pardes.git/test/macos-snapshots, branch main

Add macOS backdrop blur and preserve PDF ink opacity

2026-10-01T03:12:14Z

Align macOS rendering with Linux and establish parity regressions

2026-10-01T03:12:14Z

Give macOS scrolling momentum

2026-10-01T03:12:14Z

A swipe that was still moving when the fingers lifted stopped dead. The dial next to it has thrown properly since pardes_rotate_end, and this is that same curve on the scroll accumulator: velocity sampled between events off the monotonic clock, weighted toward the newest sample because a flick is decided by how the hand was moving when it left, and a fling that ramps up from zero at the floor rather than switching on at it. A coast stops at the end of a document rather than spinning its remaining velocity against the edge, which is why spendScroll now reports whether the pane moved and why paneAt is public. Only a step that delivered a press can report an edge — the many steps between two rows cross nothing. Nothing suppresses AppKit's own momentum, and nothing needs to: its momentum events are ordinary pardes_scroll calls, every one of which cancels the coast before spending its travel, so on a real trackpad the system takes the gesture over about a frame after the lift and the tail it ends on is below the floor. This is the path for devices AppKit does not fling for — and the only one a script can reach, since NSEvent phases have no public constructor. momentum.snap asserts both halves, which is why it needs the long scrollback: a flick with no document left proves nothing. A slow swipe is byte-identical after the release; a hard one coasts twenty-four rows further on its own, and a finger back on the pad stops it there. Co-Authored-By: Claude Opus 5 (1M context)

Let WindowOpacity through on macOS

2026-10-01T03:12:14Z

The builtin was gated on the SDL platform, so the macOS shell never registered it and the setting had nowhere to land. macOS already had transparency, but only the binary kind a theme decides — a theme with no background of its own drops the ground and the blur shows through. WindowOpacity is the graded, theme-independent version, and the window had no way to hear about it. pardes_window_opacity reports the percentage and acknowledges the request in one read, beside pardes_theme_bg: an SDL surface can arrive without an alpha channel and has to be able to refuse, while an AppKit window always composites per pixel, so there is nothing here to refuse and no rollback to perform. The view then paints the rule shaders/ui.frag.glsl states for the SDL shell: backgrounds — ground, cell and band fills, and the chrome rules over them — take the alpha, glyph ink never does, and the block cursor is exempt because it is foreground chrome that happens to be carried in a cell background. That exemption is why the cursor joins the colour in the background run key; it can no longer share a fill with the cells beside it. Two things the harness was missing fall out of testing it: it never adopted the theme background or the opacity, so a Theme or WindowOpacity in a script moved the core and never reached a pixel. draw-opacity records both ends of the alpha range, which is what makes the two-sided contract assertable — 60% reads 153..255, and 0% reads 0..255 with the ink still standing. Co-Authored-By: Claude Opus 5 (1M context)

Composite tag layers into the macOS e2e captures

2026-10-01T03:12:14Z

Tag text stopped living in the frame grid when the core moved the topbar, the column bar and every pane tag into tag layers of their own; the grid keeps the chrome and the physical grips. readFrame still read the grid alone, so every capture recorded blank header rows and each script timed out waiting for text that was on screen but not in the dump. Flatten the layers back over the grid at one cell per column — the tty backend's spelling, which these goldens exist to stay comparable with — and follow pardes_font_take to its size out-param. The goldens were recorded before the column bar and before the tag gap, so they are rerecorded here: the captures now agree column for column with what the tty renderer paints today. Co-Authored-By: Claude Opus 5 (1M context)

trunk: resume before the Reload experiment

2026-09-15T20:24:42Z

Empty marker on the last pre-Reload change. Keep the Reload experiment on reload (3801914), its first change on reload-start (200a1fc), and the unfinished performance investigation on reload-perf-wip.

macos: use two-finger click for Exec and deep press for Look

2026-09-07T17:10:43Z

macos: one tagline rule for both hosts, a kqueue beside the inotify, and effects that compile

2026-09-01T16:57:41Z

Three things this shell had its own copy of, and in each case the fix is that it stops having one. **The tagline band.** A pane tag draws at `gui_tagline_font_percent` of the body face and the band it sits on shrinks with it, while the grid row stays body-sized — so something has to decide where the shorter band sits in the taller row. This shell decided by centring, always, which is precisely the case `config.gui_topbar_pane_border_px` exists to prevent: the topbar's unused half-band meets the first pane tag's unused half-band and the window background shows through the seam. The strip is as wide as the bands are short — on a 20-pixel cell, 4 physical pixels at the default 82%, 10 at 50%, 14 at 30% — so it grew as the tagline face shrank and read as "the tagline is wrong on the mac" rather than as one missing rule. The rule is `pardes.taglineBandOffset` in the core now and both pixel hosts call it: row zero bottom-aligned, the first pane-tag row top-aligned, the two joined by `gui_topbar_pane_border_px` in the theme's scrollbar-track colour, every row between centred, and a `Tagbottom` band on the final row flush with the window edge — with the sub-cell strip beneath it painted in that band's own colour, because the core grid holds only whole cells and a window is any height it likes. `pardes_tagline_band_offset`, `pardes_topbar_pane_border_px` and `pardes_topbar_pane_border_rgb` carry it over the C ABI as PHYSICAL pixels: the host multiplies its points by the backing scale going in and divides coming out, which is the snapping `Metrics` already does for the cell, and is what keeps a one-pixel rule one pixel instead of a two-pixel smear. **The watch.** `file_watch.zig` was one mark/reconcile transaction over `inotify`, so the tty shell, the SDL window and the detached daemon all watched nothing off Linux: an edit made outside pardes never reached the pane, and a PDF replaced on disk kept rendering the old inode. It is the same transaction over two kernels now — `init`, `wait`, `stop`, `drain`, `markDir` and `unmarkDir` are still the whole of it, and the hosts wait on a kqueue and poll it exactly as they did the old descriptor. A macOS mark is TWO filters, because a kqueue directory filter reports its entries changing and never a write to a file already inside it: the parent mark follows rename-over saves, `markFile` catches in-place writes, and `remarkFile` re-arms the file filter once a rename has moved the inode. That is the same pair the AppKit host's DispatchSources already used for the same reason. Directory marks are deduplicated here by device and inode, because each `EVFILT_VNODE` filter needs a descriptor of its own and inotify did that deduplication itself; `stop` and `drain` wake through the one `EVFILT_USER` filter, since a kqueue cannot simply be read the way an inotify descriptor can. **The effects.** The three `crt.ci.metal` entry points are `extern "C" [[stitchable]]`. `CIKernel.kernels(withMetalString:)` compiles that source at runtime, looks for stitchable functions, and rejects the WHOLE source with "cannot find a valid stitchable Metal function in the source" when it finds none — so `ScenePostprocessor.init?` returned nil and every scene effect and panel transition silently degraded to the plain CoreText draw. The `effect_sources.zig` test pins the exact spelling of all three, and `draw-effect` in the e2e suite catches the degradation rather than the spelling. Beside them, the offscreen harness owes the core a PRESENTATION. Its window is borderless and never ordered front, so AppKit runs no display cycle and `pardes_frame_presented` — whose only caller is `draw(_:)` — never fired. The core holds pointer gestures inert while a layout mutation has not reached a backend, which for an unpresenting harness is the rest of the script: the first pane a script opened silently killed every later click, drag and Look. So `readFrame` presents what it just rendered, into a bitmap nobody reads. `PARDES_CHROME` also looks under `/Applications`, where a browser's executable lives inside an application bundle and never on `PATH`. The macOS goldens are regenerated; docs/macos.md, config.md, detached.md, web.md and the design PDF follow.

term_pane + builtins: terminal pane work, builtins/config additions, snapshots

2026-08-19T02:45:58Z

big slow change: prebuilt shaders (SPIR-V/Metal), core gui reflow, docs, web + snapshot refresh

2026-08-19T02:44:42Z