Live Projectors

[disconcision]

Closes #702, closes #829.
Projector fixes: closes #1359, closes #1396, closes #1436, closes #1489
Hazel selection fixes: closes #1448, , closes #1147, closes #890, closes #880, closes #1351, closes #879
General Hazel fixes: closes #1300, closes #1471, closes #933, closes #1421.

For an introduction, see the Probes slide here: https://hazel.org/build/projectors-live/.

Core Feature: Probe Projectors

• Use the projector panel or alt/option-v to apply a probe projector to any expression or pattern
• When a term is probed, live values will show in cells to the right of the line
• Cells in function bodies are grouped by function call
• If the dynamics are still calculating, or if the expression is never executed, the probe icon will spin
• If closures are found for that expression, the number of closures are indicated in the probe superscript
• By default, live cells based on all gathered closures are shown; double-clicking on any cell restricts that probe to show only that one cell.
• Hovering over a cell of the currently indicated probe which is not a pattern of variable reference shows the (co-contextual) environment (i.e. the values of variables in the expression)
• Clicking on a live cell sets a global (across-probe) closure cursor powered by that closure's call stack. Other cells (across all probes) will get decorated based on their 'relative positions' within the call graph.
• To expand/contract a cell, hold shift and drag left/right to progressively un/summarize the value. This also un/summarizes values in the corresponding environment view.
• Double-clicking on a probe on a function application 'pins' that application; for applications with more than one execution (call), the currently selected call will be pinned
• If there are more than 30 ish cells, the others will be cut off; this is indicated by ellipses to the right. When such a probe is indicated, small arrows will appear to the left of its cells to allow moving the window of cells shown.
• Probes have limited keyboard controls. When a cell is clicked, the group of cells gets keyboard focus. Left/right arrows move the cursor, left/right+shift resizes the cursor cell, and space toggle between show-one and show-all cells.

Probe caveats:

• Adding two probes on the same line will currently result in their cells overlapping; you can manually add a linebreak to work around this

Hazel Features:

• Probe system. Instrumenting a term by wrapping it in as Wrap(term, Probe(options)) will make the evaluator collect information about the term's evaluations. This currently includes a value, a (filtered) environment, and a call stack. The options on Probe are intended to be added during statics, and can refine what is collected. Currently, they are used to filter the collected environment to the values occurring in the expression's co-context.
• Clipboard cache: When you copy or cut a selection corresponding to a complete segment (a selection that contains no shards with corresponding backpack items), the underlying segment is statefully retained. If text is pasted which is identical to the text serialization of that stored segment, the segment is used instead. This speeds up pasting, and also retains projectors.
• Selection improvements: Click-drag now persists when the mouse cursor is outside the browser window. Various other selection fixes and features (see closed issues list)

Projector System Features:

• The projector API has changed. There are new entries, described below, and some of the method arguments have been changed to provide new information to projectors.
• Projectors can get dynamic info in their infopack. This collection is gated by API flag (dynamics=true), for performance reasons. This information consists of a list of 'closures', each of which has a value, a (co-contextual) environment, and a call stack corresponding to one evaluation of the probed term or pattern.
• Can now project all terms and patterns, not just operand-shaped ones. When un-projecting non-operand-shaped projectors, the previously projected syntax will be left selected, as this now enable back-and-forth toggleability thanks to:
• Can now project based on selection (not just indication). Any selection which corresponds to a full term can be projected. Caveats: Term must not be whitespace-padded on the outside. It also suffices that the term be a term in isolation, even if it's not a subterm of the program, e.g. 1+2 in 1+2*3 can be projected. If this makes people too angry I'll consider changing it.
• Projectors are retained on cut/copy/paste (see Clipboard cache below)
• The can_project method and init constant have been merged into a single method (called init). This method must decide, based on a provided term, whether or not the projector applies, and if so, returns the initial model. So the initial model can now be based on the initial syntax.
• Projectors can now specify optional additional views. The offside view renders at the end of the line where the projector is located (the probe projector uses this for value displays). The overlay view renders above all code text, code decoration, and base projector views (the probe projector uses this for closure count and function pin indicators).
• All projectors now have keyboard shortcuts. Alt/option-v for probes, -t for type, and -l for livelit-style projectors (chooses the first that matches the indicated syntax)
• Projector keyboard shortcuts now show when you hover over the projector panel when the relevant projector is active
• Projectors are now provided (through info.utility) with some helper functions for piece/segment/term conversion. These are included to avoid cyclic dependencies, and will hopefully be deprecated by the upcoming Splice Projectors PR.
• Projector view functions are now provided with a view_seg callback to render non-interactive view of syntax. This is currently used for probe projector cell value displays. This will be deprecated by the upcoming Splice Projectors PR.

Updates to existing projectors:

• Fold projector now show their hidden syntax on-hover
• Type projectors now displays at the end of the line like probes. New keyboard shortcut option/alt-t
• All livelit-style projectors now have combined keyboard shortcut option/alt-l

Hazel Internal changes:

• The code completion buffer is now implemented as a comment inside a selection (instead of a tile). This should be noticeable but should be less brittle wrt code changes as remolding/erasure will now never have to care about its shape.

Future work:

These are features that are either cut for time or are waiting on in-pipeline subsystems.

Deferred fixes (will create issues for these):

• Fix: Selection: There is still an outstanding drag-selection bug (which exists on dev but is masked behind selection bugs fixed above). When dragging down specifically, sometimes the selection will get 'stuck' and require the cursor be jostled from side to side. I don't know why.

Upcoming Probe enhancement PR will include:

• Feature: Probes control panel. This will show a legend / configuration panel which can be used to select which decoration overlays are active
• Feature: Probe keyboard control. Move between cells with arrows, add new probes with arrows_modifiers
• Feature: A way to reset the closure cursor. Needs probe keyboard control.
• Feature: Highlight function names corresponding to pinned/indicated calls associate caller/callee more closely
• Feature: Grey out unevaluated probes (current deco arch makes this awkward)
• Feature: Adding a new probe on the same line as another removes the first or induces a linebreak or displays on the next line

Other future features (will create issues for these):

• Feature: Projectors: ExplainThis Override
• Feature: Projectors: Overlay projector option: Don't replace the syntax, but augment it over top. Could be used for probe, test.
• Feature: Projectors: Projectors should somehow indicate strictly contained static errors. Might want to wait on a more general statics mechanism to associate errors which impact a term with that term.
• Feature: Probes: Enable editing of probed expressions (requires splices or overlays)
• Feature: Improve & Extend value/syntax abbreviation logic and add multiline value display (drag down)
• Feature: Refine probe window navigation arrows UI. Integrate arrows with ellipses. Drag ellipses to rescale space available for cells (ideally snapping to cell, not just cut-off)
• Feature: Probes: 'Step' display option showing environment substituted into expression (but not evaluated)

[disconcision]

@Negabinary this is only tenuously out of draft but i'd appreciate a look-over when you get the chance, especially regarding the dynamics bits. Currently I'm doing some things a bit hackily and trying to decide how bad they actually are.

There are two dynamics things in particular I'm wondering about:

1. Regarding exercise mode dynamics work. I was able to hook up to it in a basic sense, but I'm confused about dynamics across cells. Currently, probes in the correct implementation reflect values from the tests, but the user implementation probes don't reflect values from the user tests, only for executions in the user implementation editor itself. Changing the ExercisesMode piping so that the user implementation cell gets the user test dynamics also doesn't seem to make this work. (As a larger issue, it's not 100% clear to me what should get what. Also, I think I/we should consider, possibly as part of this PR, changing the way dynamics output works to make it slightly more in-line with statics. in particular, maybe dynamics should output a map, and there is just implictly a probe on the syntax root. i.e, the previous dynamics results is now 'just' an entry in the map. This clarifies my interfaces a bit but maybe fucks with stepper stuff?)
2. Another thing I don't quite understand is that in the current implementation, i'm not actually really accumulating a stack trace... it seems to be more of a lexical equivalent. i.e. I would have expected the current implementation to, if i instrument a variable x inside a function bound to f, defined at the top level, which is called inside a function g, also defined at the top level, and then g is called at the top level, to have entries corresponding to both f and g for variable x. but it only has the immediate call. this doesn't actually immediately matter but I want to understand why my understanding of the code is wrong.

P.S. I did some dead code removal in ClosureEnvironment as I was finding the interface was drifting a bit

[disconcision]

@cyrus- all issues discussed yesterday have been addressed except as discussed below.

Should be fixed:

• Selection bug: Click, then quickly move and click, and then drag
• Selection bug: Doubleclick, then quickly move and doubleclick (same cause as above)
• Selection bug: Selection dragging isn't triggered by wheel events
• Probe bug: After toggling one vs many cells shown, sometimes less cells are shown than expected
• Probe tweak: Make patterns and expression closure cells the same color
• Probe tweak: Make one vs many cells shown toggle global for all probes
• Internal tweak: Separate Parens and Probe constructors in Exp
• Probe data leak: Probe data is now removed from the test validation editor. (What was the specific correct implementation exfiltration you found with probes here? i can't immediately recreate one.)

Won't fix but will file issues:

• Need to juggle the mouse when dragging out-of-bounds: looks like it would require a new stateful subsystem.
• Make selections include linebreaks to improve selection behavior around end of lines (a lot of work and careful testing)