@variance-authority/core 0.1.1 → 0.3.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +245 -10
- package/README.md +279 -110
- package/dist/attribute/call-site.d.ts +20 -2
- package/dist/attribute/call-site.js +17 -20
- package/dist/attribute/call-site.js.map +1 -1
- package/dist/attribute/closure.d.ts +105 -0
- package/dist/attribute/closure.js +180 -0
- package/dist/attribute/closure.js.map +1 -0
- package/dist/attribute/composition.d.ts +5 -3
- package/dist/attribute/composition.js +67 -11
- package/dist/attribute/composition.js.map +1 -1
- package/dist/attribute/divergence.js +7 -1
- package/dist/attribute/divergence.js.map +1 -1
- package/dist/attribute/index.d.ts +7 -1
- package/dist/attribute/index.js +4 -1
- package/dist/attribute/index.js.map +1 -1
- package/dist/attribute/landmark.d.ts +97 -0
- package/dist/attribute/landmark.js +144 -0
- package/dist/attribute/landmark.js.map +1 -0
- package/dist/attribute/lexicon.d.ts +112 -0
- package/dist/attribute/lexicon.js +210 -0
- package/dist/attribute/lexicon.js.map +1 -0
- package/dist/attribute/source.d.ts +19 -2
- package/dist/attribute/source.js +12 -0
- package/dist/attribute/source.js.map +1 -1
- package/dist/attribute/stack.d.ts +62 -1
- package/dist/attribute/stack.js +116 -5
- package/dist/attribute/stack.js.map +1 -1
- package/dist/compare/derive.d.ts +17 -0
- package/dist/compare/derive.js.map +1 -1
- package/dist/compare/diff/index.d.ts +25 -1
- package/dist/compare/diff/index.js +8 -5
- package/dist/compare/diff/index.js.map +1 -1
- package/dist/compare/explain.d.ts +1 -1
- package/dist/compare/explain.js +15 -8
- package/dist/compare/explain.js.map +1 -1
- package/dist/compare/index.d.ts +2 -2
- package/dist/compare/index.js +1 -1
- package/dist/compare/index.js.map +1 -1
- package/dist/compare/observability.d.ts +25 -1
- package/dist/compare/observability.js +54 -0
- package/dist/compare/observability.js.map +1 -1
- package/dist/compare/parting.d.ts +24 -2
- package/dist/compare/parting.js +14 -9
- package/dist/compare/parting.js.map +1 -1
- package/dist/compare/slice.d.ts +15 -2
- package/dist/compare/slice.js +6 -2
- package/dist/compare/slice.js.map +1 -1
- package/dist/format/capture.d.ts +15 -0
- package/dist/format/capture.js +30 -1
- package/dist/format/capture.js.map +1 -1
- package/dist/format/document.d.ts +65 -5
- package/dist/format/document.js +34 -0
- package/dist/format/document.js.map +1 -1
- package/dist/format/hash.d.ts +26 -0
- package/dist/format/hash.js +41 -0
- package/dist/format/hash.js.map +1 -1
- package/dist/format/index.d.ts +4 -3
- package/dist/format/index.js +3 -2
- package/dist/format/index.js.map +1 -1
- package/dist/format/sha256.d.ts +10 -2
- package/dist/format/sha256.js +42 -9
- package/dist/format/sha256.js.map +1 -1
- package/dist/format/stabilize.d.ts +15 -0
- package/dist/format/stabilize.js +22 -1
- package/dist/format/stabilize.js.map +1 -1
- package/dist/index.d.ts +3 -11
- package/dist/index.js +3 -11
- package/dist/index.js.map +1 -1
- package/dist/judge/inspect.js +57 -4
- package/dist/judge/inspect.js.map +1 -1
- package/dist/relate/before.d.ts +137 -0
- package/dist/relate/before.js +155 -0
- package/dist/relate/before.js.map +1 -0
- package/dist/relate/graph.d.ts +53 -16
- package/dist/relate/graph.js +33 -15
- package/dist/relate/graph.js.map +1 -1
- package/dist/relate/index.d.ts +3 -2
- package/dist/relate/index.js +2 -1
- package/dist/relate/index.js.map +1 -1
- package/dist/relate/merkle.d.ts +14 -5
- package/dist/relate/merkle.js +14 -6
- package/dist/relate/merkle.js.map +1 -1
- package/dist/relate/reach.d.ts +19 -5
- package/dist/relate/reach.js +10 -8
- package/dist/relate/reach.js.map +1 -1
- package/dist/relate/records.d.ts +95 -4
- package/dist/relate/records.js +93 -13
- package/dist/relate/records.js.map +1 -1
- package/dist/segment/index.d.ts +123 -0
- package/dist/segment/index.js +265 -0
- package/dist/segment/index.js.map +1 -0
- package/dist/share/index.d.ts +148 -0
- package/dist/share/index.js +190 -0
- package/dist/share/index.js.map +1 -0
- package/package.json +10 -2
package/CHANGELOG.md
CHANGED
|
@@ -1,13 +1,248 @@
|
|
|
1
1
|
# @variance-authority/core
|
|
2
2
|
|
|
3
|
+
## 0.3.0
|
|
4
|
+
|
|
5
|
+
### Minor Changes
|
|
6
|
+
|
|
7
|
+
- 05d6683: Changes before and beyond reach
|
|
8
|
+
|
|
9
|
+
A run reads left to right: the harness starts it, the tests enter your code,
|
|
10
|
+
your code goes out into the install and never comes back. Selection lives in the
|
|
11
|
+
middle, and both ends were invisible for opposite reasons.
|
|
12
|
+
|
|
13
|
+
The far right already arrived — a package is a node, a bump is a seed, the same
|
|
14
|
+
backwards walk answers it. The far left is this. Nothing imports a
|
|
15
|
+
`vitest.config.ts`, a setup module, a CI workflow or a `.nvmrc`, so no walk
|
|
16
|
+
reaches one and the honest structural answer about a change to one is *no
|
|
17
|
+
component moved*: a skipped suite over the file that decides how every test in
|
|
18
|
+
it runs. A diff that was *wholly* outside the graph already widened. The hole
|
|
19
|
+
was a config edited beside an ordinary source file, where the walk had a seed
|
|
20
|
+
and answered confidently about a change it never looked at.
|
|
21
|
+
|
|
22
|
+
`source.before` names those files, repository-root-relative, and a directory
|
|
23
|
+
claims everything under it. What a declaration buys beyond its own name is
|
|
24
|
+
everything below it: `beforeReach` in `@variance-authority/core/relate` walks
|
|
25
|
+
*along* the arrows from each entry — the one question whose subject has no
|
|
26
|
+
dependents — and collects the setup module, the fixture only that setup
|
|
27
|
+
imports, and the packages the environment rests on. The descent stops at the
|
|
28
|
+
first file `source.dirs` already covers, because that file has dependents and a
|
|
29
|
+
change to it is answered exactly by walking them; everything below it is
|
|
30
|
+
reached through it and does not arrive either.
|
|
31
|
+
|
|
32
|
+
The two ends meet there. A `jsdom` bump is named by the install comparison,
|
|
33
|
+
reaches `jest-environment-jsdom`, and reaches a config no file in the
|
|
34
|
+
repository imports — a change beyond reach arriving before it.
|
|
35
|
+
|
|
36
|
+
`scanRelations` takes `before` to seed those paths by name, since a harness
|
|
37
|
+
lives above every directory a component scan is pointed at. A named path that
|
|
38
|
+
is absent or has no reader is dropped rather than recorded unreadable: an
|
|
39
|
+
unknown file seeds every walk forever, so a typo would otherwise widen every
|
|
40
|
+
run in the repository. A declared entry the graph does not hold contributes
|
|
41
|
+
only its own name, which is the whole answer for a `.nvmrc` and a symptom for a
|
|
42
|
+
harness config, so the run reports it as a note rather than guessing.
|
|
43
|
+
|
|
44
|
+
`source.before` requires `source.relations: true`.
|
|
45
|
+
|
|
46
|
+
## 0.2.0
|
|
47
|
+
|
|
48
|
+
### Minor Changes
|
|
49
|
+
|
|
50
|
+
- e546e21: A collector's complaints travel into the observation
|
|
51
|
+
|
|
52
|
+
A finding produced at collection reached the CLI's record and nobody else. A
|
|
53
|
+
suite driving this from its own runner asks the observation what is wrong with a
|
|
54
|
+
subject, and the one diagnostic that explained the moved pixels — the host, not
|
|
55
|
+
the stylesheet, chose the typeface — was produced, carried, and dropped one call
|
|
56
|
+
short of the person reading the failure.
|
|
57
|
+
|
|
58
|
+
`mergeDiagnostics` dedupes on every field, so the CLI, which now meets the same
|
|
59
|
+
list twice, says a shared complaint once.
|
|
60
|
+
- f4b328b: Mainline's evaluation can travel, so the next machine does not derive it again
|
|
61
|
+
|
|
62
|
+
A suite index is bytes addressed by the commit they were written at, which is
|
|
63
|
+
exactly the shape a cache already wants. What was missing was anywhere to put
|
|
64
|
+
them: every checkout that wanted to know what `main` looked like had to run
|
|
65
|
+
`main`, and in CI that is the whole suite, twice, on every pull request.
|
|
66
|
+
|
|
67
|
+
`@variance-authority/core/share` is the transport, and it has two backends
|
|
68
|
+
because there are two. A **directory** is what `actions/cache`, `aws s3 sync`, a
|
|
69
|
+
network mount and a laptop all are once the bytes are on disk; an **HTTP
|
|
70
|
+
endpoint** is what a bucket, a presigned URL and a tribunal deployment all are.
|
|
71
|
+
`createDirectoryShare` in `@variance-authority/store/share` is the filesystem
|
|
72
|
+
half, kept out of `core` because `core` takes no platform.
|
|
73
|
+
|
|
74
|
+
A share never fails a run. `get` answers `null` and `put` resolves, whatever the
|
|
75
|
+
network did — a share that is down, misconfigured or empty is indistinguishable
|
|
76
|
+
from a cold one, and the run derives its own index and continues. The cost is
|
|
77
|
+
deliberate: a typo in an endpoint is a suite that quietly got slower rather than
|
|
78
|
+
a build that went red, and the `behind` count `variance share` prints is the
|
|
79
|
+
signal that publishing has stopped.
|
|
80
|
+
|
|
81
|
+
The CLI gained a `share` section in its config, a `variance share` command that
|
|
82
|
+
says what the share holds for mainline or publishes what this run derived, and a
|
|
83
|
+
line on `variance run` naming where the index went. A lookup walks the lineage —
|
|
84
|
+
`merge-base` with the configured `mainline`, then first-parent — and asks this
|
|
85
|
+
machine's own cache for every commit before it asks the share, because the run
|
|
86
|
+
that just published is usually the one asking.
|
|
87
|
+
|
|
88
|
+
`docs/sharing.md` is the arrangement end to end, GitHub Actions first.
|
|
89
|
+
- e546e21: A subject is allowed to reference something that is not there
|
|
90
|
+
|
|
91
|
+
A fixture that points an `<img>` at a path nobody serves is testing the fallback,
|
|
92
|
+
and the broken state is the subject. Capture had no way to say so: the resolver
|
|
93
|
+
returned bytes or the capture was refused, so Material UI's CardMedia,
|
|
94
|
+
ImageListItem and Avatar suites — which all reference a deliberately absent
|
|
95
|
+
`/fake.png` — could not be captured at all.
|
|
96
|
+
|
|
97
|
+
`resolveResource` may answer `{ absent: true }`. The resource is recorded with
|
|
98
|
+
its status, and the renderer answers that status instead of counting the request
|
|
99
|
+
as one the document failed to carry.
|
|
100
|
+
- e546e21: A subject with no pixels is a baseline without an image, not a subject that got away
|
|
101
|
+
|
|
102
|
+
A wrapper whose only child went to a portal, or a conformance mount with no
|
|
103
|
+
children, occupies nothing. Refusing to photograph it is right; refusing the
|
|
104
|
+
subject was not. On Material UI's unit tier that was 1109 of 4371 subjects
|
|
105
|
+
reported as unobserved while the capture held their markup, their rules, their
|
|
106
|
+
component hashes and their accessibility tree — none of which was in doubt.
|
|
107
|
+
|
|
108
|
+
`Raster` makes the image optional: `bytes`, `width` and `height` are absent
|
|
109
|
+
together or present together, and `pictured` is the one place that narrows all
|
|
110
|
+
three. Absent means *this subject has no pixels*, which is a measurement — it
|
|
111
|
+
never means the image was lost. `occupiesPixels` asks the same question of a
|
|
112
|
+
record read without bytes, which is the only form a sidecar takes. `observe`
|
|
113
|
+
gets a second tier in `unpictured.ts`, where the comparison such a subject can
|
|
114
|
+
still take — document digest, component hashes, accessibility — is the whole
|
|
115
|
+
verdict. `promotionOf` promotes the sidecar alone when there is no `after`,
|
|
116
|
+
because the subject reached a verdict and the only missing half is the one a
|
|
117
|
+
camera would have produced.
|
|
118
|
+
|
|
119
|
+
The file-backed store carries the same nullable pair, and with it the split of a
|
|
120
|
+
baseline's two halves into two path prefixes. `identities` scans the record root,
|
|
121
|
+
because a subject with no pixels has no image directory to be found in and the
|
|
122
|
+
sibling scan would otherwise call another machine's baseline new.
|
|
123
|
+
|
|
124
|
+
Capture stops handing this to Playwright to fail on. Both screenshot paths used
|
|
125
|
+
to refuse a zero-area subject in terms of their own arguments — the clip path
|
|
126
|
+
with `Expected options.clip.height to be greater than 0`, the element path by
|
|
127
|
+
spending the full actionability timeout and then complaining about visibility —
|
|
128
|
+
so a reader had a component that rendered nothing and a sentence about a
|
|
129
|
+
rectangle. `captureSubject` decides it now, and the renderer names the subject.
|
|
130
|
+
|
|
131
|
+
**Operators:** this is schema **18**. `baselines` and `render_cache` drop
|
|
132
|
+
`NOT NULL` from `width` and `height`, shipped as `0016_pixel-less-baselines.sql`.
|
|
133
|
+
Apply it before pointing a CLI of this version at the deployment; `GET /version`
|
|
134
|
+
reports the schema a build expects.
|
|
135
|
+
- f4b328b: A run leaves behind what its suite looked like, in bytes another checkout can read
|
|
136
|
+
|
|
137
|
+
A run report answers about *this run*. Nothing answered about the suite from
|
|
138
|
+
somewhere else: which components exist, which subjects hold them, which of them
|
|
139
|
+
has an example of its own, and every name the run indexed. A branch asking any
|
|
140
|
+
of that against `main` had nothing to ask, because `main`'s answers were computed
|
|
141
|
+
on a machine that has since gone away — which is why *this component is not
|
|
142
|
+
covered anywhere* was a question with no reader, and why a second run could only
|
|
143
|
+
re-derive what the first already knew.
|
|
144
|
+
|
|
145
|
+
`@variance-authority/report/suite-index` is that artifact. `suiteIndexOf` takes
|
|
146
|
+
the baseline-bearing half of a report — the census, the subject denominator it is
|
|
147
|
+
counted against, and the lexicon — and `encodeSuiteIndex` writes it as one
|
|
148
|
+
segment, with `readSuiteIndex` and `writeSuiteIndex` beside the report's own file
|
|
149
|
+
functions. The other half of a composition does not travel: `movements` is what
|
|
150
|
+
moved since a comparison nobody else made, and `divergences` and `echoes` are
|
|
151
|
+
readings of one commit's snapshots. None of the three is a fact about the suite.
|
|
152
|
+
|
|
153
|
+
It carries the commit it was written at and nothing else about where it is, which
|
|
154
|
+
is the rule the selection index already settled — a commit answers *what changed
|
|
155
|
+
since this was written* exactly, and a timestamp says when a machine was rather
|
|
156
|
+
than where a tree was. An index that cannot name itself is the honest record of a
|
|
157
|
+
run that could not, and a reader holding one has nothing to diff.
|
|
158
|
+
|
|
159
|
+
It is not JSON. A lexicon is the same few thousand strings written once per
|
|
160
|
+
subject that holds them, and in JSON the repetition *is* the payload. Interned
|
|
161
|
+
once and referenced by number it stops being one, and equal facts encode to equal
|
|
162
|
+
bytes — a sorted dictionary, so a cache that keys on content actually hits.
|
|
163
|
+
|
|
164
|
+
`@variance-authority/core/segment` is the arithmetic underneath: named columns,
|
|
165
|
+
alignment, interned strings, and the validation a decode performs before it
|
|
166
|
+
believes a file. A column's width comes from the array's own type, so a mismatch
|
|
167
|
+
is a compile error rather than a decode against the wrong stride, and a malformed
|
|
168
|
+
reference rejects the whole file — a segment is a cache or a baseline, both of
|
|
169
|
+
which may be rebuilt, and half of either is worse than neither. The source index
|
|
170
|
+
now reads and writes through it and keeps its own bytes; what it still owns is the
|
|
171
|
+
part that is about source.
|
|
172
|
+
- e546e21: Two readings of two places are not a flake
|
|
173
|
+
|
|
174
|
+
`flake` is an accusation — *this reading is not repeatable* — and it was being
|
|
175
|
+
made about a comparison that never claimed repeatability. Lifting two instances
|
|
176
|
+
out of two subjects at one commit and parting them produces a case where every
|
|
177
|
+
input agrees, the tree holds, and the output moved; the old table had exactly one
|
|
178
|
+
row for that.
|
|
179
|
+
|
|
180
|
+
Where a component sits is decided by the boxes around it, and no component
|
|
181
|
+
receives its own position as a prop. Two instances that agreed on everything they
|
|
182
|
+
were handed and landed at different coordinates have contradicted nothing.
|
|
183
|
+
|
|
184
|
+
`partingOf` takes a `PartingPlace`: `same` for one subject read twice, across
|
|
185
|
+
revisions or across two moments of one scenario, and `elsewhere` for two
|
|
186
|
+
instances at one commit. It decides only between `flake` and the new `placed`
|
|
187
|
+
rung and slice, and it defaults to `same`, which is what every comparison across
|
|
188
|
+
revisions is. `divergencesOf` is the one caller that says `elsewhere`. The
|
|
189
|
+
sentence points outwards, at what put the component there, rather than at the
|
|
190
|
+
component.
|
|
191
|
+
|
|
192
|
+
### Patch Changes
|
|
193
|
+
|
|
194
|
+
- e546e21: Four ways a large unit tier lost subjects, none of which said so
|
|
195
|
+
|
|
196
|
+
**The styling was gone before the capture looked.** A capture taken from a setup
|
|
197
|
+
file runs in the outermost `afterEach` a run has; every hook registered inside a
|
|
198
|
+
`describe` has already finished, and `onTestFinished` runs later still. CSS-in-JS
|
|
199
|
+
teardown lives in exactly those inner hooks — emotion's test renderer removes each
|
|
200
|
+
`<style>` tag it inserted — so the capture read the page with the styling taken
|
|
201
|
+
back off it. The class names were all still in the markup, they matched nothing,
|
|
202
|
+
and every subject captured, compared and passed against a photograph of unstyled
|
|
203
|
+
DOM. Measured on Material UI's unit tier: 399 of 400 captures carried no CSS at
|
|
204
|
+
all, and 1201 subjects laid out to zero height because nothing was sizing them.
|
|
205
|
+
`retainStyles` records style elements as they are inserted and puts the removed
|
|
206
|
+
ones back, in insertion order, for the length of one capture. The next test still
|
|
207
|
+
gets a clean page.
|
|
208
|
+
|
|
209
|
+
**A subject id can be longer than a filename.** A suite that names subjects after
|
|
210
|
+
the test that produced them — a file path and a full test name — passes 255 bytes
|
|
211
|
+
on ordinary tests, and the point of that convention is that the id says where the
|
|
212
|
+
subject came from. `fileNameFor` in core is now the one rule: a readable prefix,
|
|
213
|
+
plus a digest of the whole id when the id does not fit, because truncation alone
|
|
214
|
+
merges two tests into one file. The baseline store, the capture archive, the
|
|
215
|
+
CLI's image directory, the Playwright evidence directory and the Eyes journal all
|
|
216
|
+
use it; before this, each was one long subject id away from a raw `ENAMETOOLONG`
|
|
217
|
+
that never mentions a subject. `@variance-authority/eyes` declares
|
|
218
|
+
`@variance-authority/core` directly rather than reaching it through `react` — the
|
|
219
|
+
host stacks it keeps optional are Playwright and Testing Library, and a filename
|
|
220
|
+
rule is not one of them.
|
|
221
|
+
|
|
222
|
+
**An inline `url()` is spelled in entities.** `style="background-image:url("/a.png")"`
|
|
223
|
+
reaches the scan through HTML serialization, and reading it literally asked the
|
|
224
|
+
caller for bytes at a URL that exists nowhere but in the escaping. The HTML half
|
|
225
|
+
is scanned with the attribute's entities undone; the stylesheet half, which was
|
|
226
|
+
never escaped, is scanned as before.
|
|
227
|
+
|
|
228
|
+
**A stubbed `getComputedStyle` is not a broken asset scan.** Replacing
|
|
229
|
+
`window.getComputedStyle` with a map of the two properties a component reads is
|
|
230
|
+
the only way to drive some layouts in jsdom. The scan called `getPropertyValue`
|
|
231
|
+
on the plain object it got back, and every test in the file died on a `TypeError`
|
|
232
|
+
raised four frames below anything you wrote. There is nothing to scan and nothing
|
|
233
|
+
to complain about — whatever those styles would have named, the stub already
|
|
234
|
+
removed from the page. The attributes on the element are still read.
|
|
235
|
+
|
|
3
236
|
## 0.1.1
|
|
4
237
|
|
|
238
|
+
Lockstep release — nothing in this package changed. Every `@variance-authority/*` package shares one version.
|
|
239
|
+
|
|
5
240
|
## 0.1.0
|
|
6
241
|
|
|
7
242
|
### Minor Changes
|
|
8
243
|
|
|
9
244
|
- 5c34e6d: Say which input moved, not just which tag did
|
|
10
|
-
|
|
245
|
+
|
|
11
246
|
`partingOf(baseline, candidate)` climbs from a set of deltas to the component
|
|
12
247
|
boundary that owns them and names the input that carried the decision:
|
|
13
248
|
`Cart chose differently — useState #0 moved`, with `Summary` reported as having
|
|
@@ -16,7 +251,7 @@
|
|
|
16
251
|
as lines. Two runs of one page that differ in a `<p>` where the other has a
|
|
17
252
|
`<span>` previously produced a structural delta and nothing to say about it;
|
|
18
253
|
the tag is now the symptom and the hook cell is the report.
|
|
19
|
-
|
|
254
|
+
|
|
20
255
|
`holdingOf` from `@variance-authority/react` supplies the evidence — props,
|
|
21
256
|
context values and hook cells as digests, per boundary, including
|
|
22
257
|
`useSyncExternalStore` snapshots so a store that moved outside React is
|
|
@@ -28,28 +263,28 @@
|
|
|
28
263
|
caller-supplied reader, opt in separately from `wiringOf`, and `capture` in
|
|
29
264
|
`@variance-authority/unit-test` now accepts `provenanceOf`, `wiringOf` and
|
|
30
265
|
`holdingOf` so a unit test can read one.
|
|
31
|
-
|
|
266
|
+
|
|
32
267
|
A boundary whose own input could not be read is reported as `unread` and never
|
|
33
268
|
as nondeterminism: `undetermined` is reserved for a component whose every input
|
|
34
269
|
was read and agreed. A wrapper that roots a component boundary no longer
|
|
35
270
|
collapses, because collapsing it discarded the holding — the cost is that a run
|
|
36
271
|
reading holdings keeps wrappers a run without them removes, which is why both
|
|
37
272
|
sides of a comparison must be read the same way.
|
|
38
|
-
|
|
273
|
+
|
|
39
274
|
A component that ran no hooks is read as having run none, rather than as
|
|
40
275
|
unreadable. React writes `_debugHookTypes = null` on every fiber and fills it on
|
|
41
276
|
the first hook call, so the property's absence is the only silence — and
|
|
42
277
|
collapsing the two reported the one shape most worth calling nondeterministic, a
|
|
43
278
|
component with no props and no hooks that renders differently twice, as
|
|
44
279
|
something nothing could be said about.
|
|
45
|
-
|
|
280
|
+
|
|
46
281
|
`PartedBoundary.moved` names the properties the owned deltas named — `color`,
|
|
47
282
|
`padding-top`, `width` — and `explainParting` spends them on the delta line.
|
|
48
283
|
That is the last joint of the chain the rungs climb: a hook cell moved, a prop
|
|
49
284
|
carried it down, and this is what the prop turned into on the page. A count and
|
|
50
285
|
a band stop one link short of what somebody chasing a visual regression is
|
|
51
286
|
trying to name.
|
|
52
|
-
|
|
287
|
+
|
|
53
288
|
`Parting.slice` answers the question asked before which input moved: is this
|
|
54
289
|
worth opening. Three facts are read independently — did the component tree
|
|
55
290
|
move, did any input move, did the output move — and the combinations collapse
|
|
@@ -58,24 +293,24 @@
|
|
|
58
293
|
since a pixel differ can say the screenshots match and nothing about what was
|
|
59
294
|
rewritten underneath. `flake` is `settled`'s opposite number and is refused on
|
|
60
295
|
silence — a run that read no boundary reports `unread`, never `flake`.
|
|
61
|
-
|
|
296
|
+
|
|
62
297
|
`explainParting` leads with that line, and stops enumerating manifestations
|
|
63
298
|
past three: one input at a fork can put a boundary on every component beneath
|
|
64
299
|
it, and nine lines carrying one decision bury the one line worth reading.
|
|
65
|
-
|
|
300
|
+
|
|
66
301
|
The whole layer is documented in `docs/parting.md`: the six slices, the seven
|
|
67
302
|
rungs, what a holding carries and what it deliberately does not.
|
|
68
303
|
|
|
69
304
|
### Patch Changes
|
|
70
305
|
|
|
71
306
|
- e8fee66: Refuse a value whose state is not in its own enumerable keys.
|
|
72
|
-
|
|
307
|
+
|
|
73
308
|
`shapeValue` documented a `Date` as refused and encoded it instead. `typeof`
|
|
74
309
|
answers `object`, `Object.entries` answers empty, and the canonical text came
|
|
75
310
|
out `{}` — so a snapshot holding a timestamp addressed to the digest of an empty
|
|
76
311
|
object, and a later run comparing a different timestamp reported unchanged. The
|
|
77
312
|
same hole swallowed `Map`, `Set`, `URL`, and `RegExp`.
|
|
78
|
-
|
|
313
|
+
|
|
79
314
|
An object with no own enumerable keys and a prototype other than `Object`'s is
|
|
80
315
|
now refused by name at its pointer. A genuinely empty object is still a value,
|
|
81
316
|
and an instance carrying its own fields still serializes them.
|