Commit Graph

439 Commits

Author SHA1 Message Date
lilleman c400ed7908 Make great documentation goal 1.2, and renumber goals 13 to 16 as 12 to 15 wherever they are cited
Tests / vet + fmt + tests (pull_request) Has been cancelled
Tests / Known vulnerabilities in this Go version and the required modules (pull_request) Has been cancelled
Tests / Release from CHANGELOG.md (pull_request) Has been cancelled
Tests / packages (pull_request) Has been cancelled
2026-10-14 22:55:20 +02:00
lilleman 79dafa38f0 Make goal 12 great documentation, state how the README meets it, and plan the README's reorganisation for 0.2.0
Tests / vet + fmt + tests (pull_request) Has been cancelled
Tests / Known vulnerabilities in this Go version and the required modules (pull_request) Has been cancelled
Tests / Release from CHANGELOG.md (pull_request) Has been cancelled
Tests / packages (pull_request) Has been cancelled
2026-10-14 22:50:30 +02:00
lilleman 7739c09163 Check the plan before filing into it, keep filed items in order, name the filer's tests in the gate, and record the merge over a worse ruling 2026-10-14 13:51:22 +02:00
lilleman 5a6928750c Give the vulnerability filer a folder of its own, and leave its item's score to rank_todo.py 2026-10-14 13:51:22 +02:00
lilleman 15965e7a77 State the filer's contract with the plan where it is read, name the step an outage fails, and mark an unfixed vulnerability by name 2026-10-14 13:51:22 +02:00
lilleman 2d86f4f0c0 Answer architecture round 2: 1 🔴, 3 🟡; read the report and edit the plan in two functions, propose only while no pull request is open, and test filing into the repository's own todo.md 2026-10-14 13:51:22 +02:00
lilleman 273dd45128 Answer architecture round 1: 0 🔴, 5 🟡; a job of its own that only releases wait for, filing each vulnerability atop the plan by a pull request, a recorded decision, one gate list, and Renovate on the go run tools 2026-10-14 13:51:22 +02:00
lilleman 117f685939 Fail the gate, daily and before each release, on a known vulnerability in the Go release or the module 2026-10-14 13:51:22 +02:00
lilleman fbadb99634 Answer architecture round 2: 0 🔴, 0 🟡; tie the token choice to the account type, and keep GHCR's reason in the registry decision 2026-10-14 11:33:08 +02:00
lilleman 508e93d11a Answer architecture round 1: 0 🔴, 3 🟡; copy the image to Docker Hub in a job of its own, give the registries their own decision, and name the narrowest token 2026-10-14 11:33:08 +02:00
lilleman f3fbbac0d9 Push the image to Docker Hub beside GHCR, so docker run larvit/fejkdata works 2026-10-14 11:33:08 +02:00
lilleman 717f9f2abc Drop the AUR until it reopens registration, and plan it as item 252 2026-10-14 09:49:01 +02:00
lilleman 2885f7b571 Answer prose round 1: 0 🔴, 4 🟡; name the package-repository checks, list the release secrets, and point each sentence at its actor 2026-10-14 09:49:01 +02:00
lilleman b9ae602399 Answer stability round 1: 0 🔴, 3 🟡; fail when the latest release can't be read, build every platform's binaries when Go changes, and test main() and the Unreleased heading 2026-10-14 09:49:01 +02:00
lilleman a82279c6a4 Answer product-owner round 1: 0 🔴, 1 🟡; check the apk repository on Alpine 3.24's apk-tools 3, and say how macOS runs a downloaded binary 2026-10-14 09:49:01 +02:00
lilleman 895c0e977f Answer architecture round 4: 0 🔴, 1 🟡; run the repository checks on every file they read, and check each repository installs the release's own version 2026-10-14 09:49:01 +02:00
lilleman e6e06d007b Answer architecture round 3: 0 🔴, 4 🟡; keep signing keys out of image layers, and build and install from the package repositories in the gate with throwaway keys 2026-10-14 09:49:01 +02:00
lilleman f08182fb94 Answer architecture round 2: 1 🔴, 0 🟡; self-host signed apt, dnf and apk repositories on GitHub Pages in place of Cloudsmith, and keep an older tag's image while leaving latest alone 2026-10-14 09:49:01 +02:00
lilleman cc7d4fb972 Answer architecture round 1: 0 🔴, 7 🟡; build with the gate's Go, guard single-version channels on older tags, record the Cloudsmith tie, and return the tag from publish() 2026-10-14 09:49:01 +02:00
lilleman e67c858d41 Ship prebuilt binaries, deb, rpm and apk packages on Cloudsmith, a GHCR image, and Homebrew, Scoop and AUR packages with each release 2026-10-14 09:49:01 +02:00
lilleman b8b4188af7 Answer prose round 1: the id bullet sets apart the test series and the real-looking SSNs, the own-data bullet names a path rather than a category, the stdin heading names all it covers, and doc.go's namespace sentence has its subject 2026-10-13 00:42:45 +02:00
lilleman 5f8d55e2bb Answer product-owner round 2: the reproducible bullet states its goroutine condition, and the table of contents names the Data format heading it links 2026-10-13 00:42:45 +02:00
lilleman df46340af2 Answer product-owner round 1: the front page's data and id bullets say what ships and what may be real, the Library section links the terms Data defines, and doc.go's synopsis matches the tagline 2026-10-13 00:42:45 +02:00
lilleman 82fb611c74 Open the README with badges, what ships and a table of contents, put the stdin byte rules under a heading of their own, and move the Library section before Data 2026-10-13 00:42:45 +02:00
lilleman bf61527363 Answer prose round 1: the seed advice says to log a new seed in the error, the doc and the README alike, the changelog says each run takes a new seed, New's failures read in one pass, and the concurrency lines say who draws 2026-10-12 18:34:15 +02:00
lilleman e0f5493daa Answer product-owner round 1: the seed docs say a new seed each run, not draws afresh, the README says to log a drawn seed to replay a run, and the changelog names WithSeed and the data options apart 2026-10-12 18:34:15 +02:00
lilleman cde9e354c5 Answer architecture round 2: New's doc comment wraps whole, the README's stream needs no seeded, and the system-seed test reports each run's exit code and stderr 2026-10-12 18:34:15 +02:00
lilleman 8e8fa92e77 Answer architecture round 1: Main draws the CLI's default seed and passes it to run, New names every option mistake at once, dataGiven says what it is, and the docs drop seeded where every sequence is 2026-10-12 18:34:15 +02:00
lilleman 90c804f6e6 Fail New without WithSeed, naming it, and seed the CLI from the system without --seed 2026-10-12 18:34:15 +02:00
lilleman 1a5b632b90 Answer prose round 1: List gives direct descents only, the library and CLI lines name their verbs, the template line explains its values, the geo line says which places ship, and the README's pointer rule reads in one pass 2026-10-12 18:07:39 +02:00
lilleman 143311270c Answer product-owner round 1: the release notes link the tagged README, say how to install and pin, name 64-bit targets, List and IsTemplate, the README says a nullable column needs a pointer, and item 89 names the changelog lines it changes 2026-10-12 18:07:39 +02:00
lilleman 70fdfbd591 Answer prose round 1: the README lists the release tooling's tests in the gate, and the decision says the size test checks the premise 2026-10-12 17:50:30 +02:00
lilleman 870082c12f Read an existing tag through git/ref/tags in publish_release.py, following an annotated tag to its commit 2026-10-12 17:50:30 +02:00
lilleman 71b0754679 Answer prose round 1: the custom-CLI test advice says why an indexed module needs each path rendered, the help ruling says the help points to --list and who ruled, ErrNoColumns's doc says what it matches, and the filed items read in one pass 2026-10-11 23:39:14 +02:00
lilleman 5e6330c75b Answer product-owner round 1: the hint for an empty stdin points to --list, the help's examples are templates only, the README's record id is an integer, and a test over an indexed module renders each path 2026-10-11 23:39:14 +02:00
lilleman 90d1948a2b Answer stability round 1: the tests precede the code they pin, ErrNoColumns's doc speaks of itself alone, the edited docs are reflowed, and the binding hint and a JSON-shaped template's parse position are filed as items 228 and 229 2026-10-11 23:39:14 +02:00
lilleman 1edc3b57dd Show a struct filling an address from one pick, a record template compiled once, a counter per table and how a custom CLI's embed maps to paths, and record the custom-CLI builder's ruling on the help 2026-10-11 23:39:14 +02:00
lilleman adba2510db Answer prose round 1: the CLI decision says what keeping cli off internal/ buys, the changelog and Library sentences hold one idea each, and item 185 names the three wordings its builder rules on 2026-10-11 21:19:57 +02:00
lilleman 5b2bb7e681 Answer architecture round 2: every error a custom CLI prints names it, the help marks where the name goes, the version line checks the main module first, and the README embed compiles for data without a table 2026-10-11 21:19:57 +02:00
lilleman 9528ecd8b0 Answer architecture round 1: a custom CLI names itself by its file, --version names its module beside fejkdata's and follows a replace, the module path comes from the library's type, the end-to-end test runs main in a subprocess, the CLI tests find the shipped modules by their manifests, and Your own CLI sits under Library 2026-10-11 21:19:57 +02:00
lilleman 35f1eeffb2 Make the CLI the library call cli.Main, with fejkdata's own main one line over data.Modules() 2026-10-11 21:19:57 +02:00
lilleman 9b7c3f0dc2 Answer prose round 1: goal 2.1 is dropped for goal 2, the no-street decision rests on goal 2.2 and names the item ending it, and en_US stays out of the move 2026-10-11 20:37:49 +02:00
lilleman 66cd1e0605 Answer prose round 1: the sources name Wikidata, the decision title reads in one pass, and the street rule's threshold says why it is five 2026-10-11 19:13:48 +02:00
lilleman 1f9b201633 Answer product-owner round 2: a street joining names with a bare comma is dropped, and the US places read as a drawn tenth plus the capitals and big places 2026-10-11 19:13:48 +02:00
lilleman d3d9547876 Answer stability round 1: the docs say the US tenth is drawn from every place, a capital matches by a name only one place in its state carries, and the test comment says what it checks 2026-10-11 19:13:48 +02:00
lilleman 2132bde2ba Record the shares, the capitals' source and the measured cost, and close items 199 and 200 2026-10-11 19:13:48 +02:00
lilleman af49b83fa9 Re-pin the seeded sv_SE.address example 2026-10-11 19:13:48 +02:00
lilleman 7f5a0cb63c Answer prose round 1: one name for a place holding most of a postal code's land, the decision leaves the seeding to the code, and the test says what it checks 2026-10-11 16:42:23 +02:00
lilleman 66b5d699fa Answer stability round 1: each US place's draw is seeded by its GEOID, so a changed source moves only the places near the cut; item 78 leaves the plan and the multi-county fetch is filed 2026-10-11 16:42:23 +02:00
lilleman 5db699469b Record the shipped places and streets as a decision, and say so in the README and changelog 2026-10-11 16:42:23 +02:00