Add all-outputs mode for #78#79
Closed
jason-curtis wants to merge 8 commits into
Closed
Conversation
When --all-outputs (or general.all_outputs=true in config) is set, wl-kbptr creates a layer-shell surface on every connected output simultaneously. The first surface receives exclusive keyboard focus; the rest are visual-only. All surfaces share a single label namespace spanning the global bounding box of all outputs, so the user can type any visible label regardless of which display currently has window focus. The tile mode works without changes: it renders in global coordinates and each per-output cairo context is translated by (-output.x, -output.y) so cells naturally clip to each output's bounds. After a selection, resolve_result_output() identifies which output contains the result centre, sets state->current_output, and converts the result rect to output-local coordinates for move_pointer. Backwards compatible: defaults to single-output mode. The existing -O, -r, and single-output behaviour are unchanged. Closes moverest#78
Cell density: compute sub_area_size from average logical monitor area rather than the full bounding box so each monitor gets the same cell density as in single-output mode (same ~676 cells worth of size per monitor, distributed across the bounding box). Dead zones: when the result centre lands in a gap between monitors (empty bounding box space with no output), snap it to the nearest output boundary rather than falling back to the first output at arbitrary coordinates.
Documents the --all-outputs / -A flag, its tile-mode-only scope, cell density behaviour, and dead zone handling. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
In all-outputs mode, cells whose centre falls in a gap between monitors (empty bounding-box space not covered by any output) no longer get labels assigned. This keeps the full label budget for reachable cells, reducing keystroke count on multi-monitor setups with unaligned displays. A cell_idx_map[] array maps label-index -> linear cell-index for the valid subset. The map is NULL in single-output mode (identity mapping). Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Owner
|
I haven't read the code in details yet, however I have a few thoughts. There are two ways to think about it, either:
Now for the dead-zones... I'm not sure this is the right approach. Having tested your changes for one it doesn't cover the full screens which means areas that can't be selected. On top of that, this means having to keep track of these. Looking at the problem a better way to think about it would be to think in terms of regions. Each screen is a region and each region has a number of identifiers associated with it (we want a similar density for each). Then we render all the regions with continuous indexes. No need to keep track of dead zones with this method. Where things get more interesting is if regions overlap:
This way we cover the whole displays and it should make things easier. We "just" need to determine the regions then the number of selectable area per region and then render with these parameters. FYI, I've tried your code and it segfault when I chose an area. Not sure it's worth investigating if it's going to change. |
Instead of computing a global bounding-box grid and skipping cells whose centre falls in a gap between monitors, treat each monitor as an independent region with its own rows×cols grid. Labels are indexed continuously across all regions with no dead zones. - Add struct tile_region (area, rows/cols, cell dimensions, label offset) - tile_mode_enter: build a regions[] array in all-outputs mode; keep the single-output flat-grid path unchanged - tile_mode_key: look up the selected label's region and compute its cell rect directly - tile_mode_render: extract render_cell() helper; iterate per-region in all-outputs mode; keep the flat-grid render path for single output - tile_mode_state_free: free regions[] Benefits over the previous approach: - Every label maps to a real, clickable cell on a real monitor - Full label budget is used (no wasted labels for inter-monitor gaps) - Naturally handles overlapping/mirrored displays (each output is its own region regardless of spatial overlap) - No segfault-prone global-coordinate dead-zone snap logic needed Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Remove cell_idx_map: it was always NULL after the region-based rewrite; drop from tile_mode_state, the constructor, render loop, key handler, and free - calloc instead of malloc in tile_mode_enter so sub_area_* fields are zero-initialised in the multi-output path - Fix overlay_surface pointer alignment: wp_fractional_scale_v1 was one column off from the other pointer fields - handle_layer_surface_configure: remove redundant double-set of overlay->configured; set it once before the branch and use a local was_configured for the else-if guard - render_cell comment: fix reference to ms->label_selection (not a parameter); rename to match the actual argument - resolve_result_output comment: describe the snap fallback accurately rather than calling it a dead-zone case - Revert two accidental char* spacing changes in main.c that diverged from upstream style without reason - README: update "How it works" bullet to say per-region labels, not global bounding box; remove stale Dead zones section; add config file example for all_outputs; tighten Cell density paragraph - config: add parse_bool and register all_outputs so it can be set in the config file as well as via -A Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Author
|
Thanks for the feedback. Implemented a new approach with a grid per output. Now the grid for each display, lines up with the monitor edges. Supporting arbitrary overlap cases is substantially more complicated but I'll give it a shot. It may be reasonable to skip support for overlapping outputs, or maybe just for simple and more common cases (mirroring or strict subset only?). Thoughts? |
Author
|
I think we should probably close this in favor of #78, which addresses your concerns and also adds floating mode support. |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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 implements a multi-monitor simultaneous mode.
wl-kbptr --all-outputs -o modes=tile,clicknow shows a tiled grid across all monitors, allowing user to select a tile across multiple displays.Currently works in my testing in tile mode. If you like it I can look at adding floating mode as well. It's not obvious to me how this would work in bisect/split modes so my inclination is to leave those out.