studio-engine 0.53.1 → 0.54.1
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.
- checksums.yaml +4 -4
- data/CHANGELOG.md +71 -0
- data/app/assets/tailwind/studio_engine/engine-motion.css +46 -9
- data/app/views/studio/fields/_date_of_birth.html.erb +80 -0
- data/app/views/studio/modals/blocks/_age_verify.html.erb +5 -29
- data/app/views/studio/profiles/_birthday_fields.html.erb +39 -186
- data/app/views/studio/profiles/_birthday_picker_script.html.erb +115 -164
- data/app/views/style/_tricks.html.erb +13 -3
- data/lib/studio/version.rb +1 -1
- metadata +2 -1
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: b79896d29716fe1408c39d86c9e2d19d4e6dc12dd28749ff85054c29e5ba9a77
|
|
4
|
+
data.tar.gz: 033a9a7e0e4799ef5ed0522e3fb6b6d1652ab4239d7de625add1e24ac81a5386
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: dcd13075e85ee744c86ec901f62e749320f916a72c7c04b1530a5715dcdbcec3adcaa39dc1b8236327bc5a2de4238ef1585cf42853444ff92eeae0ff7cfcc702
|
|
7
|
+
data.tar.gz: 6d242c947ec1d582202bb53652091f6ab4ed111cadfa721aab578df1f89f7b8b9a9bdb4bcd9f57b5b8a15ba9d8c9200b2358668ec2058d59f0c210072e93a79a
|
data/CHANGELOG.md
CHANGED
|
@@ -4,6 +4,77 @@ The format is [Keep a Changelog](https://keepachangelog.com/en/1.1.0/). This pro
|
|
|
4
4
|
|
|
5
5
|
## Unreleased
|
|
6
6
|
|
|
7
|
+
### Fixed
|
|
8
|
+
|
|
9
|
+
- **`.conic-surface` keeps its wash inside its own box.** The wash was an
|
|
10
|
+
OVERSIZED pseudo (`inset: -50%`) spun by a `transform` keyframe, cut back to
|
|
11
|
+
the element by the host's `overflow: hidden`. That clip is not dependable: the
|
|
12
|
+
rotating child is composited, and a `backdrop-filter` in its compositing path
|
|
13
|
+
drops it. The `.surface-glass` specimen — a backdrop-filter child of the wash
|
|
14
|
+
itself — is exactly that path, and on admin/style in Chromium the wash escaped
|
|
15
|
+
as a rotated slab across BOTH conic specimens.
|
|
16
|
+
|
|
17
|
+
The element no longer moves. The pseudo sits at `inset: 0` with
|
|
18
|
+
`border-radius: inherit`, and the sweep animates a registered
|
|
19
|
+
`--conic-surface-angle` inside the gradient — the technique §9b already used.
|
|
20
|
+
Nothing extends past the box, the rounded corners come from the pseudo, and no
|
|
21
|
+
clip has to hold. Same picture, one fewer thing that can fail.
|
|
22
|
+
`overflow: hidden` stays for the host's CONTENT.
|
|
23
|
+
|
|
24
|
+
The trade, stated: a registered-property animation repaints on the main thread
|
|
25
|
+
where a transform rode the compositor. At the default 16s sweep that is cheap,
|
|
26
|
+
and §9b already pays it at 4s.
|
|
27
|
+
|
|
28
|
+
- **`.studio-team-glow` is demoed and documented as the wrapper it is.** Its two
|
|
29
|
+
wedge layers are pseudos at z-index -1/-2, and CSS paints an element's own
|
|
30
|
+
background BEFORE its negative-z descendants — so a background on the glow HOST
|
|
31
|
+
is covered BY the wedges, not the other way round, and the ring becomes a wash
|
|
32
|
+
across the whole card face. What covers the middle is an opaque CHILD, which
|
|
33
|
+
paints later; that is TM's `holo-wrap` / `holo-card` split, and what both real
|
|
34
|
+
consumers already ship.
|
|
35
|
+
|
|
36
|
+
The §9b comment claimed the opposite ("host needs an opaque background"), and
|
|
37
|
+
the Effects specimen followed it, so the style guide showed two washed slabs
|
|
38
|
+
where the selection ring belongs. The specimen, its copyable snippet, and the
|
|
39
|
+
contract comment now all state the same thing. The primitive's geometry is
|
|
40
|
+
unchanged — the consumers tuned against it are untouched.
|
|
41
|
+
|
|
42
|
+
### Added
|
|
43
|
+
|
|
44
|
+
- **`studio/fields/_date_of_birth`** — the engine's one date-of-birth field,
|
|
45
|
+
three selects (Month / Day / Year). Rendered by
|
|
46
|
+
`studio/modals/blocks/_age_verify` and by `/profile/edit`'s birthday row.
|
|
47
|
+
|
|
48
|
+
**It was already duplicated**, which is the real argument for extracting it:
|
|
49
|
+
those selects lived in the age-verify block, and turf-monster forked a second
|
|
50
|
+
copy into its own `modals/_age_verify`. Two copies of one field is how they
|
|
51
|
+
drift.
|
|
52
|
+
|
|
53
|
+
### Changed
|
|
54
|
+
|
|
55
|
+
- **`/profile/edit`'s birthday drops the calendar popover** for that shared
|
|
56
|
+
field (operator's call). The calendar shipped **four defects in three days**:
|
|
57
|
+
its day grid used `grid-cols-7`, which exists in no consumer bundle, so the
|
|
58
|
+
weekday letters stacked into one column and it grew to the height of the page;
|
|
59
|
+
a flipped popover positioned from a hardcoded 340px height when the real one
|
|
60
|
+
is ~245px; `place()` re-decided the side on every scroll so it snapped
|
|
61
|
+
mid-scroll; and it came back open after a browser Back.
|
|
62
|
+
|
|
63
|
+
Three selects have none of those failure modes — no position, no flip, no
|
|
64
|
+
popover, no scroll listener. Their utilities are also **proven across the
|
|
65
|
+
fleet**: `grid-cols-3` and `label-upper` are present in all three consumers'
|
|
66
|
+
compiled bundles, which `grid-cols-7` never was.
|
|
67
|
+
|
|
68
|
+
The **no-JavaScript path is unchanged**: the native `<input type="date">` still
|
|
69
|
+
renders under `template x-if`, so only one branch is ever in the DOM.
|
|
70
|
+
|
|
71
|
+
Nine browser specs retired with the popover — all of them about positioning —
|
|
72
|
+
and six replaced them, testing where a select-based field can actually be
|
|
73
|
+
wrong: the day list following the month (30 in April, 29 in a leap February),
|
|
74
|
+
clearing a day the new month lost, refusing a partial date, and a year list
|
|
75
|
+
that stops at this year.
|
|
76
|
+
|
|
77
|
+
|
|
7
78
|
### Changed
|
|
8
79
|
|
|
9
80
|
- **Each `/profile` row is its own card.** They were rendered inside one merged
|
|
@@ -521,10 +521,22 @@
|
|
|
521
521
|
transition, so a consumer can fade the glow in/out (the living style guide
|
|
522
522
|
slides it between step cards as a modal's step machine advances).
|
|
523
523
|
|
|
524
|
-
|
|
525
|
-
|
|
526
|
-
|
|
527
|
-
|
|
524
|
+
THE HOST IS A BARE WRAPPER, NOT THE CARD. Both wedge layers are pseudo-elements
|
|
525
|
+
at z-index -1 / -2, and CSS paints an element's own background FIRST, then its
|
|
526
|
+
negative-z descendants on top of it (CSS 2.1 Appendix E, steps 1 and 2). So a
|
|
527
|
+
background on the HOST cannot cover the wedges — they wash its whole face, and
|
|
528
|
+
the ring is gone. What covers the middle is an opaque CHILD in front: in-flow
|
|
529
|
+
descendants paint later (step 3), leaving only the thickness-wide band around
|
|
530
|
+
the edge. That is TM's holo-wrap / holo-card split, and it is what both
|
|
531
|
+
consumers ship — the engine's style/_modal_specimen (bare glow wrapper + the
|
|
532
|
+
opaque .card article) and the hub's release-phase-meter (bare host + the bar's
|
|
533
|
+
own fill). Demoing it the other way is what shipped a washed card to
|
|
534
|
+
admin/style; the host also must NOT be clipped by an overflow-hidden ancestor
|
|
535
|
+
near its border.
|
|
536
|
+
|
|
537
|
+
Usage: <div class="studio-team-glow rounded-xl">
|
|
538
|
+
<div class="rounded-xl bg-surface p-5">Selected</div>
|
|
539
|
+
</div>
|
|
528
540
|
red glow: style="--studio-team-glow-color: #E11D48"
|
|
529
541
|
Tune: --studio-team-glow-color / -opacity / -speed / -thickness / -bloom
|
|
530
542
|
========================================================================== */
|
|
@@ -604,12 +616,36 @@
|
|
|
604
616
|
11. conic-surface — slow rotating conic-gradient wash (authored fresh)
|
|
605
617
|
A subtle animated ambient background for hero panels / empty states: a conic
|
|
606
618
|
gradient sweeps between two themed stops (CTA -> surface by default) on a slow
|
|
607
|
-
rotation. The gradient lives on a ::before
|
|
608
|
-
it stays crisp.
|
|
619
|
+
rotation. The gradient lives on a ::before behind the content, so content kept
|
|
620
|
+
above it stays crisp.
|
|
621
|
+
|
|
622
|
+
THE WASH IS CONTAINED BY CONSTRUCTION, NOT BY A CLIP. The first cut rotated an
|
|
623
|
+
OVERSIZED child (`inset: -50%` + `transform: rotate`) and leaned on the host's
|
|
624
|
+
`overflow: hidden` to cut it back to the box. That clip is not dependable: the
|
|
625
|
+
rotating child is composited, and a `backdrop-filter` anywhere in its
|
|
626
|
+
compositing path (its OWN .surface-glass child on the style guide, or an
|
|
627
|
+
unrelated overlay elsewhere on the page) drops the clip — the wash then paints
|
|
628
|
+
outside the box as a rotated slab. Observed on admin/style in Chromium, on both
|
|
629
|
+
conic specimens, with the .surface-glass demo the reliable trigger.
|
|
630
|
+
|
|
631
|
+
So the element never moves. The ::before sits at `inset: 0` with
|
|
632
|
+
`border-radius: inherit`, and the sweep animates a registered @property angle
|
|
633
|
+
inside the gradient — the same technique §9b uses. Nothing extends past the
|
|
634
|
+
box, the rounded corners come from the pseudo itself, and no clip has to hold.
|
|
635
|
+
`overflow: hidden` stays for the host's CONTENT, not for the wash.
|
|
636
|
+
|
|
637
|
+
The cost of the trade, stated: a registered-property animation repaints the
|
|
638
|
+
gradient on the main thread each frame, where a transform rode the compositor.
|
|
639
|
+
At the default 16s sweep that is cheap, and §9b already pays it at 4s.
|
|
609
640
|
|
|
610
641
|
Usage: <div class="conic-surface rounded-2xl p-8"> ...content... </div>
|
|
611
642
|
Tune: --conic-surface-from / --conic-surface-to / --conic-surface-duration
|
|
612
643
|
========================================================================== */
|
|
644
|
+
@property --conic-surface-angle {
|
|
645
|
+
syntax: "<angle>";
|
|
646
|
+
inherits: false;
|
|
647
|
+
initial-value: 0deg;
|
|
648
|
+
}
|
|
613
649
|
.conic-surface {
|
|
614
650
|
position: relative;
|
|
615
651
|
isolation: isolate;
|
|
@@ -619,10 +655,11 @@
|
|
|
619
655
|
.conic-surface::before {
|
|
620
656
|
content: "";
|
|
621
657
|
position: absolute;
|
|
622
|
-
inset:
|
|
658
|
+
inset: 0;
|
|
659
|
+
border-radius: inherit;
|
|
623
660
|
z-index: -1;
|
|
624
661
|
background: conic-gradient(
|
|
625
|
-
from
|
|
662
|
+
from var(--conic-surface-angle),
|
|
626
663
|
var(--conic-surface-from, var(--color-cta)),
|
|
627
664
|
var(--conic-surface-to, var(--color-surface)),
|
|
628
665
|
var(--conic-surface-from, var(--color-cta))
|
|
@@ -631,7 +668,7 @@
|
|
|
631
668
|
}
|
|
632
669
|
|
|
633
670
|
@keyframes conic-surface-spin {
|
|
634
|
-
to {
|
|
671
|
+
to { --conic-surface-angle: 360deg; }
|
|
635
672
|
}
|
|
636
673
|
|
|
637
674
|
/* =============================================================================
|
|
@@ -0,0 +1,80 @@
|
|
|
1
|
+
<%# A date of birth, as three selects. The engine's one DOB field.
|
|
2
|
+
|
|
3
|
+
Locals (all optional):
|
|
4
|
+
label_month / label_day / label_year — the field labels. Defaults below.
|
|
5
|
+
on_change — Alpine expression run when ANY of the three changes, so a call
|
|
6
|
+
site can recompute whatever it derives. Defaults to the
|
|
7
|
+
age-gate's onMonthChange(), which every current caller has.
|
|
8
|
+
|
|
9
|
+
ONE HOOK FOR ALL THREE, where age_verify used onMonthChange()
|
|
10
|
+
for month/year and `error = ''` for day. That is not a
|
|
11
|
+
behaviour change: onMonthChange() clears an out-of-range day
|
|
12
|
+
and then clears the error, and a day just chosen FROM
|
|
13
|
+
dayOptions is never out of range — so on the day select it
|
|
14
|
+
reduces to exactly `error = ''`. Checked rather than assumed,
|
|
15
|
+
because an extraction that quietly alters a call site is worse
|
|
16
|
+
than no extraction.
|
|
17
|
+
|
|
18
|
+
WHAT THE ENCLOSING ALPINE COMPONENT MUST SUPPLY. This partial is markup
|
|
19
|
+
only; it binds to the component it is rendered inside:
|
|
20
|
+
|
|
21
|
+
month, day, year — x-model targets
|
|
22
|
+
months — [{ n: 1, label: "January" }, …]
|
|
23
|
+
dayOptions — the days valid for the selected month/year
|
|
24
|
+
years — the year range this caller offers
|
|
25
|
+
|
|
26
|
+
EXTRACTED BECAUSE IT WAS ALREADY DUPLICATED: these selects lived in
|
|
27
|
+
studio/modals/blocks/_age_verify, and turf-monster forked a second copy into
|
|
28
|
+
its own modals/_age_verify. Two copies of one field is how they drift.
|
|
29
|
+
|
|
30
|
+
AND BECAUSE IT REPLACED SOMETHING WORSE. /profile/edit used a custom calendar
|
|
31
|
+
popover, which shipped four defects in three days — its grid collapsed in
|
|
32
|
+
every consumer (grid-cols-7 exists in no consumer bundle), a flipped popover
|
|
33
|
+
used a hardcoded height and landed ~95px wrong, place() re-decided the side on
|
|
34
|
+
every scroll so it snapped mid-scroll, and it came back OPEN after a browser
|
|
35
|
+
Back. Three selects have none of those failure modes: no positioning, no
|
|
36
|
+
flip, no popover, no scroll listener.
|
|
37
|
+
|
|
38
|
+
ITS UTILITIES ARE PROVEN, which the calendar's were not. Measured across the
|
|
39
|
+
fleet: grid-cols-3 and label-upper are present in turf-monster,
|
|
40
|
+
mcritchie-studio AND mcritchie-industries' compiled bundles. The engine ships
|
|
41
|
+
a prebuilt bundle, so a utility exists in a consumer only if that consumer
|
|
42
|
+
already emitted it — this field's do. %>
|
|
43
|
+
<%
|
|
44
|
+
label_month = local_assigns.fetch(:label_month, "Month")
|
|
45
|
+
label_day = local_assigns.fetch(:label_day, "Day")
|
|
46
|
+
label_year = local_assigns.fetch(:label_year, "Year")
|
|
47
|
+
on_change = local_assigns.fetch(:on_change, "onMonthChange()")
|
|
48
|
+
%>
|
|
49
|
+
<div class="grid grid-cols-3 gap-2">
|
|
50
|
+
<div>
|
|
51
|
+
<label class="label-upper block mb-1"><%= label_month %></label>
|
|
52
|
+
<select x-model="month" @change="<%= on_change %>"
|
|
53
|
+
class="input-field !px-3 !py-2.5 text-sm appearance-none cursor-pointer">
|
|
54
|
+
<option value="" disabled><%= label_month %></option>
|
|
55
|
+
<template x-for="m in months" :key="m.n">
|
|
56
|
+
<option :value="m.n" x-text="m.label"></option>
|
|
57
|
+
</template>
|
|
58
|
+
</select>
|
|
59
|
+
</div>
|
|
60
|
+
<div>
|
|
61
|
+
<label class="label-upper block mb-1"><%= label_day %></label>
|
|
62
|
+
<select x-model="day" @change="<%= on_change %>"
|
|
63
|
+
class="input-field !px-3 !py-2.5 text-sm appearance-none cursor-pointer">
|
|
64
|
+
<option value="" disabled><%= label_day %></option>
|
|
65
|
+
<template x-for="d in dayOptions" :key="d">
|
|
66
|
+
<option :value="d" x-text="d"></option>
|
|
67
|
+
</template>
|
|
68
|
+
</select>
|
|
69
|
+
</div>
|
|
70
|
+
<div>
|
|
71
|
+
<label class="label-upper block mb-1"><%= label_year %></label>
|
|
72
|
+
<select x-model="year" @change="<%= on_change %>"
|
|
73
|
+
class="input-field !px-3 !py-2.5 text-sm appearance-none cursor-pointer">
|
|
74
|
+
<option value="" disabled><%= label_year %></option>
|
|
75
|
+
<template x-for="y in years" :key="y">
|
|
76
|
+
<option :value="y" x-text="y"></option>
|
|
77
|
+
</template>
|
|
78
|
+
</select>
|
|
79
|
+
</div>
|
|
80
|
+
</div>
|
|
@@ -49,35 +49,11 @@
|
|
|
49
49
|
<%= intro %>
|
|
50
50
|
</p>
|
|
51
51
|
|
|
52
|
-
<%# Month / Day / Year
|
|
53
|
-
|
|
54
|
-
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
<option value="" disabled>Month</option>
|
|
58
|
-
<template x-for="m in months" :key="m.n">
|
|
59
|
-
<option :value="m.n" x-text="m.label"></option>
|
|
60
|
-
</template>
|
|
61
|
-
</select>
|
|
62
|
-
</div>
|
|
63
|
-
<div>
|
|
64
|
-
<label class="label-upper block mb-1">Day</label>
|
|
65
|
-
<select x-model="day" @change="error = ''" class="input-field !px-3 !py-2.5 text-sm appearance-none cursor-pointer">
|
|
66
|
-
<option value="" disabled>Day</option>
|
|
67
|
-
<template x-for="d in dayOptions" :key="d">
|
|
68
|
-
<option :value="d" x-text="d"></option>
|
|
69
|
-
</template>
|
|
70
|
-
</select>
|
|
71
|
-
</div>
|
|
72
|
-
<div>
|
|
73
|
-
<label class="label-upper block mb-1">Year</label>
|
|
74
|
-
<select x-model="year" @change="onMonthChange()" class="input-field !px-3 !py-2.5 text-sm appearance-none cursor-pointer">
|
|
75
|
-
<option value="" disabled>Year</option>
|
|
76
|
-
<template x-for="y in years" :key="y">
|
|
77
|
-
<option :value="y" x-text="y"></option>
|
|
78
|
-
</template>
|
|
79
|
-
</select>
|
|
80
|
-
</div>
|
|
52
|
+
<%# Month / Day / Year — the engine's one DOB field, shared with the profile
|
|
53
|
+
page's birthday row. It lived here first and turf-monster forked a second
|
|
54
|
+
copy; extracting it is what stops the two drifting. %>
|
|
55
|
+
<div class="mb-4">
|
|
56
|
+
<%= render "studio/fields/date_of_birth" %>
|
|
81
57
|
</div>
|
|
82
58
|
|
|
83
59
|
<%# Live too-young hint — shown once a full date is picked and it fails the
|
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
<%# Birthday —
|
|
1
|
+
<%# Birthday — THREE SELECTS over three integer columns.
|
|
2
2
|
|
|
3
3
|
Locals: user (required).
|
|
4
4
|
|
|
@@ -9,26 +9,29 @@
|
|
|
9
9
|
(needs month and day, and no year at all). A single date column answers the
|
|
10
10
|
second one badly.
|
|
11
11
|
|
|
12
|
-
|
|
13
|
-
|
|
12
|
+
THE SAME FIELD THE AGE GATE USES (operator's call, 2026-08-15). It renders
|
|
13
|
+
studio/fields/_date_of_birth, which studio/modals/blocks/_age_verify also
|
|
14
|
+
renders — one field, two surfaces. It replaced a custom calendar popover that
|
|
15
|
+
shipped four defects in three days; see the component script for the list.
|
|
14
16
|
|
|
15
17
|
TWO INPUTS, ONE NAME, AND ONLY EVER ONE IN THE DOM. `<template x-if>` is what
|
|
16
18
|
makes that true rather than a claim: the branch Alpine does not take is never
|
|
17
19
|
instantiated, so the form can never carry two `profile[birthday]` fields and
|
|
18
20
|
the server can never receive the wrong one. Toggling with x-show would leave
|
|
19
21
|
BOTH present and submit the hidden one's value too — the last duplicate wins
|
|
20
|
-
in Rack's params
|
|
22
|
+
in Rack's params.
|
|
21
23
|
|
|
22
|
-
no JS — the native `<input type="date">`.
|
|
24
|
+
no JS — the native `<input type="date">`. Not a downgrade: it is the
|
|
23
25
|
control every platform already knows how to render, and the reason
|
|
24
26
|
a JS-less page can still set a birthday at all.
|
|
25
|
-
JS — the
|
|
27
|
+
JS — the three selects. `alpine` comes from the enclosing
|
|
26
28
|
studioProfileForm and is true only once that component has booted.
|
|
27
|
-
|
|
28
|
-
The operator asked for a calendar rather than an open field (2026-08-15).
|
|
29
29
|
%>
|
|
30
|
-
|
|
31
|
-
|
|
30
|
+
<%# NO FIELD LABEL HERE, deliberately (operator's call, 2026-08-15). The row is
|
|
31
|
+
already titled "Birthday" by the section wrapper, and the three selects carry
|
|
32
|
+
their own MONTH / DAY / YEAR labels — a "Date of birth" line between the two
|
|
33
|
+
said the same thing a third time. The no-JS <input type="date"> keeps its id
|
|
34
|
+
so a label elsewhere could still point at it. %>
|
|
32
35
|
<%
|
|
33
36
|
y = user.birth_year
|
|
34
37
|
m = user.birth_month
|
|
@@ -36,7 +39,7 @@
|
|
|
36
39
|
value = (y.present? && m.present? && d.present?) ? format("%04d-%02d-%02d", y, m, d) : nil
|
|
37
40
|
%>
|
|
38
41
|
|
|
39
|
-
<div x-data="
|
|
42
|
+
<div x-data="studioBirthdayFields('<%= value %>')">
|
|
40
43
|
<%# --- the JS-less path ------------------------------------------------- %>
|
|
41
44
|
<template x-if="!alpine">
|
|
42
45
|
<input type="date" name="profile[birthday]" id="profile_birthday"
|
|
@@ -44,187 +47,37 @@
|
|
|
44
47
|
autocomplete="bday" max="<%= Date.current.iso8601 %>">
|
|
45
48
|
</template>
|
|
46
49
|
|
|
47
|
-
<%# --- the
|
|
50
|
+
<%# --- the three selects -------------------------------------------------- %>
|
|
48
51
|
<template x-if="alpine">
|
|
49
52
|
<div>
|
|
50
|
-
<%# The value the form actually submits. The
|
|
51
|
-
types into it.
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
|
|
58
|
-
|
|
59
|
-
|
|
60
|
-
|
|
61
|
-
|
|
62
|
-
|
|
63
|
-
|
|
64
|
-
|
|
65
|
-
|
|
66
|
-
|
|
67
|
-
|
|
68
|
-
|
|
69
|
-
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
|
|
73
|
-
|
|
74
|
-
class="studio-birthday-popover"
|
|
75
|
-
:style="'top:' + top + 'px; left:' + left + 'px; min-width:' + width + 'px;'">
|
|
76
|
-
|
|
77
|
-
<%# Month and year JUMP rather than step. Stepping from today to 1985 is
|
|
78
|
-
about four hundred clicks, which is the whole reason this is not the
|
|
79
|
-
contest picker. %>
|
|
80
|
-
<div class="studio-birthday-nav">
|
|
81
|
-
<select x-model.number="viewMonth" class="input-field studio-birthday-month"
|
|
82
|
-
aria-label="Month">
|
|
83
|
-
<template x-for="(name, i) in months" :key="name">
|
|
84
|
-
<option :value="i" x-text="name"></option>
|
|
85
|
-
</template>
|
|
86
|
-
</select>
|
|
87
|
-
<select x-model.number="viewYear" class="input-field studio-birthday-year"
|
|
88
|
-
aria-label="Year">
|
|
89
|
-
<template x-for="y in years" :key="y">
|
|
90
|
-
<option :value="y" x-text="y"></option>
|
|
91
|
-
</template>
|
|
92
|
-
</select>
|
|
93
|
-
</div>
|
|
94
|
-
|
|
95
|
-
<div class="studio-birthday-grid studio-birthday-weekdays">
|
|
96
|
-
<template x-for="(wd, i) in weekdays" :key="i">
|
|
97
|
-
<div x-text="wd"></div>
|
|
98
|
-
</template>
|
|
99
|
-
</div>
|
|
100
|
-
|
|
101
|
-
<div class="studio-birthday-grid">
|
|
102
|
-
<template x-for="(cell, i) in calDays()" :key="i">
|
|
103
|
-
<div>
|
|
104
|
-
<template x-if="cell">
|
|
105
|
-
<button type="button" @click="pick(cell)" :disabled="isFuture(cell)"
|
|
106
|
-
:class="isSelected(cell) && 'is-selected'"
|
|
107
|
-
class="studio-birthday-day"
|
|
108
|
-
x-text="cell"></button>
|
|
109
|
-
</template>
|
|
110
|
-
</div>
|
|
111
|
-
</template>
|
|
112
|
-
</div>
|
|
113
|
-
|
|
114
|
-
<%# A birthday already set is the only case where clearing is meaningful,
|
|
115
|
-
and without this the calendar is a one-way door — every day cell sets
|
|
116
|
-
a value and none of them unset one. %>
|
|
117
|
-
<div class="mt-3 flex justify-end" x-show="value">
|
|
118
|
-
<button type="button" @click="clear()"
|
|
119
|
-
class="text-xs text-secondary hover:text-heading transition">
|
|
120
|
-
Clear
|
|
121
|
-
</button>
|
|
122
|
-
</div>
|
|
123
|
-
</div>
|
|
53
|
+
<%# The value the form actually submits. The selects write here; nothing
|
|
54
|
+
types into it.
|
|
55
|
+
|
|
56
|
+
THE NAME IS BOUND, AND THAT IS THE WHOLE POINT. A partly-filled date
|
|
57
|
+
has no ISO value, and submitting a BLANK birthday is not "no change" to
|
|
58
|
+
this controller — it is DELETE, and it clears all three columns on
|
|
59
|
+
purpose (see #birthday_attributes: someone deleting their birthday
|
|
60
|
+
means it). Those two facts met in a data-loss bug: a stored Jan 31,
|
|
61
|
+
switch the month to February, onMonthChange blanks the 31st because
|
|
62
|
+
February has no 31st, and the form is now one Save away from wiping a
|
|
63
|
+
birthday the person never asked to delete.
|
|
64
|
+
|
|
65
|
+
On THIS path a blank is always an edit in progress and never an
|
|
66
|
+
intentional clear, because the selects offer no empty option — once a
|
|
67
|
+
part is set the user cannot unset it. The intentional clear lives on
|
|
68
|
+
the no-JS <input type="date"> above, where it is a real gesture.
|
|
69
|
+
|
|
70
|
+
So while the date is incomplete the field submits NOTHING: Alpine
|
|
71
|
+
removes an attribute bound to false, an input with no name is not
|
|
72
|
+
submitted, and the controller's `return {} if raw.nil?` makes an absent
|
|
73
|
+
param a no-op. The birthday is left exactly as it was. %>
|
|
74
|
+
<input type="hidden" :name="complete ? 'profile[birthday]' : false" :value="isoValue">
|
|
75
|
+
|
|
76
|
+
<%= render "studio/fields/date_of_birth" %>
|
|
124
77
|
</div>
|
|
125
78
|
</template>
|
|
126
79
|
</div>
|
|
127
80
|
|
|
128
81
|
<p class="text-muted text-xs mt-2">Stored as day, month and year — we use it for birthdays and age checks.</p>
|
|
129
82
|
|
|
130
|
-
<style>
|
|
131
|
-
/* EVERY RULE BELOW IS OWNED CSS, NOT TAILWIND UTILITIES, and that is not a
|
|
132
|
-
style preference — it is the bug the operator reported.
|
|
133
|
-
|
|
134
|
-
The engine ships a PREBUILT bundle to consumers. A Tailwind utility exists in
|
|
135
|
-
a consuming app only if that app's OWN views already emitted it, because its
|
|
136
|
-
build scans its own source, not the gem's. This calendar first shipped using
|
|
137
|
-
`grid grid-cols-7`, and `grid-cols-7` is rare enough that NO consumer had
|
|
138
|
-
ever emitted it: measured in mcritchie-studio's compiled bundle, zero
|
|
139
|
-
occurrences. With no grid, the seven weekday letters stacked into a single
|
|
140
|
-
vertical column and the popover grew to the height of the page.
|
|
141
|
-
`hover:bg-surface-alt` was missing for the same reason.
|
|
142
|
-
|
|
143
|
-
The same note is on studio/profiles/_identity_styles for the same reason.
|
|
144
|
-
|
|
145
|
-
AND THE BROWSER LANE COULD NOT SEE IT. e2e/tailwind_input.css carries
|
|
146
|
-
`@source "../app/views"`, so the lane compiles the engine's OWN views and
|
|
147
|
-
emits grid-cols-7 — every spec passed against a page no consumer renders.
|
|
148
|
-
That blindness is recorded in docs/E2E_LANE.md; the fix here is to stop
|
|
149
|
-
depending on the utility at all. */
|
|
150
|
-
.studio-birthday-nav {
|
|
151
|
-
display: flex;
|
|
152
|
-
align-items: center;
|
|
153
|
-
gap: 0.5rem;
|
|
154
|
-
margin-bottom: 0.75rem;
|
|
155
|
-
}
|
|
156
|
-
|
|
157
|
-
.studio-birthday-month { flex: 1 1 auto; }
|
|
158
|
-
|
|
159
|
-
.studio-birthday-month,
|
|
160
|
-
.studio-birthday-year {
|
|
161
|
-
font-size: 0.875rem;
|
|
162
|
-
padding-top: 0.25rem;
|
|
163
|
-
padding-bottom: 0.25rem;
|
|
164
|
-
}
|
|
165
|
-
|
|
166
|
-
/* THE ONE THAT BROKE. Seven columns, owned outright. */
|
|
167
|
-
.studio-birthday-grid {
|
|
168
|
-
display: grid;
|
|
169
|
-
grid-template-columns: repeat(7, minmax(0, 1fr));
|
|
170
|
-
text-align: center;
|
|
171
|
-
}
|
|
172
|
-
|
|
173
|
-
.studio-birthday-weekdays {
|
|
174
|
-
margin-bottom: 0.25rem;
|
|
175
|
-
font-size: 0.75rem;
|
|
176
|
-
font-weight: 500;
|
|
177
|
-
color: var(--color-text-muted, #9ca3af);
|
|
178
|
-
}
|
|
179
|
-
|
|
180
|
-
.studio-birthday-weekdays > div { padding: 0.25rem 0; }
|
|
181
|
-
|
|
182
|
-
.studio-birthday-day {
|
|
183
|
-
width: 2rem;
|
|
184
|
-
height: 2rem;
|
|
185
|
-
border-radius: 9999px;
|
|
186
|
-
font-size: 0.875rem;
|
|
187
|
-
color: var(--color-text-body, inherit);
|
|
188
|
-
transition: background-color 150ms ease, color 150ms ease;
|
|
189
|
-
}
|
|
190
|
-
|
|
191
|
-
.studio-birthday-day:hover:not(:disabled) {
|
|
192
|
-
background: var(--color-surface-alt, rgba(127, 127, 127, 0.15));
|
|
193
|
-
}
|
|
194
|
-
|
|
195
|
-
.studio-birthday-day:disabled {
|
|
196
|
-
opacity: 0.4;
|
|
197
|
-
cursor: not-allowed;
|
|
198
|
-
}
|
|
199
|
-
|
|
200
|
-
.studio-birthday-day.is-selected {
|
|
201
|
-
background: var(--color-primary);
|
|
202
|
-
color: #fff;
|
|
203
|
-
font-weight: 700;
|
|
204
|
-
}
|
|
205
|
-
|
|
206
|
-
/* FIXED, not absolute, for TWO reasons — and the second is the one a spec can
|
|
207
|
-
actually catch.
|
|
208
|
-
|
|
209
|
-
1. Clipping. The edit page's rows sit inside `card p-0 overflow-hidden`. An
|
|
210
|
-
absolutely-positioned popover only escapes that clip while NO ancestor is
|
|
211
|
-
positioned; the day a consumer's card grows a `position: relative` it is
|
|
212
|
-
clipped out of sight, with perfectly correct markup.
|
|
213
|
-
2. Anchoring. place() writes VIEWPORT coordinates from the trigger's rect.
|
|
214
|
-
Fixed keeps them meaningful as the page scrolls; absolute reinterprets
|
|
215
|
-
them as document coordinates and the popover drifts by the scroll offset.
|
|
216
|
-
|
|
217
|
-
At scroll 0 the two are indistinguishable, which is why the spec that proves
|
|
218
|
-
this scrolls first. */
|
|
219
|
-
.studio-birthday-popover {
|
|
220
|
-
position: fixed;
|
|
221
|
-
z-index: 40;
|
|
222
|
-
background: var(--color-surface);
|
|
223
|
-
border: 1px solid var(--color-border-subtle);
|
|
224
|
-
border-radius: 0.75rem;
|
|
225
|
-
box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.25);
|
|
226
|
-
padding: 0.75rem;
|
|
227
|
-
}
|
|
228
|
-
</style>
|
|
229
|
-
|
|
230
83
|
<%= render "studio/profiles/birthday_picker_script" %>
|
|
@@ -1,40 +1,48 @@
|
|
|
1
|
-
<%# The birthday
|
|
2
|
-
|
|
3
|
-
|
|
4
|
-
|
|
5
|
-
|
|
6
|
-
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
10
|
-
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
|
|
1
|
+
<%# The birthday field's Alpine component.
|
|
2
|
+
|
|
3
|
+
Supplies exactly what studio/fields/_date_of_birth binds to — month, day,
|
|
4
|
+
year, months, dayOptions, years, onMonthChange() — plus the one thing the
|
|
5
|
+
profile page needs on top: it writes the assembled ISO date into the
|
|
6
|
+
enclosing studioProfileForm's `fields.birthday`, which is what raises the
|
|
7
|
+
save bar.
|
|
8
|
+
|
|
9
|
+
THIS REPLACED A CALENDAR POPOVER, and the replacement is the point. That
|
|
10
|
+
component shipped four defects in three days:
|
|
11
|
+
|
|
12
|
+
· its day grid used `grid-cols-7`, which exists in NO consumer bundle, so
|
|
13
|
+
the seven weekday letters stacked into one column and the popover grew to
|
|
14
|
+
the height of the page;
|
|
15
|
+
· a flipped popover positioned itself from a hardcoded 340px height when
|
|
16
|
+
the real popover is ~245px, landing ~95px too high;
|
|
17
|
+
· place() ran on every scroll and RE-DECIDED the side, so it snapped
|
|
18
|
+
between above and below mid-scroll — a 254px jump between two 60px steps;
|
|
19
|
+
· and it came back OPEN after a browser Back.
|
|
20
|
+
|
|
21
|
+
Three selects have none of those failure modes. There is no position, no
|
|
22
|
+
flip, no popover, and no scroll listener — so there is nothing to place
|
|
23
|
+
wrongly, nothing to re-decide, and nothing to restore open.
|
|
21
24
|
%>
|
|
22
25
|
<script>
|
|
23
|
-
window.
|
|
26
|
+
window.studioBirthdayFields = function (initial) {
|
|
27
|
+
var parts = /^(\d{4})-(\d{2})-(\d{2})$/.exec(String(initial || ""));
|
|
28
|
+
|
|
24
29
|
return {
|
|
25
|
-
|
|
26
|
-
|
|
27
|
-
|
|
28
|
-
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
30
|
+
year: parts ? String(+parts[1]) : "",
|
|
31
|
+
month: parts ? String(+parts[2]) : "",
|
|
32
|
+
day: parts ? String(+parts[3]) : "",
|
|
33
|
+
|
|
34
|
+
months: [
|
|
35
|
+
{ n: 1, label: "January" }, { n: 2, label: "February" },
|
|
36
|
+
{ n: 3, label: "March" }, { n: 4, label: "April" },
|
|
37
|
+
{ n: 5, label: "May" }, { n: 6, label: "June" },
|
|
38
|
+
{ n: 7, label: "July" }, { n: 8, label: "August" },
|
|
39
|
+
{ n: 9, label: "September" }, { n: 10, label: "October" },
|
|
40
|
+
{ n: 11, label: "November" }, { n: 12, label: "December" }
|
|
41
|
+
],
|
|
35
42
|
|
|
36
43
|
// Oldest year offered. 120 covers every living person with room to spare;
|
|
37
|
-
// the
|
|
44
|
+
// the newest is THIS year, because a birthday cannot be in the future and
|
|
45
|
+
// the cheapest way to say so is not to offer one.
|
|
38
46
|
get years() {
|
|
39
47
|
var now = new Date().getFullYear();
|
|
40
48
|
var out = [];
|
|
@@ -42,153 +50,96 @@
|
|
|
42
50
|
return out;
|
|
43
51
|
},
|
|
44
52
|
|
|
45
|
-
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
|
|
49
|
-
|
|
50
|
-
|
|
51
|
-
this.
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
|
|
55
|
-
|
|
56
|
-
// current year means everyone starts 30+ years from where they are going.
|
|
57
|
-
syncViewFromValue() {
|
|
58
|
-
var parts = this.parse(this.value);
|
|
59
|
-
var now = new Date();
|
|
60
|
-
this.viewYear = parts ? parts.y : now.getFullYear() - 30;
|
|
61
|
-
this.viewMonth = parts ? parts.m - 1 : 0;
|
|
62
|
-
},
|
|
63
|
-
|
|
64
|
-
parse(iso) {
|
|
65
|
-
var m = /^(\d{4})-(\d{2})-(\d{2})$/.exec(String(iso || ""));
|
|
66
|
-
if (!m) return null;
|
|
67
|
-
return { y: +m[1], m: +m[2], d: +m[3] };
|
|
68
|
-
},
|
|
69
|
-
|
|
70
|
-
get displayValue() {
|
|
71
|
-
var p = this.parse(this.value);
|
|
72
|
-
if (!p) return "";
|
|
73
|
-
return this.months[p.m - 1] + " " + p.d + ", " + p.y;
|
|
53
|
+
// The days that exist in the chosen month — 30 in April, and 29 in a
|
|
54
|
+
// February only when the chosen year is a leap year. Day 0 of the NEXT
|
|
55
|
+
// month is the last day of this one, which is the standard trick and
|
|
56
|
+
// handles the leap rule without one being written down.
|
|
57
|
+
get dayOptions() {
|
|
58
|
+
var m = parseInt(this.month, 10);
|
|
59
|
+
var y = parseInt(this.year, 10) || 2000; // a leap year, so Feb offers 29 before a year is picked
|
|
60
|
+
if (!m) return Array.from({ length: 31 }, function (_, i) { return i + 1; });
|
|
61
|
+
|
|
62
|
+
var count = new Date(y, m, 0).getDate();
|
|
63
|
+
return Array.from({ length: count }, function (_, i) { return i + 1; });
|
|
74
64
|
},
|
|
75
65
|
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
var first = new Date(this.viewYear, this.viewMonth, 1).getDay();
|
|
79
|
-
var count = new Date(this.viewYear, this.viewMonth + 1, 0).getDate();
|
|
80
|
-
var cells = [];
|
|
81
|
-
for (var i = 0; i < first; i++) cells.push(null);
|
|
82
|
-
for (var d = 1; d <= count; d++) cells.push(d);
|
|
83
|
-
return cells;
|
|
66
|
+
get complete() {
|
|
67
|
+
return !!(this.year && this.month && this.day);
|
|
84
68
|
},
|
|
85
69
|
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
today.setHours(0, 0, 0, 0);
|
|
92
|
-
return candidate > today;
|
|
70
|
+
get isoValue() {
|
|
71
|
+
if (!this.complete) return "";
|
|
72
|
+
return this.year + "-" +
|
|
73
|
+
String(this.month).padStart(2, "0") + "-" +
|
|
74
|
+
String(this.day).padStart(2, "0");
|
|
93
75
|
},
|
|
94
76
|
|
|
95
|
-
|
|
96
|
-
|
|
97
|
-
|
|
77
|
+
// Named to match what the shared field binds to. Clears a day that the new
|
|
78
|
+
// month or year does not have — pick 31 January then switch to February and
|
|
79
|
+
// the 31st stops existing — then publishes.
|
|
80
|
+
onMonthChange() {
|
|
81
|
+
if (this.day && parseInt(this.day, 10) > this.dayOptions.length) this.day = "";
|
|
82
|
+
this.publish();
|
|
98
83
|
},
|
|
99
84
|
|
|
100
|
-
|
|
101
|
-
|
|
102
|
-
|
|
103
|
-
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
|
|
116
|
-
|
|
117
|
-
|
|
118
|
-
|
|
85
|
+
// WHAT THE DIRTY CHECK IS SHOWN, and why it is not isoValue.
|
|
86
|
+
//
|
|
87
|
+
// isoValue is "" for EVERY incomplete date, so publishing it made a whole
|
|
88
|
+
// class of edit invisible: on an empty birthday, choosing a month wrote ""
|
|
89
|
+
// over "" and the save bar never appeared. The operator reported exactly
|
|
90
|
+
// that. Choosing a month IS a change — the form just had no way to see it.
|
|
91
|
+
//
|
|
92
|
+
// So the signal is the three parts, which differ whenever any one of them
|
|
93
|
+
// does. It stays equal to isoValue for a COMPLETE date (same string, same
|
|
94
|
+
// shape) so a finished edit compares correctly against the value the page
|
|
95
|
+
// loaded with, and Discard restores it to that same string.
|
|
96
|
+
//
|
|
97
|
+
// `fields` belongs to the enclosing studioProfileForm — Alpine's scope
|
|
98
|
+
// chain makes it reachable and it is the SAME object, so assigning through
|
|
99
|
+
// it is what raises the save bar.
|
|
100
|
+
// THE EMPTY CASE IS NOT A PARTIAL ONE, and conflating them makes the page
|
|
101
|
+
// arrive already dirty. studioProfileForm's initial.birthday is "" for a
|
|
102
|
+
// person with no birthday on file, so an untouched empty field must also
|
|
103
|
+
// read "" — joining three blanks gives "--", which compares as a change and
|
|
104
|
+
// would raise the save bar on load, before anyone touched anything.
|
|
105
|
+
get signature() {
|
|
106
|
+
if (this.complete) return this.isoValue;
|
|
107
|
+
if (!this.year && !this.month && !this.day) return "";
|
|
108
|
+
return [this.year, this.month, this.day].join("-");
|
|
119
109
|
},
|
|
120
110
|
|
|
121
|
-
|
|
122
|
-
this.
|
|
123
|
-
if (!this.open) return;
|
|
124
|
-
|
|
125
|
-
this.syncViewFromValue();
|
|
126
|
-
|
|
127
|
-
// Placed twice, deliberately. The first call stops the popover appearing
|
|
128
|
-
// at 0,0 for a frame; the second is the one that can MEASURE, because
|
|
129
|
-
// offsetHeight is 0 while x-show still has the element hidden — and it is
|
|
130
|
-
// also the only place the SIDE is chosen, for the life of this open.
|
|
131
|
-
this.place();
|
|
132
|
-
this.$nextTick(function () {
|
|
133
|
-
this.chooseSide();
|
|
134
|
-
this.place();
|
|
135
|
-
}.bind(this));
|
|
111
|
+
publish() {
|
|
112
|
+
if (this.fields) this.fields.birthday = this.signature;
|
|
136
113
|
},
|
|
137
114
|
|
|
138
|
-
//
|
|
139
|
-
//
|
|
140
|
-
//
|
|
115
|
+
// DISCARD MUST REACH THE SELECTS. studioProfileForm#discard reassigns its
|
|
116
|
+
// `fields` object, which restores every input bound with x-model — and the
|
|
117
|
+
// three selects are not, because they are bound to THIS component's own
|
|
118
|
+
// month/day/year. So Discard used to leave them showing the abandoned edit
|
|
119
|
+
// while the bar disappeared, and the hidden input went on carrying it: the
|
|
120
|
+
// next save would write the change the user had just discarded.
|
|
141
121
|
//
|
|
142
|
-
//
|
|
143
|
-
//
|
|
144
|
-
//
|
|
145
|
-
//
|
|
146
|
-
|
|
147
|
-
|
|
148
|
-
|
|
149
|
-
|
|
150
|
-
|
|
151
|
-
|
|
152
|
-
|
|
153
|
-
place() {
|
|
154
|
-
var el = this.$refs.trigger;
|
|
155
|
-
if (!el) return;
|
|
156
|
-
|
|
157
|
-
var r = el.getBoundingClientRect();
|
|
158
|
-
var popover = this.$refs.popover;
|
|
159
|
-
var height = (popover && popover.offsetHeight) || 340;
|
|
160
|
-
var GAP = 4;
|
|
161
|
-
|
|
162
|
-
this.left = r.left;
|
|
163
|
-
this.width = r.width;
|
|
164
|
-
|
|
165
|
-
// THE SIDE IS NOT RE-DECIDED HERE. place() runs on every scroll and
|
|
166
|
-
// resize event, and choosing the side each time makes the popover SNAP
|
|
167
|
-
// between above and below the moment the available space crosses the
|
|
168
|
-
// threshold. Measured on turf-monster at 60px scroll steps: ABOVE, ABOVE,
|
|
169
|
-
// ABOVE, ABOVE, then below — a 254px jump between two steps, mid-scroll,
|
|
170
|
-
// while the reader was looking at it. That is what "the date picker
|
|
171
|
-
// freaks out on scroll" was.
|
|
172
|
-
//
|
|
173
|
-
// chooseSide() runs ONCE per open; this only moves the popover.
|
|
174
|
-
this.top = this.flipped ? r.top - height - GAP : r.bottom + GAP;
|
|
122
|
+
// Watching `fields.birthday` is what joins them. Discard sets it back to
|
|
123
|
+
// the initial string, this fires, and the selects follow — and because the
|
|
124
|
+
// guard below returns early when the value already matches, the publish →
|
|
125
|
+
// watch → publish path cannot loop.
|
|
126
|
+
init() {
|
|
127
|
+
var self = this;
|
|
128
|
+
this.publish();
|
|
129
|
+
this.$watch("fields.birthday", function (value) {
|
|
130
|
+
if (value === self.signature) return;
|
|
131
|
+
self.adopt(value);
|
|
132
|
+
});
|
|
175
133
|
},
|
|
176
134
|
|
|
177
|
-
//
|
|
178
|
-
//
|
|
179
|
-
//
|
|
180
|
-
|
|
181
|
-
|
|
182
|
-
|
|
183
|
-
|
|
184
|
-
|
|
185
|
-
var r = el.getBoundingClientRect();
|
|
186
|
-
var popover = this.$refs.popover;
|
|
187
|
-
var height = (popover && popover.offsetHeight) || 340;
|
|
188
|
-
var GAP = 4;
|
|
189
|
-
var below = window.innerHeight - r.bottom;
|
|
190
|
-
|
|
191
|
-
this.flipped = below < height + GAP && r.top > height + GAP;
|
|
135
|
+
// Takes an ISO date (or a blank) and puts the selects on it. Anything that
|
|
136
|
+
// is not a full date empties all three, which is what an initial blank and
|
|
137
|
+
// a cleared birthday both mean.
|
|
138
|
+
adopt(value) {
|
|
139
|
+
var m = /^(\d{4})-(\d{2})-(\d{2})$/.exec(String(value || ""));
|
|
140
|
+
this.year = m ? String(+m[1]) : "";
|
|
141
|
+
this.month = m ? String(+m[2]) : "";
|
|
142
|
+
this.day = m ? String(+m[3]) : "";
|
|
192
143
|
}
|
|
193
144
|
};
|
|
194
145
|
};
|
|
@@ -297,12 +297,22 @@
|
|
|
297
297
|
</div>
|
|
298
298
|
<% end %>
|
|
299
299
|
|
|
300
|
+
<%# The glow host is a BARE wrapper and the opaque face is a CHILD. The ring
|
|
301
|
+
pseudos sit at z-index -1/-2, and a negative-z descendant paints ABOVE its
|
|
302
|
+
host's own background — so a background on the HOST washes the whole card
|
|
303
|
+
face and there is no ring left. The child in front is what covers the
|
|
304
|
+
middle, exactly as TM's holo-wrap / holo-card split does. See
|
|
305
|
+
engine-motion.css §9b. %>
|
|
300
306
|
<%= render layout: "style/specimen",
|
|
301
307
|
locals: { klass: ".studio-team-glow (knob: --studio-team-glow-color)",
|
|
302
|
-
usage: %(<div class="studio-team-glow rounded-xl
|
|
308
|
+
usage: %(<div class="studio-team-glow rounded-xl">\n <div class="rounded-xl bg-surface p-5">Selected</div>\n</div>) } do %>
|
|
303
309
|
<div class="flex flex-wrap items-center justify-center gap-6">
|
|
304
|
-
<div class="studio-team-glow rounded-xl
|
|
305
|
-
|
|
310
|
+
<div class="studio-team-glow rounded-xl">
|
|
311
|
+
<div class="rounded-xl bg-surface border border-subtle px-5 py-4 text-sm font-semibold text-heading text-center">Selected</div>
|
|
312
|
+
</div>
|
|
313
|
+
<div class="studio-team-glow rounded-xl" style="--studio-team-glow-color: #E11D48">
|
|
314
|
+
<div class="rounded-xl bg-surface border border-subtle px-5 py-4 text-sm font-semibold text-heading text-center">Ravens</div>
|
|
315
|
+
</div>
|
|
306
316
|
</div>
|
|
307
317
|
<% end %>
|
|
308
318
|
|
data/lib/studio/version.rb
CHANGED
metadata
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: studio-engine
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.
|
|
4
|
+
version: 0.54.1
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Alex McRitchie
|
|
@@ -298,6 +298,7 @@ files:
|
|
|
298
298
|
- app/views/studio/emails/index.html.erb
|
|
299
299
|
- app/views/studio/emails/orphan.html.erb
|
|
300
300
|
- app/views/studio/emails/show.html.erb
|
|
301
|
+
- app/views/studio/fields/_date_of_birth.html.erb
|
|
301
302
|
- app/views/studio/links/confirm.html.erb
|
|
302
303
|
- app/views/studio/local_emails/index.html.erb
|
|
303
304
|
- app/views/studio/mailers/_layered_banner.html.erb
|