak and were getting added to the main monitor still cloaked creating ghost tiles. This commit fixes that.
This commit makes use of both `serial_number_id` and `device_id` with the first taking priority on all monitor reconciliator code, monitor cache and on postload and reload. This should allow using the serial numbers on the `display_index_preferences` config, while keeping compatibility with the use of `device_id`. Although using `device_id` should be discourage since that value can change on restart while serial number doesn't appear to do so.
This commit makes sure that any workspace_rules that tried to move a window to a monitor that has been disconnected are removed. If the monitor is latter reconnected the workspace_rules should be added from the cached config.
This commit creates a new field on the `WindowManager` called `monitor_usr_idx_map` which contains a map of user intended index for monitors to their actual monitors' index. It will be rebuilt on `load_monitor_information` by taking into account the `display_index_preferences`.
This commit makes so a bar can be disabled when it's monitor is disconnected and re-enabled when said monitor reconnects. This commit also uses the new `monitor_usr_idx_map` to properly map the monitor index given by the users on config to the actual monitor index on the `WindowManager`, since in case some middle monitor gets disconnected the index of the monitors to the "right" of that one will be lowered by 1. This should allow for cases with monitor A, B and C, if monitor B disconnects, its bar will be disabled and monitor C will properly change its monitor index internally to 1 so it still gets the infos from monitor C (which now will have index 1 on the `WindowManager`).
This commit changes the MONITOR_CACHE to hold an actual monitor with its state instead of the config. This allows us to keep the state of a disconnected monitor, so that when the monitor reconnects we still have access to all its workspaces and windows as it was before. While the monitor is disconnected, all the windows from that monitor are not considered as handled by komorebi and so they're free to be handled at any point, if for example the user uses `alt+tab` or presses the taskbar icon of one of these windows that window will show up producing it's `Show` event and will become handled by the currently focused workspace of the focused monitor. When the disconnected monitor reconnects it checks all windows and once it notices that this specific window is now being shown on another monitor/workspace then it ignores it.
This commit makes sure that if the state on file isn't up to date with the expected `State` struct (maybe after an update) it doesn't panic komorebi entirely, instead it ignores the state and continues with a clean state.
When there is a monitor disconnect/reconnect, usually it produces multiple monitor events along with it, like the monitor resolution change and the work area change. With the bound set to 1 it would sometimes result in missed events. This commit increases the bound to 10 to prevent this from happening.
alex-ds13
force-pushed
the
fix/monitor-disconnect
branch
from
8225cde to
f6fa86e
Compare
|
@LGUG2Z Should be ready to merge! 🚀 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR builds upon #1241
This PR changes the monitor cache from caching a config to caching the actual state of the monitor itself.
With this PR when a monitor disconnects instead of "throwing" all windows to the main monitor, right now it does nothing and keeps your other monitors as is and simply stores the disconnected monitor state.
If then you reconnect (or wake up) the monitor again, then you'll get back to the same monitor with all the windows like you had before!
If while the monitor is disconnected you need to access some window that was on the disconnected monitor, you can simply
alt + tabto it or press the taskbar icon and the window will show on the focused workspace of the focused monitor and it becomes handled by this monitor/workspace. When the previously disconnected monitor reconnects it will see that this specific window is already handled by another workspace so it ignores it.