Skip to content

docs(readme): section links under the badges and a star-history chart - #43

Merged
alexshapalov merged 10 commits into
mainfrom
docs/readme-nav-star-history
Nov 3, 2026
Merged

alexshapalov merged 10 commits into
mainfrom
docs/readme-nav-star-history

Conversation

@alexshapalov

@alexshapalov alexshapalov commented Nov 2, 2026 •

Copy link
Copy Markdown
Contributor

README changes, matching the pgbot and orca README layouts:

  • A centered row of section links right under the badges: Quickstart · Agents · Why PGX · Benchmarks · Architecture · Status · Docs. All anchors point at existing headings.
  • A star-history chart (star-history.com, pgrundev/pgx, as of 2026-11-02) at the very end of the README, committed as docs/pgx/img/star-history.png and linked to the live star-history page.
  • A compact README: 661 → ~480 lines, with no text dropped from the repo.
    • Why PGX, At a glance, Disposable Postgres and Built for agents merge into one Why PGX section.
    • Benchmarks keeps the headline numbers and the caveats (regression figure is not a compat %, not faster than PG in general); the tables, isolation results and reproduction steps move verbatim to docs/pgx/benchmarks.md.
    • Build from source, Run a PGX runtime and Make a template and mint databases move verbatim to docs/pgx/local-dev/build-from-source.md; install.md now points there instead of at the README.
    • Status absorbs PGX + PGRun; Quickstart folds the branching subsection into the command list.
  • Architecture replaces How it works as an orca-style feature wall: five rows, text left, visual right.
    • One runtime, thousands of databases carries a recording of the density grid as a GIF (docs/pgx/img/arch-density.gif, 4.2 MB, 600 px, 6 fps) with a JPG poster for clients that do not play GIFs.
    • Mint-on-connect, Copy-on-write storage, Limits at every level and A branch per agent use text diagrams with the measured numbers; swap in a recording for each when one exists.
    • The limits example uses the real error text and the real pgrust_runtime_status() field names from the docs.

All relative links and nav anchors were checked by script. No code changes.

🤖 Generated with Claude Code

alexshapalov and others added 10 commits November 2, 2026 13:41
… at the end

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…urce to docs

README goes from 661 to 374 lines with no text dropped from the repo:

- Why PGX, At a glance, Disposable Postgres and Built for agents merge into
  one Why PGX section.
- Benchmarks keeps the headline numbers and the caveats; the tables,
  isolation results and the reproduction steps move verbatim to
  docs/pgx/benchmarks.md.
- Build from source, Run a PGX runtime and Make a template and mint
  databases move verbatim to docs/pgx/local-dev/build-from-source.md;
  install.md now points there instead of at the README.
- Status absorbs PGX + PGRun; Quickstart folds the branching subsection
  into the command list.
- The nav row shrinks to seven links; the Documentation table lists the
  two new pages.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…y demo as a GIF

Replaces How it works with an Architecture table in the style of the
orca README: five rows, text on the left and a visual on the right. The
first row, one runtime holding 1,000 databases, carries a recording of
the density grid (docs/pgx/img/arch-density.gif, 600 px, 6 fps, with a
JPG poster for clients that do not play GIFs). The other four rows,
mint-on-connect, copy-on-write storage, limits at every level and a
branch per agent, use text diagrams with the measured numbers until a
recording exists for each.

The limits example uses the real error text from memory-limits.md and
the real pgrust_runtime_status() field names from pgrun-interface.md.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…one run

Drops the two-column table for one item after another: heading, text,
visual, link. Adds a second item, 1,000 databases on a MacBook in
46.6 seconds (16 GB M1 Pro, pgx clone of a 31 MB app database: 41 ms
median, 0 errors, engine RSS 20 MB idle with all 1,000, 340 MB at 100
concurrent readers), which carries the density recording; the first
item, 1,000 databases in one process, uses the measured memory table.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…tion

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
@alexshapalov
alexshapalov merged commit 66cc4b2 into main Nov 3, 2026
10 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant