docs: events in the dock and on its page

The dock's Events block, the events slide-over page, the availability-rule deviation, the probe lines and the removal of the left pane are described; the false remembered-width claim is gone.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

README.md

@@ -94,7 +94,7 @@ switches trigger mode on, waits for an event, then leaves trigger mode the way t
 "Go live" button does, closes the window, and prints on stdout the status line in
 both states, the frame and per-channel sample counts, each plot's painted-frame
 count, which channel each trace slot holds, and the allocated size and reveal
-state of the dock and the events pane.
+state of the dock, and the events page's controls (present and nested in the page).
 
 Every size it prints is read from a value the frame tick captured while the window was alive, not
 from the widget: the dump runs from the destroy handler, where the widget pointer is already freed
@@ -146,8 +146,8 @@ panel laid over the plot's right edge, opaque, with its own title, a *Close* but
 page. The Trigger badge opens the trigger form; the Source badge (when shown) opens the source kind's own panels
 (the E-1608's *Analog out* and *Digital I/O*, behind a segmented switch at the top of the panel). It
 closes with a second click on its badge, with `Esc`, with *Close*, or with a click on the plot; it never
-resizes the plot. The right-hand **dock** (the sidebar, **280 px**, fixed) holds the **readouts**: a
-compact, read-only copy of the cursors table (a header row `A  B  delta`, then one line per cursor
+resizes the plot. The right-hand **dock** (the sidebar, **280 px**, fixed) holds the **readouts** and, under
+them, the **Events** block (below): the readouts are a compact, read-only copy of the cursors table (a header row `A  B  delta`, then one line per cursor
 line) and of the measurements (one line each: the signal's colour chip, the type, the value). It is
 shown from the start and insensitive until the stream has signals. Each of the two blocks has an
 **Edit** button, which opens the full panel as a slide-over page -- `cursors` or `measure` -- over the
@@ -159,8 +159,12 @@ cells are a fixed number of characters wide and ellipsized, so the dock's width
 depend on what a value reads: the **who** cell of a cursor line is 5 characters (the full text is its
 tooltip), a value cell 9 (its tooltip is the full text). The dock and the pages show **one string**:
 `update_cursors_pane()` and `update_measures()` format each value once and hand that variable to the
-page's label and to the dock's, so the two cannot disagree. The remaining part of the redesign is
-listed under "Review findings not (yet) implemented".
+page's label and to the dock's, so the two cannot disagree. The **Events** block is a heading with a
+**Manage** button over a short list of the cached events, newest first -- one compact row each, never a
+picture -- and one click on a row replays that event (see "Captured events"). The full list, with the
+pictures and the comparison actions, is the `events` slide-over page that **Manage** opens. There is no
+events pane beside the plot and no Events button in the header: events never take any of the plot's width.
+The rest of the redesign is listed under "Review findings not (yet) implemented".
 
 Threading: one producer thread owns the data source (the UDPStreamer client is the built-in
 one) and calls `poll()`; the GTK main thread owns every widget and the GL contexts.
@@ -838,20 +842,36 @@ time to look at is still there:
   20261009-123250-0001.bmp   # a picture of it, drawn from the samples
 ```
 
-The **left pane** is opened on demand from the toolbar's event-history button, is collapsed by
-default, and its **divider can be dragged**: it starts at 240 pixels and remembers the width you last
-chose, and while it is closed it takes no room at all. Both the button and the pane exist while **recording is on**, or while an event is on
-screen: with the cache disabled and nothing being examined there is nothing to show, and a visible
-empty column beside the plot is what *off* must not look like. Recording is switched in the **trigger
-panel**, with the trigger's other options rather than in the pane whose contents it decides, and it is
-**off by default** -- writing every latch to disk is the user's decision about their disk.
+Events are read in two places. The **dock's Events block** lists them compactly -- one row per event, newest
+first, at most **24** with a dim "... N more, see Manage" row after them: a marker (a filled circle on the
+event on screen, a hollow one on each member of the comparison or accumulation on screen), the time and
+duration, and the event's sequence number (`#7`). **One click on a row replays it.** The block's **Manage**
+button opens the **`events` page**, a slide-over page over the plot with the full list (below), the *Live*
+button and the two group actions. The page replaces the left pane that used to sit beside the plot behind a
+header button and a draggable divider; both are gone, and so is the toolbar's Events button -- opening the
+list no longer resizes the plot at all.
+
+The block and the page exist all the time and are **insensitive, not hidden**, when there is nothing to show:
+they are available while **recording is on**, while an event is on screen (a replay or a group), or while
+the cache already holds events. **Deviation from the redesign spec,** which has the block insensitive
+"outside trigger mode": a replay or a group *leaves* trigger mode (a file has no edge to wait for), so a block
+that went grey outside trigger mode would grey itself under the very click that started the replay. Trigger
+mode is therefore not one of the conditions. Recording is switched in the **trigger panel**, with the
+trigger's other options rather than in the list whose contents it decides, and it is **off by default** --
+writing every latch to disk is the user's decision about their disk. (The session file keeps nothing about
+this list: there is no remembered width, because there is no pane to size.)
+
+The dock follows every latch; the page's list is refreshed per latch only while it is open, and rebuilt when
+**Manage** opens it. **The dock list has no thumbnails** -- its rows are fixed-width, ellipsized cells, so the
+dock's size does not depend on what an event is called -- the page has them, and its rows are replayed with a
+**double** click, because on the page a single click builds the selection that *Compare* and *Accumulate* read.
 
 Each row carries **its own picture** beside the numbers -- time, duration, signal count, rate -- 138
-pixels wide whatever the pane's width, and as tall as that event's own proportions make it: the
+pixels wide whatever the page's width, and as tall as that event's own proportions make it: the
 preview is one 44-pixel strip per signal, so a one-signal event is a short row and a six-signal one a
 tall one (capped, so that a row stays a row). **Hovering it shows the whole preview at full size**,
 through a tooltip -- the small picture says *which* event this is, the zoom says what is in it.
-Clicking a row **replays** it: the same loader `--open` uses, on the event's file. The click **leaves
+Double-clicking a row on the page (or clicking one in the dock) **replays** it: the same loader `--open` uses, on the event's file. The click **leaves
 the list alone** -- your scroll position and the pair you had selected for a comparison both survive
 it, because marking which event is on screen rewrites the rows' titles in place instead of rebuilding
 the list. **The layout does
@@ -862,7 +882,7 @@ rearrange the user's work under a click. What changes is the data, and the trigg
 file has no edge to wait for. Cursors, measurements, math channels and another export all work on a
 stored event. **Live** takes you back to the stream.
 
-Two buttons under the pane's header read the **selection**:
+Two buttons under the page's header read the **selection**:
 
 - *Compare* puts the two selected events in **two plots side by side**, each with **the layout you
   already have** -- the same rows, the same signals in the same slots, the same colours -- one showing
@@ -880,8 +900,8 @@ model whose channel count no longer matches the stream), the trigger is left, an
 
 A replay leaves trigger mode, because a file has no edge to wait for -- an armed trigger keeps a
 *frozen capture* on screen (`display_mode()` stays `Held` while the model holds a snapshot), and a
-held display draws that snapshot rather than the file that was just loaded. The events pane stays
-open through it: it is available while a stored event is on screen.
+held display draws that snapshot rather than the file that was just loaded. The dock's Events block and
+the `events` page stay available through it: they are available while a stored event is on screen.
 
 **Going back to *Live* puts the live view back as it was.** Opening a file changes two things that
 are not the user's choice -- the time window becomes the event's own span, because that is what
@@ -912,7 +932,7 @@ Three things worth knowing:
 - **The picture is decoded synchronously** (`gdk_texture_new_from_filename()` into
   `gtk_picture_set_paintable()`), not through `gtk_picture_set_filename()`: that path uses the image
   loader's sandboxed, asynchronous decoder, which here fails silently -- the picture keeps a
-  paint-able and an allocation, so a probe reports success while the pane shows an empty rectangle.
+  paint-able and an allocation, so a probe reports success while the page shows an empty rectangle.
   It is then **scaled once to the row's size** and cached by file: a `GtkPicture`'s natural size is
   its content, so the stored 480-pixel preview given straight to a row claims the row's whole width
   and squeezes the text out of it -- a size request on the picture is a floor, not a cap.
@@ -929,8 +949,8 @@ Three things worth knowing:
   one at a time was eight of those, which is most of what made the switch feel slow. They are now
   enabled in a single call, and a rebuild whose plan is unchanged clears the buffer instead of
   allocating a new one.
-- The switch in the pane's header stops the recording; the cache is only useful if it was already
-  on when the event you wanted happened, which is why it defaults to on.
+- The switch in the trigger panel starts and stops the recording; the cache is only useful if it was
+  already on when the event you wanted happened, so switch it on before the event you are waiting for.
 
 ## Opening an exported waveform
 
@@ -1241,8 +1261,10 @@ reveals it, while one whose *pane is collapsed* (the dock, in a window too narro
 plot) is dropped, because it is unmapped with its pane (a hidden page keeps its own *visible* flag
 true -- the widget's own flag, which is all `is_visible()` walks -- so the old test walked over
 surfaces that were not on screen); a narrow window therefore walks five steps. The dock's stop
-describes the readouts and the **Edit** buttons; it names no slide-over page, so it closes the panel
-like the plot's. The cursors and measurements pages have no stop of their own, and their contextual
+describes the readouts and their **Edit** buttons and the Events block under them (one click on a row
+replays it, **Manage** opens the list); it names no slide-over page, so it closes the panel like the plot's.
+There is no stop for the `events` page, so nothing in the tour opens it; the contextual help's
+event-history hint is attached to the page's heading the way the cursors' and measurements' are. The cursors and measurements pages have no stop of their own, and their contextual
 help "(i)" hints are attached to the pages' headings the way the trigger's and the signals' are: they
 are drawn while the page is open, and the dock has none. The chooser's
 tour works the same way over its own two pages: its tabs step shows the *Waveforms* page and
@@ -3021,11 +3043,16 @@ now — they are listed so they are not mistaken for anything else:
   **who** cell ellipsizes at **5 characters** -- the width the 280 px pin allows; with 8 the readouts'
   minimum width was 304 px and widened the sidebar -- and its tooltip carries the full text. (6) The dock and
   the pages **share one string** for every value (see `docs/claude/invariants-readouts-and-cursors.md`).
+  (7) The **Events block and the `events` page are available** while recording is on, a replay is on screen, or
+  the cache holds events -- **not** "outside trigger mode" as the spec has it, because a replay or a group leaves
+  trigger mode and the block would grey itself under the click that started it. (8) The **dock's event rows carry no
+  thumbnails** (the page's do) and show a **hollow marker on the members of the group on screen**, which the page's
+  rows do not. (9) The dock lists at most 24 events; the page lists them all.
   The signal badges, the signal pages and the drag (step 2) are done. Still to come, in the spec's
-  order: the events section in the dock, the deletion of the left pane and the Events toggle, and
-  `USER-GUIDE.md` and the manual's figures (4). Until then the events pane, the header's Events toggle
-  and the manual (which still describes the side panel's tabs, and the cursors and measurements as
-  panes of it) still describe the previous layout.
+  order: `USER-GUIDE.md` and the manual's figures (4). The events section in the dock and the `events`
+  page are done, and the left events pane, its divider and the header's Events toggle are deleted. Until
+  the manual is swept, it (which still describes the side panel's tabs, and the cursors, measurements
+  and events as panes of it) still describes the previous layout.
 - **A plot holds at most eight signals** (`RowSet::kSlotsPerRow`), because a `Plot`
   cannot give up a trace and the slots are therefore created up front.
 - **One toggle drives both cursor orientations.** The vertical markers and the