diff --git a/public/og/enumerate-zip-iota-or-views-enumerate.png b/public/og/enumerate-zip-iota-or-views-enumerate.png new file mode 100644 index 00000000..68c9b910 Binary files /dev/null and b/public/og/enumerate-zip-iota-or-views-enumerate.png differ diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/README.md b/social/facebook/enumerate-zip-iota-or-views-enumerate/README.md new file mode 100644 index 00000000..85dc2af2 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/README.md @@ -0,0 +1,130 @@ +# LinkedIn Square Post (1200x1200) + +Standard LinkedIn feed post at 1:1 square aspect ratio. Works well as a scroll-stopping image in the LinkedIn feed. + +## Usage + +```bash +brand-gen init social/linkedin-post my-post +cd my-post +# edit content.md +brand-gen build # -> index.html +brand-gen image # -> my-post.png (2400x2400 at 2x DPR) +``` + +## Directives + +All templates use one of four post types. Each uses the same directive pattern: + +### Quote + +```markdown +::::post{type=quote theme=dark logo=top-left} + +:::quote{author="Filip Sajdak" role="Founder & CEO, ScudoAI"} + +Sovereign AI is not a feature you bolt on later. It is the architecture. + +::: + +:::: +``` + +### Stat + +```markdown +::::post{type=stat theme=light accent=brand} + +:::stat{value="USD 80B" label="Worldwide sovereign cloud spend" context="Europe is the fastest-growing region at 83% YoY." source="Gartner, Feb 2026"} +::: + +:::: +``` + +### Announcement + +```markdown +::::post{type=announcement theme=light} + +:::announcement{kind="launch" date="April 2026"} + +# ScudoAI joins NVIDIA Inception + +We are thrilled to join the NVIDIA Inception program to accelerate sovereign AI deployment across European regulated industries. + +::: + +:::: +``` + +### Insight + +```markdown +::::post{type=insight theme=dark} + +:::insight{citation="EU AI Act full enforcement: 2 August 2026"} + +# The AI compliance deadline most companies are missing + +Full EU AI Act enforcement for high-risk AI systems begins in 2 August 2026. Most regulated enterprises are not ready for the documentation and audit requirements. Teams running cloud AI will face the hardest transition. + +::: + +:::: +``` + +## Attributes reference + +### `::::post` attributes + +- `type` - quote | stat | announcement | insight (required) +- `theme` - dark | light (default light) +- `accent` - brand | amber | nv (optional colored gradient background) +- `logo` - top-left | top-right | bottom-left | bottom-right (default bottom-right) + +### `:::quote` attributes + +- `author` - attribution name (appears below the quote) +- `role` - attribution role/title (appears under the author) + +### `:::stat` attributes + +- `value` - the big number (e.g. "USD 80B", "3/4", "Hours to minutes") +- `label` - short label under the value +- `context` - 1-2 sentences providing context +- `source` - small citation + +### `:::announcement` attributes + +- `kind` - launch | milestone | event | hire | partnership +- `date` - free-form date string + +### `:::insight` attributes + +- `citation` - small footer citation (e.g. regulation reference) + +## Character limits + +- **Quote body**: ~120 characters for maximum readability at thumbnail size +- **Stat label**: ~40 characters +- **Stat context**: ~200 characters +- **Announcement headline**: ~80 characters +- **Announcement body**: ~250 characters +- **Insight hook**: ~100 characters +- **Insight body**: ~300 characters + +Posts that exceed these limits will still render but typography may shrink or content may clip. + +## Media + +Add a `media/` directory to the project and reference local files: + +```markdown +:::background{image="nvidia-thor.jpg" overlay="dark"} +::: + +:::image{src="product-screenshot.png" alt="ScudoAI compliance dashboard" position="top"} +::: +``` + +Paths are relative to the project's `media/` directory. diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json new file mode 100644 index 00000000..a8a4bc01 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json @@ -0,0 +1,11 @@ +{ + "template": "social/linkedin-post", + "bundleSource": "scudoai-social.css", + "brandKitVersion": "0.1.0", + "vendoredAt": "2026-11-17T20:50:47.502Z", + "files": [ + "scudoai.css", + "layout.css", + "logo-symbol.svg" + ] +} diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/layout.css b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/layout.css new file mode 100644 index 00000000..e69de29b diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg new file mode 100644 index 00000000..1fc44966 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg @@ -0,0 +1,22 @@ + diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css new file mode 100644 index 00000000..612b5a12 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css @@ -0,0 +1,343 @@ +/* ============================================================ + * wro.cpp social card -- standalone stylesheet + * + * Self-contained styling for brand-gen's `social/linkedin-post` + * output. Replaces (does NOT extend) the scudoai-social.css that + * brand-gen ships with -- the design language here is the + * wro.cpp 'Editorial Magnet' look used by src/styles/global.css, + * not the scudoai blue-indigo deck aesthetic. + * + * Source of truth for tokens: src/styles/global.css:12-30 in the + * wrocpp.github.io repo. + * + * Brand-gen markup we style: + *
+ *
(decorative, hidden) + *
(decorative, hidden) + *
+ *

...

+ *

...

+ *
...
+ *
+ *
+ * + * ScudoAI (hidden; replaced via ::after) + *
+ *
+ * + * Other card types (card-quote, card-stat, card-announcement) + * get analogous rules at the bottom. + * ============================================================ */ + +@import url('https://fonts.googleapis.com/css2?family=Quicksand:wght@400;500;600;700&family=Source+Serif+4:ital,opsz,wght@0,8..60,400;0,8..60,600;1,8..60,400&family=JetBrains+Mono:wght@400;500;700&display=swap'); + +:root { + --paper: #FCFAF5; + --paper-2: #F3EFE6; + --paper-3: #E8E1D0; + --ink: #1B1B1D; + --ink-muted:#55555A; + --ink-faint:#8A8A90; + --rule: #D7D2C4; + --orange: #EA541E; + --orange-2: #D6481A; + --blue: #3EB5E9; + --blue-deep:#2E8FB8; + + --font-display: 'Quicksand', system-ui, -apple-system, sans-serif; + --font-body: 'Source Serif 4', Georgia, serif; + --font-mono: 'JetBrains Mono', ui-monospace, SFMono-Regular, Menlo, monospace; +} + +html, body { + margin: 0; + padding: 0; + background: var(--paper); + color: var(--ink); + -webkit-font-smoothing: antialiased; + -moz-osx-font-smoothing: grayscale; + font-family: var(--font-body); + font-size: 18px; + line-height: 1.6; +} + +/* Hide brand-gen decorative noise -- the wro.cpp aesthetic is calm + * and editorial, not a dark deck with glow + grid backgrounds. */ +.grid-bg, .glow { display: none !important; } + +/* ---------------- Card frame (1200x1200) ---------------- */ +.card { + position: relative; + width: 1200px; + height: 1200px; + background: var(--paper); + /* Faint warm radial wash, mirrors body { background-image } in + * src/styles/global.css:44-49 -- gives the paper texture without + * looking flat. */ + background-image: + radial-gradient(900px 700px at 80% -10%, rgba(234, 84, 30, 0.06), transparent 60%), + radial-gradient(800px 600px at -10% 110%, rgba(62, 181, 233, 0.05), transparent 60%); + overflow: hidden; + box-sizing: border-box; + padding: 130px 110px 130px 110px; + display: flex; + flex-direction: column; + /* Faint inner border gives the card a printed-page feel. */ + border-bottom: 6px solid var(--orange); +} + +/* Override: theme=dark cards still use paper. We keep the editorial + * light look across the board so the wro.cpp brand stays consistent. + * If the user really wants a dark variant later, this is the hook. */ +.card.dark { background: var(--paper) !important; color: var(--ink) !important; } +.card.dark h1, .card.dark h2, .card.dark p, +.card.dark blockquote { color: var(--ink) !important; } + +/* ---------------- INSIGHT card ---------------- */ +.card-insight { + justify-content: center; + padding-top: 200px; /* reserve space for logo watermark (top:90px + 64px + safety) */ +} +.card-insight .insight-content { + display: flex; + flex-direction: column; + gap: 60px; + max-width: 980px; +} +.card-insight h1 { + font-family: var(--font-display); + font-weight: 700; + font-size: 92pt; + line-height: 1.02; + letter-spacing: -0.025em; + color: var(--ink); + margin: 0; +} +.card-insight h1.title-long { + font-size: 72pt; + line-height: 1.06; + letter-spacing: -0.02em; + max-height: 420px; +} +.card-insight h1.title-xlong { + font-size: 56pt; + line-height: 1.10; + letter-spacing: -0.015em; + max-height: 440px; +} +/* Auto-accent: ++ in C++/C++26 turns orange, sentence-final period + * (when wrapped in ) turns blue. The skill writes + * the heading with these spans for the wro.cpp signature look. */ +.card-insight h1 .accent { color: var(--orange); } +.card-insight h1 .dot { color: var(--blue-deep); } +.card-insight p { + font-family: var(--font-body); + font-size: 30pt; + line-height: 1.5; + color: var(--ink-muted); + margin: 0; + max-width: 920px; +} +.card-insight .insight-citation { + font-family: var(--font-mono); + font-size: 18pt; + letter-spacing: 0.04em; + color: var(--ink-faint); + margin-top: 40px; + padding-top: 30px; + border-top: 1px solid var(--rule); + text-transform: lowercase; +} +.card-insight .insight-citation::before { + content: '// '; + color: var(--orange); + font-weight: 700; +} + +/* ---------------- QUOTE card ---------------- */ +.card-quote { justify-content: center; padding: 140px 110px; } +.card-quote .quote-content { display: flex; flex-direction: column; align-items: flex-start; } +.card-quote .quote-mark { + font-family: var(--font-display); + font-weight: 700; + font-size: 200pt; + line-height: 0.8; + color: var(--orange); + opacity: 0.20; + /* In-flow above the text (not floated in the corner) with a negative + * bottom margin so the mark nearly touches the first quote line. */ + margin: 0 0 -0.62em; +} +.card-quote .quote-body { + font-family: var(--font-display); + font-weight: 600; + font-size: 64pt; + line-height: 1.15; + letter-spacing: -0.02em; + color: var(--ink); + max-width: 980px; + margin: 0; +} +.card-quote .quote-attribution { + margin-top: 50px; + font-family: var(--font-mono); + font-size: 18pt; + color: var(--ink-muted); +} +.card-quote .quote-attribution::before { content: '-- '; color: var(--orange); } +/* author and role are adjacent spans with no separator in the markup; + * add one so they do not run together ("wro.cppC++26 ..."). */ +.card-quote .quote-attribution .role::before { content: ' / '; color: var(--ink-faint); } + +/* ---------------- STAT card ---------------- */ +.card-stat { justify-content: center; } +/* Value and label sit side by side (label wraps to stacked lines whose + * combined height roughly matches the value); context + source span the + * full width underneath. */ +.card-stat .stat-content { + display: grid; + grid-template-columns: auto 1fr; + grid-template-areas: + "value label" + "context context" + "source source"; + column-gap: 64px; + align-items: center; +} +.card-stat .stat-value { + grid-area: value; + font-family: var(--font-display); + font-weight: 700; + font-size: 240pt; + line-height: 0.82; + letter-spacing: -0.04em; + color: var(--orange); + margin: 0; +} +.card-stat .stat-label { + grid-area: label; + font-family: var(--font-display); + font-weight: 700; + font-size: 90pt; + line-height: 1.0; + letter-spacing: -0.02em; + color: var(--ink); + /* min-content forces a per-word wrap, so "codegen steps" stacks into + * two lines next to the value instead of running on one long line. */ + width: min-content; + margin: 0; +} +.card-stat .stat-context { + grid-area: context; + font-family: var(--font-body); + font-size: 26pt; + line-height: 1.45; + color: var(--ink-muted); + margin: 56px 0 0; + max-width: 880px; +} +.card-stat .stat-source { + grid-area: source; + margin-top: 44px; + font-family: var(--font-mono); + font-size: 18pt; + color: var(--ink-faint); +} +.card-stat .stat-source::before { content: '// '; color: var(--orange); } + +/* ---------------- ANNOUNCEMENT card ---------------- */ +.card-announcement { justify-content: center; } +.card-announcement .announcement-meta { + font-family: var(--font-mono); + font-size: 20pt; + text-transform: uppercase; + letter-spacing: 0.08em; + color: var(--orange); + margin-bottom: 30px; +} +.card-announcement .announcement-content h1, +.card-announcement .announcement-content h2 { + font-family: var(--font-display); + font-weight: 700; + font-size: 78pt; + line-height: 1.05; + letter-spacing: -0.02em; + color: var(--ink); + margin: 0; +} +.card-announcement .announcement-content p { + font-family: var(--font-body); + font-size: 28pt; + line-height: 1.5; + color: var(--ink-muted); + margin: 40px 0 0; + max-width: 900px; +} + +/* ---------------- LOGO WATERMARK ---------------- */ +/* brand-gen hard-codes "ScudoAI" in .logo-wordmark; we hide it and + * draw "wro.cpp" via ::after on the watermark. The orange ".cpp" mirrors + * the brand-mark in the site header. */ +.logo-watermark { + position: absolute; + display: flex; + align-items: center; + gap: 18px; + height: 64px; + z-index: 10; + color: var(--ink); +} +.logo-watermark.top-left { top: 90px; left: 110px; } +.logo-watermark.top-right { top: 90px; right: 110px; } +.logo-watermark.bottom-left { bottom: 60px; left: 110px; } +.logo-watermark.bottom-right { bottom: 60px; right: 110px; } +.logo-watermark svg { + width: 64px; + height: 64px; + flex-shrink: 0; + /* Magnet logo PNG already has transparent paper-tone background; + * no rounded box needed -- it sits on the paper card directly. */ +} +.logo-wordmark { + font-size: 0 !important; + line-height: 0 !important; + /* Hide the upstream "ScudoAI" text node entirely. */ +} +.logo-watermark::after { + content: 'wro.cpp'; + font-family: var(--font-display); + font-weight: 700; + font-size: 28pt; + letter-spacing: -0.025em; + color: var(--ink); + line-height: 1; + /* The .cpp portion is given orange via a ::first-line trick below + * isn't reliable for inline -- so the SKILL writes the wordmark + * with explicit spans when it owns the markup. For this default + * rendering, the whole wordmark is ink and the orange comes through + * the bottom border + eyebrow. */ +} +.logo-watermark.top-right::after, +.logo-watermark.bottom-right::after { order: -1; margin-right: 18px; } + +/* ---------------- AI DISCLOSURE BADGE ---------------- */ +/* A small, honest "AI-GENERATED" label so the disclosure rides with the + * card image on social. Bottom-right corner (logo is top-left, citation is + * center-bottom, so this corner is free). Injected into index.html by + * inject.sh. See the /ai policy page. */ +.ai-badge { + position: absolute; + bottom: 74px; /* clears the 6px orange accent bar; sits in the bottom padding band */ + right: 110px; /* matches the card side padding / logo inset */ + z-index: 10; + font-family: var(--font-mono); + font-size: 14pt; + font-weight: 500; + letter-spacing: 0.10em; + text-transform: uppercase; + color: var(--ink-faint); +} + +/* ---------------- Defensive overrides ---------------- */ +/* Anything brand-gen might have set inline using its tokens. */ +* { box-sizing: border-box; } diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/caption.md b/social/facebook/enumerate-zip-iota-or-views-enumerate/caption.md new file mode 100644 index 00000000..5d4a5f89 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/caption.md @@ -0,0 +1,16 @@ +# views::enumerate is sized and common, and zip(iota, r) is neither + +## Body +A hand-written enumerate built as zip(iota, r), shown by Alex Dathskovsky, is not sized and not common on GCC 16.2 and libc++ trunk. Building a vector from it takes 2000 iterator steps over 1000 elements, and views::enumerate takes 1000. A bounded iota brings size and a common end back. libc++ 22.1 has no views::enumerate. + +https://wrocpp.github.io/posts/enumerate-zip-iota-or-views-enumerate/ + +## Hashtags +#cpp #cpp23 #ranges #wrocpp + +## Alt-text +wro.cpp card with the magnet logo. Headline reads "A zip-based enumerate has no size and no common end" with a subtitle about 2000 iterator steps against 1000 for views::enumerate. Citation footer: wro.cpp, 2026-12-26. + +## Suggested post time +Saturday 2026-12-26, 08:00 UTC +Reason: the morning slot on the publication day. diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/config.yaml b/social/facebook/enumerate-zip-iota-or-views-enumerate/config.yaml new file mode 100644 index 00000000..36f6b571 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/config.yaml @@ -0,0 +1,12 @@ +template: social/linkedin-post + +size: + width: 1200 + height: 1200 + dpr: 2 + +project: + name: "social/facebook/enumerate-zip-iota-or-views-enumerate" + author: "Filip Sajdak, Founder & CEO" + email: "office@scudoai.com" + date: "2026-01-01" diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/content.md b/social/facebook/enumerate-zip-iota-or-views-enumerate/content.md new file mode 100644 index 00000000..aa17ea01 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/content.md @@ -0,0 +1,12 @@ +--- +template: social/linkedin-post +--- + +::::post{type=insight theme=dark logo=top-left} + +:::insight{citation="wro.cpp -- 2026-12-26"} +# A zip-based enumerate has no size and no common end +Over a vector, zip(iota, r) is neither sized nor common. ranges::to takes 2000 iterator steps against 1000 for views::enumerate. A bounded iota fixes both. +::: + +:::: diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/image.png b/social/facebook/enumerate-zip-iota-or-views-enumerate/image.png new file mode 100644 index 00000000..59d9d422 Binary files /dev/null and b/social/facebook/enumerate-zip-iota-or-views-enumerate/image.png differ diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/index.html b/social/facebook/enumerate-zip-iota-or-views-enumerate/index.html new file mode 100644 index 00000000..218b3e59 --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/index.html @@ -0,0 +1,28 @@ + + + + + + social/facebook/enumerate-zip-iota-or-views-enumerate + + + + + + + + + + +

A zip-based enumerate has no size and no common end

Over a vector, zip(iota, r) is neither sized nor common. ranges::to takes 2000 iterator steps against 1000 for views::enumerate. A bounded iota fixes both.

wro.cpp -- 2026-12-26
ScudoAI
AI-generated
+ + \ No newline at end of file diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/media/.gitkeep b/social/facebook/enumerate-zip-iota-or-views-enumerate/media/.gitkeep new file mode 100644 index 00000000..e69de29b diff --git a/social/facebook/enumerate-zip-iota-or-views-enumerate/media/README.md b/social/facebook/enumerate-zip-iota-or-views-enumerate/media/README.md new file mode 100644 index 00000000..bb81c7fa --- /dev/null +++ b/social/facebook/enumerate-zip-iota-or-views-enumerate/media/README.md @@ -0,0 +1,9 @@ +# Media directory + +Drop images here (PNG, JPG, SVG) for use in your post. + +Reference them in your content.md with: + + :::image{src="screenshot.png" alt="Alt text" position="top"} + +The renderer resolves paths relative to this media/ directory. diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/README.md b/social/linkedin/enumerate-zip-iota-or-views-enumerate/README.md new file mode 100644 index 00000000..85dc2af2 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/README.md @@ -0,0 +1,130 @@ +# LinkedIn Square Post (1200x1200) + +Standard LinkedIn feed post at 1:1 square aspect ratio. Works well as a scroll-stopping image in the LinkedIn feed. + +## Usage + +```bash +brand-gen init social/linkedin-post my-post +cd my-post +# edit content.md +brand-gen build # -> index.html +brand-gen image # -> my-post.png (2400x2400 at 2x DPR) +``` + +## Directives + +All templates use one of four post types. Each uses the same directive pattern: + +### Quote + +```markdown +::::post{type=quote theme=dark logo=top-left} + +:::quote{author="Filip Sajdak" role="Founder & CEO, ScudoAI"} + +Sovereign AI is not a feature you bolt on later. It is the architecture. + +::: + +:::: +``` + +### Stat + +```markdown +::::post{type=stat theme=light accent=brand} + +:::stat{value="USD 80B" label="Worldwide sovereign cloud spend" context="Europe is the fastest-growing region at 83% YoY." source="Gartner, Feb 2026"} +::: + +:::: +``` + +### Announcement + +```markdown +::::post{type=announcement theme=light} + +:::announcement{kind="launch" date="April 2026"} + +# ScudoAI joins NVIDIA Inception + +We are thrilled to join the NVIDIA Inception program to accelerate sovereign AI deployment across European regulated industries. + +::: + +:::: +``` + +### Insight + +```markdown +::::post{type=insight theme=dark} + +:::insight{citation="EU AI Act full enforcement: 2 August 2026"} + +# The AI compliance deadline most companies are missing + +Full EU AI Act enforcement for high-risk AI systems begins in 2 August 2026. Most regulated enterprises are not ready for the documentation and audit requirements. Teams running cloud AI will face the hardest transition. + +::: + +:::: +``` + +## Attributes reference + +### `::::post` attributes + +- `type` - quote | stat | announcement | insight (required) +- `theme` - dark | light (default light) +- `accent` - brand | amber | nv (optional colored gradient background) +- `logo` - top-left | top-right | bottom-left | bottom-right (default bottom-right) + +### `:::quote` attributes + +- `author` - attribution name (appears below the quote) +- `role` - attribution role/title (appears under the author) + +### `:::stat` attributes + +- `value` - the big number (e.g. "USD 80B", "3/4", "Hours to minutes") +- `label` - short label under the value +- `context` - 1-2 sentences providing context +- `source` - small citation + +### `:::announcement` attributes + +- `kind` - launch | milestone | event | hire | partnership +- `date` - free-form date string + +### `:::insight` attributes + +- `citation` - small footer citation (e.g. regulation reference) + +## Character limits + +- **Quote body**: ~120 characters for maximum readability at thumbnail size +- **Stat label**: ~40 characters +- **Stat context**: ~200 characters +- **Announcement headline**: ~80 characters +- **Announcement body**: ~250 characters +- **Insight hook**: ~100 characters +- **Insight body**: ~300 characters + +Posts that exceed these limits will still render but typography may shrink or content may clip. + +## Media + +Add a `media/` directory to the project and reference local files: + +```markdown +:::background{image="nvidia-thor.jpg" overlay="dark"} +::: + +:::image{src="product-screenshot.png" alt="ScudoAI compliance dashboard" position="top"} +::: +``` + +Paths are relative to the project's `media/` directory. diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json new file mode 100644 index 00000000..84a94e58 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/brand-kit.json @@ -0,0 +1,11 @@ +{ + "template": "social/linkedin-post", + "bundleSource": "scudoai-social.css", + "brandKitVersion": "0.1.0", + "vendoredAt": "2026-11-17T20:50:47.409Z", + "files": [ + "scudoai.css", + "layout.css", + "logo-symbol.svg" + ] +} diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/layout.css b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/layout.css new file mode 100644 index 00000000..e69de29b diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg new file mode 100644 index 00000000..1fc44966 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/logo-symbol.svg @@ -0,0 +1,22 @@ + diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css new file mode 100644 index 00000000..612b5a12 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/assets/scudoai.css @@ -0,0 +1,343 @@ +/* ============================================================ + * wro.cpp social card -- standalone stylesheet + * + * Self-contained styling for brand-gen's `social/linkedin-post` + * output. Replaces (does NOT extend) the scudoai-social.css that + * brand-gen ships with -- the design language here is the + * wro.cpp 'Editorial Magnet' look used by src/styles/global.css, + * not the scudoai blue-indigo deck aesthetic. + * + * Source of truth for tokens: src/styles/global.css:12-30 in the + * wrocpp.github.io repo. + * + * Brand-gen markup we style: + *
+ *
(decorative, hidden) + *
(decorative, hidden) + *
+ *

...

+ *

...

+ *
...
+ *
+ *
+ * + * ScudoAI (hidden; replaced via ::after) + *
+ *
+ * + * Other card types (card-quote, card-stat, card-announcement) + * get analogous rules at the bottom. + * ============================================================ */ + +@import url('https://fonts.googleapis.com/css2?family=Quicksand:wght@400;500;600;700&family=Source+Serif+4:ital,opsz,wght@0,8..60,400;0,8..60,600;1,8..60,400&family=JetBrains+Mono:wght@400;500;700&display=swap'); + +:root { + --paper: #FCFAF5; + --paper-2: #F3EFE6; + --paper-3: #E8E1D0; + --ink: #1B1B1D; + --ink-muted:#55555A; + --ink-faint:#8A8A90; + --rule: #D7D2C4; + --orange: #EA541E; + --orange-2: #D6481A; + --blue: #3EB5E9; + --blue-deep:#2E8FB8; + + --font-display: 'Quicksand', system-ui, -apple-system, sans-serif; + --font-body: 'Source Serif 4', Georgia, serif; + --font-mono: 'JetBrains Mono', ui-monospace, SFMono-Regular, Menlo, monospace; +} + +html, body { + margin: 0; + padding: 0; + background: var(--paper); + color: var(--ink); + -webkit-font-smoothing: antialiased; + -moz-osx-font-smoothing: grayscale; + font-family: var(--font-body); + font-size: 18px; + line-height: 1.6; +} + +/* Hide brand-gen decorative noise -- the wro.cpp aesthetic is calm + * and editorial, not a dark deck with glow + grid backgrounds. */ +.grid-bg, .glow { display: none !important; } + +/* ---------------- Card frame (1200x1200) ---------------- */ +.card { + position: relative; + width: 1200px; + height: 1200px; + background: var(--paper); + /* Faint warm radial wash, mirrors body { background-image } in + * src/styles/global.css:44-49 -- gives the paper texture without + * looking flat. */ + background-image: + radial-gradient(900px 700px at 80% -10%, rgba(234, 84, 30, 0.06), transparent 60%), + radial-gradient(800px 600px at -10% 110%, rgba(62, 181, 233, 0.05), transparent 60%); + overflow: hidden; + box-sizing: border-box; + padding: 130px 110px 130px 110px; + display: flex; + flex-direction: column; + /* Faint inner border gives the card a printed-page feel. */ + border-bottom: 6px solid var(--orange); +} + +/* Override: theme=dark cards still use paper. We keep the editorial + * light look across the board so the wro.cpp brand stays consistent. + * If the user really wants a dark variant later, this is the hook. */ +.card.dark { background: var(--paper) !important; color: var(--ink) !important; } +.card.dark h1, .card.dark h2, .card.dark p, +.card.dark blockquote { color: var(--ink) !important; } + +/* ---------------- INSIGHT card ---------------- */ +.card-insight { + justify-content: center; + padding-top: 200px; /* reserve space for logo watermark (top:90px + 64px + safety) */ +} +.card-insight .insight-content { + display: flex; + flex-direction: column; + gap: 60px; + max-width: 980px; +} +.card-insight h1 { + font-family: var(--font-display); + font-weight: 700; + font-size: 92pt; + line-height: 1.02; + letter-spacing: -0.025em; + color: var(--ink); + margin: 0; +} +.card-insight h1.title-long { + font-size: 72pt; + line-height: 1.06; + letter-spacing: -0.02em; + max-height: 420px; +} +.card-insight h1.title-xlong { + font-size: 56pt; + line-height: 1.10; + letter-spacing: -0.015em; + max-height: 440px; +} +/* Auto-accent: ++ in C++/C++26 turns orange, sentence-final period + * (when wrapped in ) turns blue. The skill writes + * the heading with these spans for the wro.cpp signature look. */ +.card-insight h1 .accent { color: var(--orange); } +.card-insight h1 .dot { color: var(--blue-deep); } +.card-insight p { + font-family: var(--font-body); + font-size: 30pt; + line-height: 1.5; + color: var(--ink-muted); + margin: 0; + max-width: 920px; +} +.card-insight .insight-citation { + font-family: var(--font-mono); + font-size: 18pt; + letter-spacing: 0.04em; + color: var(--ink-faint); + margin-top: 40px; + padding-top: 30px; + border-top: 1px solid var(--rule); + text-transform: lowercase; +} +.card-insight .insight-citation::before { + content: '// '; + color: var(--orange); + font-weight: 700; +} + +/* ---------------- QUOTE card ---------------- */ +.card-quote { justify-content: center; padding: 140px 110px; } +.card-quote .quote-content { display: flex; flex-direction: column; align-items: flex-start; } +.card-quote .quote-mark { + font-family: var(--font-display); + font-weight: 700; + font-size: 200pt; + line-height: 0.8; + color: var(--orange); + opacity: 0.20; + /* In-flow above the text (not floated in the corner) with a negative + * bottom margin so the mark nearly touches the first quote line. */ + margin: 0 0 -0.62em; +} +.card-quote .quote-body { + font-family: var(--font-display); + font-weight: 600; + font-size: 64pt; + line-height: 1.15; + letter-spacing: -0.02em; + color: var(--ink); + max-width: 980px; + margin: 0; +} +.card-quote .quote-attribution { + margin-top: 50px; + font-family: var(--font-mono); + font-size: 18pt; + color: var(--ink-muted); +} +.card-quote .quote-attribution::before { content: '-- '; color: var(--orange); } +/* author and role are adjacent spans with no separator in the markup; + * add one so they do not run together ("wro.cppC++26 ..."). */ +.card-quote .quote-attribution .role::before { content: ' / '; color: var(--ink-faint); } + +/* ---------------- STAT card ---------------- */ +.card-stat { justify-content: center; } +/* Value and label sit side by side (label wraps to stacked lines whose + * combined height roughly matches the value); context + source span the + * full width underneath. */ +.card-stat .stat-content { + display: grid; + grid-template-columns: auto 1fr; + grid-template-areas: + "value label" + "context context" + "source source"; + column-gap: 64px; + align-items: center; +} +.card-stat .stat-value { + grid-area: value; + font-family: var(--font-display); + font-weight: 700; + font-size: 240pt; + line-height: 0.82; + letter-spacing: -0.04em; + color: var(--orange); + margin: 0; +} +.card-stat .stat-label { + grid-area: label; + font-family: var(--font-display); + font-weight: 700; + font-size: 90pt; + line-height: 1.0; + letter-spacing: -0.02em; + color: var(--ink); + /* min-content forces a per-word wrap, so "codegen steps" stacks into + * two lines next to the value instead of running on one long line. */ + width: min-content; + margin: 0; +} +.card-stat .stat-context { + grid-area: context; + font-family: var(--font-body); + font-size: 26pt; + line-height: 1.45; + color: var(--ink-muted); + margin: 56px 0 0; + max-width: 880px; +} +.card-stat .stat-source { + grid-area: source; + margin-top: 44px; + font-family: var(--font-mono); + font-size: 18pt; + color: var(--ink-faint); +} +.card-stat .stat-source::before { content: '// '; color: var(--orange); } + +/* ---------------- ANNOUNCEMENT card ---------------- */ +.card-announcement { justify-content: center; } +.card-announcement .announcement-meta { + font-family: var(--font-mono); + font-size: 20pt; + text-transform: uppercase; + letter-spacing: 0.08em; + color: var(--orange); + margin-bottom: 30px; +} +.card-announcement .announcement-content h1, +.card-announcement .announcement-content h2 { + font-family: var(--font-display); + font-weight: 700; + font-size: 78pt; + line-height: 1.05; + letter-spacing: -0.02em; + color: var(--ink); + margin: 0; +} +.card-announcement .announcement-content p { + font-family: var(--font-body); + font-size: 28pt; + line-height: 1.5; + color: var(--ink-muted); + margin: 40px 0 0; + max-width: 900px; +} + +/* ---------------- LOGO WATERMARK ---------------- */ +/* brand-gen hard-codes "ScudoAI" in .logo-wordmark; we hide it and + * draw "wro.cpp" via ::after on the watermark. The orange ".cpp" mirrors + * the brand-mark in the site header. */ +.logo-watermark { + position: absolute; + display: flex; + align-items: center; + gap: 18px; + height: 64px; + z-index: 10; + color: var(--ink); +} +.logo-watermark.top-left { top: 90px; left: 110px; } +.logo-watermark.top-right { top: 90px; right: 110px; } +.logo-watermark.bottom-left { bottom: 60px; left: 110px; } +.logo-watermark.bottom-right { bottom: 60px; right: 110px; } +.logo-watermark svg { + width: 64px; + height: 64px; + flex-shrink: 0; + /* Magnet logo PNG already has transparent paper-tone background; + * no rounded box needed -- it sits on the paper card directly. */ +} +.logo-wordmark { + font-size: 0 !important; + line-height: 0 !important; + /* Hide the upstream "ScudoAI" text node entirely. */ +} +.logo-watermark::after { + content: 'wro.cpp'; + font-family: var(--font-display); + font-weight: 700; + font-size: 28pt; + letter-spacing: -0.025em; + color: var(--ink); + line-height: 1; + /* The .cpp portion is given orange via a ::first-line trick below + * isn't reliable for inline -- so the SKILL writes the wordmark + * with explicit spans when it owns the markup. For this default + * rendering, the whole wordmark is ink and the orange comes through + * the bottom border + eyebrow. */ +} +.logo-watermark.top-right::after, +.logo-watermark.bottom-right::after { order: -1; margin-right: 18px; } + +/* ---------------- AI DISCLOSURE BADGE ---------------- */ +/* A small, honest "AI-GENERATED" label so the disclosure rides with the + * card image on social. Bottom-right corner (logo is top-left, citation is + * center-bottom, so this corner is free). Injected into index.html by + * inject.sh. See the /ai policy page. */ +.ai-badge { + position: absolute; + bottom: 74px; /* clears the 6px orange accent bar; sits in the bottom padding band */ + right: 110px; /* matches the card side padding / logo inset */ + z-index: 10; + font-family: var(--font-mono); + font-size: 14pt; + font-weight: 500; + letter-spacing: 0.10em; + text-transform: uppercase; + color: var(--ink-faint); +} + +/* ---------------- Defensive overrides ---------------- */ +/* Anything brand-gen might have set inline using its tokens. */ +* { box-sizing: border-box; } diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/caption.md b/social/linkedin/enumerate-zip-iota-or-views-enumerate/caption.md new file mode 100644 index 00000000..4054d760 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/caption.md @@ -0,0 +1,20 @@ +# views::enumerate is sized and common, and zip(iota, r) is neither + +## Body +Alex Dathskovsky showed a C++23 enumerate adaptor built as zip(iota, r) in a range_adaptor_closure. It is a clean example of that class, and this post asks what std::views::enumerate does differently. + +On GCC 16.2 and libc++ trunk, the zip version over a vector is not a sized_range and not a common_range, and views::enumerate is both. So size() does not compile on it, and neither does the iterator-pair vector constructor. ranges::to allocates once for both, but takes 2000 iterator steps over 1000 elements against 1000 for views::enumerate. Bounding the iota by the size of the range brings back size, a common end and a signed index. + +libc++ 22.1 has no views::enumerate, and libc++ trunk has it. I did not run MSVC, input-only ranges or wall-clock time. + +https://wrocpp.github.io/posts/enumerate-zip-iota-or-views-enumerate/ + +## Hashtags +#cpp #cpp23 #ranges #zip #wrocpp + +## Alt-text +wro.cpp card with the magnet logo. Headline reads "A zip-based enumerate has no size and no common end" with a subtitle about 2000 iterator steps against 1000 for views::enumerate. Citation footer: wro.cpp, 2026-12-26. + +## Suggested post time +Saturday 2026-12-26, 08:00 UTC +Reason: the morning slot on the publication day. diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/config.yaml b/social/linkedin/enumerate-zip-iota-or-views-enumerate/config.yaml new file mode 100644 index 00000000..78ed59d1 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/config.yaml @@ -0,0 +1,12 @@ +template: social/linkedin-post + +size: + width: 1200 + height: 1200 + dpr: 2 + +project: + name: "social/linkedin/enumerate-zip-iota-or-views-enumerate" + author: "Filip Sajdak, Founder & CEO" + email: "office@scudoai.com" + date: "2026-01-01" diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/content.md b/social/linkedin/enumerate-zip-iota-or-views-enumerate/content.md new file mode 100644 index 00000000..aa17ea01 --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/content.md @@ -0,0 +1,12 @@ +--- +template: social/linkedin-post +--- + +::::post{type=insight theme=dark logo=top-left} + +:::insight{citation="wro.cpp -- 2026-12-26"} +# A zip-based enumerate has no size and no common end +Over a vector, zip(iota, r) is neither sized nor common. ranges::to takes 2000 iterator steps against 1000 for views::enumerate. A bounded iota fixes both. +::: + +:::: diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/image.png b/social/linkedin/enumerate-zip-iota-or-views-enumerate/image.png new file mode 100644 index 00000000..68c9b910 Binary files /dev/null and b/social/linkedin/enumerate-zip-iota-or-views-enumerate/image.png differ diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/index.html b/social/linkedin/enumerate-zip-iota-or-views-enumerate/index.html new file mode 100644 index 00000000..a4c0e62c --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/index.html @@ -0,0 +1,28 @@ + + + + + + social/linkedin/enumerate-zip-iota-or-views-enumerate + + + + + + + + + + +

A zip-based enumerate has no size and no common end

Over a vector, zip(iota, r) is neither sized nor common. ranges::to takes 2000 iterator steps against 1000 for views::enumerate. A bounded iota fixes both.

wro.cpp -- 2026-12-26
ScudoAI
AI-generated
+ + \ No newline at end of file diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/media/.gitkeep b/social/linkedin/enumerate-zip-iota-or-views-enumerate/media/.gitkeep new file mode 100644 index 00000000..e69de29b diff --git a/social/linkedin/enumerate-zip-iota-or-views-enumerate/media/README.md b/social/linkedin/enumerate-zip-iota-or-views-enumerate/media/README.md new file mode 100644 index 00000000..bb81c7fa --- /dev/null +++ b/social/linkedin/enumerate-zip-iota-or-views-enumerate/media/README.md @@ -0,0 +1,9 @@ +# Media directory + +Drop images here (PNG, JPG, SVG) for use in your post. + +Reference them in your content.md with: + + :::image{src="screenshot.png" alt="Alt text" position="top"} + +The renderer resolves paths relative to this media/ directory. diff --git a/src/content/posts/2026-12-26-enumerate-zip-iota-or-views-enumerate.mdx b/src/content/posts/2026-12-26-enumerate-zip-iota-or-views-enumerate.mdx new file mode 100644 index 00000000..e21f661a --- /dev/null +++ b/src/content/posts/2026-12-26-enumerate-zip-iota-or-views-enumerate.mdx @@ -0,0 +1,124 @@ +--- +title: "views::enumerate is sized and common, and zip(iota, r) is neither" +slug: "enumerate-zip-iota-or-views-enumerate" +pubDate: 2026-12-26 +author: filip-sajdak +language: en +kind: short +audience: working-cpp +tags: [cpp, cpp23, ranges, zip, enumerate, stdlib, gcc] +summary: "Alex Dathskovsky's hand-written enumerate is zip(iota, r) in a range_adaptor_closure. On GCC 16.2 and libc++ trunk it has no size and no common end, so std::vector cannot take its iterator pair, and ranges::to walks it twice, 2000 iterator steps against 1000 for views::enumerate. A bounded iota brings both back. libc++ 22.1 has no views::enumerate." +draft: false +relatedPosts: [generator-enumerate-by-value, gcc-16-2-toolchain-status, views-as-input-before-filter] +--- + +import GodboltEmbed from '../../components/GodboltEmbed.astro'; +import PostLink from '../../components/PostLink.astro'; + +Alex Dathskovsky {/* LINK: Alex Dathskovsky LinkedIn post, URL from the user */} showed a C++23 enumerate adaptor that zips an `iota` with the range, inside a `range_adaptor_closure`. It is a clean example of that class. This post asks what `std::views::enumerate` does differently. +On GCC 16.2 the hand-written one has no size and no common end. Building a vector from it takes 2000 iterator steps over 1000 elements, and the standard one takes 1000. +Everything below ran on Compiler Explorer on 17 November 2026. I know the outside post only from screenshots, and the code in it is as I read it there. + +## In short + +- The hand-written zip over a `std::vector` is not sized and not common. `views::enumerate` is both. Both are random access and borrowed. +- The hand-written index is the type you pass to `iota`, unsigned in the example. The standard index is the difference type, signed. +- The hand-written view has no `size()`. It also fails with the iterator-pair vector constructor and with `std::count_if`. The errors are quoted below. +- `ranges::to` allocates once for both, but takes 2000 iterator steps for the hand-written one and 1000 for the standard one. +- A bounded `iota` brings back size, common end and the signed index. libc++ 22.1 has no `views::enumerate`, and libc++ trunk has it. + +## What the hand-written adaptor is + +The adaptor in the outside post, as I read it, passes `views::iota(Index{0})` and the container to `views::zip`. Their screenshot sets `using Index = std::size_t;`, and my copy does the same: + +```cpp +struct Enumerate : std::ranges::range_adaptor_closure { + constexpr auto operator()(Iterable auto&& c) const { + return std::views::zip(std::views::iota(std::size_t{0}), std::forward(c)); + } +}; +``` + +It gives the right elements. A loop over a vector of three printed `0:10 1:20 2:30`, and so did `views::enumerate`. +Alex's earlier `std::generator` version is in the by-value generator post. This one only compares the two ways of building the pairs from ranges. + +## The property table + +I asked the compiler about the same `std::vector` through three adaptors, with `static_assert` in the example and a printed table. The third column is a variant of the first, described further down. + +| property | `zip(iota, r)` | `views::enumerate` | `zip(bounded iota, r)` | +|---|---|---|---| +| `sized_range` | no | yes | yes | +| `common_range` | no | yes | yes | +| `random_access_range` | yes | yes | yes | +| `borrowed_range` | yes | yes | yes | +| index type | `std::size_t` | `std::ptrdiff_t` | `std::ptrdiff_t` | + +The table is the same on libstdc++ 16 with GCC 16.2, libstdc++ 15 with Clang 22.1, and libc++ 24.0 with Clang trunk. The `` feature-test macro `__cpp_lib_ranges_enumerate` is 202302 on all three. In the 22.1 release of libc++ it is not defined, and the same file fails: + +```text +error: no member named 'enumerate' in namespace 'std::ranges::views'; did you mean simply 'enumerate'? +``` + +So a project on libc++ 22.1 needs a hand-written adaptor. libc++ trunk, version 24.0 on Compiler Explorer, does not. + +## Why it is neither sized nor common + +With one argument, `iota` is unbounded: its sentinel is `std::unreachable_sentinel_t`, and [cppreference says](https://en.cppreference.com/w/cpp/ranges/iota_view) it is not sized. A `zip_view` has a size [only if every adapted range is sized](https://en.cppreference.com/w/cpp/ranges/zip_view). The unbounded `iota` is not, so the zip is not. +The same page lists when a `zip_view` is common. The case that would cover a vector requires every adapted view to be random access and sized, and again the `iota` is not sized. [`enumerate_view`](https://en.cppreference.com/w/cpp/ranges/enumerate_view) is sized when its base is, and common when its base is common, forward and sized. +That matches the table. + +## What the loss costs + +On GCC 16.2 and libc++ trunk, three ordinary calls on `auto z = v | enumerate;` do not compile. + +```text +z.size() + g++ 16.2: error: no matching function for call to 'std::ranges::zip_view, ...>::size()' +std::vector> w(z.begin(), z.end()) + libc++ trunk: candidate template ignored: deduced conflicting types for parameter '_InputIterator' ('__iterator' vs. '__sentinel') +std::count_if(z.begin(), z.end(), pred) + g++ 16.2: error: no matching function for call to 'count_if(..._Iterator, ..._Sentinel, ...)' +``` + +The cause is one fact: `begin()` and `end()` have different types, and the classic container constructors and algorithms want one iterator type. `views::enumerate` has the same type for both. The example asserts that the vector constructor accepts its pair. I did not run `count_if` on it. `std::ranges::to` accepts the hand-written view too, as the next table shows. + +A call that compiles can still cost more. I counted the steps of a counting iterator, a sized and common view over 1000 ints, wrapped in each adaptor and fed to `std::ranges::to`. Each of the three allocated once. + +| adaptor | allocations | iterator steps | +|---|---|---| +| `zip(iota, r)` | 1 | 2000 | +| `views::enumerate` | 1 | 1000 | +| `zip(bounded iota, r)` | 1 | 1001 on libstdc++ 16, 1000 on libc++ trunk | + +The hand-written view still gets a single allocation, so it does reserve. It takes twice the steps. That is consistent with a counting pass before the copy, because the view has no size to read. I did not read the library source to confirm the mechanism. +The counts did not change between three optimization levels on GCC 16.2. I did not time anything. + +## A bounded iota brings it back + +When the range is sized, bound the `iota` by its size and count in the range's difference type: + +```cpp +struct EnumerateSized : std::ranges::range_adaptor_closure { + template + requires std::ranges::sized_range + constexpr auto operator()(R&& c) const { + using Index = std::ranges::range_difference_t; + return std::views::zip(std::views::iota(Index{0}, static_cast(std::ranges::size(c))), std::forward(c)); + } +}; +``` + +The result is sized, common, random access and borrowed, with the signed index, as the third table column shows. `size()` works and the vector constructor takes the iterator pair on GCC 16.2. +On libstdc++ 15 with Clang 22.1 the constructor rejects the pair. That claim holds for libstdc++ 16.2 and libc++ trunk only. + +One difference remains. The bounded adaptor is constrained to sized ranges, so its `requires` clause rejects an input-only range or a filter view. I did not try those. `views::enumerate` takes any input range, and it is the standard's choice for new code. + + + +## Which to use + +On a library with `views::enumerate`, use it: it is sized, common, and takes half the steps in `ranges::to`. Without it, the hand-written adaptor still works for a loop. For a sized range, bound the `iota`, and keep the index signed if you compare it with sizes. +The sources, outputs and the scratch errors are in [`posts/136-enumerate-zip-or-views`](https://github.com/wrocpp/cpp26-reflection-examples/tree/main/posts/136-enumerate-zip-or-views). + +What I did not run: MSVC, a `filter_view` or an input-only source for either adaptor, wall-clock time, and the library source behind the step counts.