Wiki, CONTRIBUTING and planning documents: eleven passages describe the project before changes that have landed #406
Labels
No labels
accessibility
authentication
breaking change
bug
documentation
enhancement
interface
internationalisation
observability
performance
security
tier
1
tier
2
tier
3
tier/4
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
Postulo/postulo#406
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Found on 2026-09-29 in a read-only review of postulo at
95348908dand the wiki at de1f52467, reading the user, contributor and planning documents against the settings, forms, compose file, workflows,pyproject.tomland the issue tracker.Checked again on 2026-10-01 against
mainat0a5cad587: still present. The code is now ataccounts/forms.py:450(AccessibilityForm) and:465(the label), andjobs/forms.py:142-146(the upload's help text). InCONTRIBUTING.mdthe Scanning the image section is now at:520-552(its first sentence at:522), the two earlier passages at:391and:418, and the Translations paragraph at:285-287. In.forgejo/workflows/ci.ymlthe secondactions/cache@v4is now at:405andactions/upload-artifact@v3at:434. Assumption 3 indocs/PLAN.mdis now at:328-335.Eleven passages describe Postulo as it was before a change that has since landed. None is a defect in the code. Each sends somebody to a place that no longer exists, gives a command that fails, says something is missing that is there, or tells a contributor something false about what the build runs and what is planned; two of them contradict another section of the same file.
Related to #281, #156 and #255, all closed. #281 moved two settings to Settings → Accessibility and #156 put the image scan in the pipeline, but the pages that described the earlier state were not all updated (sections 1 and 6); #255 revised
docs/PLAN.md, and sections 8 and 9 slipped past that revision or were overtaken by #247. This issue lists what those three left behind, and the passages of the same kind left by #264, #235, #247 and #269.Paths in sections 1 to 5 are pages of the wiki; the rest are in the postulo repository.
1. Wiki: the accessibility choices are no longer under Appearance
Since #281 the order number on career entries and the single-key shortcuts live on Settings → Accessibility (
core/settings_sections.py:44-51;accounts/forms.py:449,AccessibilityForm, withshow_career_order,keyboard_shortcutsandnav_underline). Three pages still send people to Appearance:Getting-started.md:38-41describes them inside the Appearance bullet, and the list of sections has no Accessibility bullet at all.Capturing-postings.md:119says "Settings → Appearance → Keyboard shortcuts". The setting is now under Accessibility and is labelled "Let a single key do something" (accounts/forms.py:464).Your-career-record.md:507-508says "under Settings → Appearance".2. Wiki: company logos accept SVG and keep their size
Tracking-applications.md:77-79says "SVG is refused for now" and that whatever arrives is "re-encoded to a 256-pixel square". Since #264 an SVG is sanitised and kept as a vector, and a bitmap is re-encoded at its own size (jobs/forms.py:141-145, the upload's help text;jobs/logos.py:16-22).3. Wiki: tagged PDFs are listed as still wanted
Roadmap.md:102-103lists "Tagged PDFs" under Still wanted, not scheduled and links to Accessibility → Known gaps.documents/pdf.py:83renderspdf/ua-1,Accessibility.md:187says "They are tagged", and its Known gaps no longer mentions PDFs (#235).4. Wiki:
docker logs postulonames no containerTroubleshooting.md:17saysdocker logs postulo.docker/compose.yml:19sets the projectname: postulo, the service ispostuloand there is nocontainer_name, so the container ispostulo-postulo-1and the command answers "No such container". Every other page usesdocker compose -f docker/compose.yml ….5. Wiki: outbound webhooks are described as not built
The-capture-API.md:295-297says outbound webhooks "are a separate feature and not built yet".Webhooks.mddocuments them. Its event table has no deletion event, so the paragraph's point (a mirror must notice absences itself) still stands; only the reason given is wrong.6.
CONTRIBUTING.md: "Scanning the image" says nothing scans the imageCONTRIBUTING.md:518-550. Line 520 says "Nothing scans the image yet — that needs a runner that can build one, and is #156", and gives two manual commands that run unpinned scanner images (aquasec/trivy,anchore/grype) againstpostulo:latest. About a hundred lines earlier,:389and:416say thatscripts/scan-image.shscans both the dev image and the release image before they are pushed, and.forgejo/workflows/image.yml:127and.forgejo/workflows/dev-image.yml:126call it. #156 is closed. The script pinsaquasec/trivy:0.74.0andanchore/grype:v0.100.0(scripts/scan-image.sh:58,:62) and records why the pin matters.A reader who lands on this section first concludes that there is no gate, and anyone who follows its commands runs whatever scanner image is current, which the script's own comments explain is unreliable.
7.
CONTRIBUTING.md: the Translations paragraphCONTRIBUTING.md:283-285says "French and Portuguese catalogues exist and are waiting for contributors".src/postulo/locale/holds 68 catalogues, machine-drafted and flaggeddraftuntil a speaker has read them.8.
ci.yml: the header says onlyactions/checkoutis used.forgejo/workflows/ci.yml:8-11says "Onlyactions/checkoutis used, and nothing else from any action marketplace". The same file usesactions/cache@v4at:78and:396andactions/upload-artifact@v3at:425. This comment is where a reviewer would check which third-party code runs in CI with the workflow's token, and it understates that list.9.
docs/PLAN.md: the extras, and the task queuedocs/PLAN.md:166saysextras: postgres, weasyprint, chromium. WeasyPrint is a default dependency (pyproject.toml:35,"weasyprint>=70"); the extras under[project.optional-dependencies]arepostgres,chromiumandserver.docs/PLAN.md:315-322(assumption 3) says "still nothing enqueues a task" and that "#247 is where that happens, and the queue the Overview counts is structurally empty until it does". #247 is closed:core/errands.py:105callsperform_errand.enqueue(errand.pk), anddocker/compose.yml:88defines aworkerservice behind theworkerprofile.10.
docs/TRANSLATING.md: the phase tabledocs/TRANSLATING.md:19-27still plans "0.4.0 | Asia and South America (#71) | Not started" and "0.5.0 | The rest of the world (#72) | Not started". #71 is closed, absorbed into #269, which orders languages by number of speakers rather than by continent;docs/PLAN.md:38already states that rule. The Africa row puts #70 in 0.3.0, but #70 is open on the 1.0.0 milestone, and #72 is on 0.6.0.docker compose -f docker/compose.yml logs postuloin Troubleshooting../scripts/scan-image.sh <image>.docs/TRANSLATING.md.ci.ymluses in its header, with the reason each one is acceptable (both extra steps are alreadycontinue-on-error, and the upload step says why it stays on v3).extras: postgres, chromium, server, and rewrite assumption 3 to say what is queued now and what theworkerprofile is for, or close the assumption.Done when
tests/test_wiki_surface.pythat no wiki page says "Settings → Appearance" together with "Keyboard shortcuts" or "order number", and that none containsdocker logs postulo. Both fail today.tests/test_wiki_surface.pyor a newtests/test_docs.py, that every action named by auses:line in.forgejo/workflows/ci.ymlis named in the file's header. It fails today foractions/cacheandactions/upload-artifact.CONTRIBUTING.mdhas one scanning section, which namesscripts/scan-image.sh, and its Translations paragraph no longer names two languages.docs/PLAN.mdgives the real extras and the current state of the task queue, anddocs/TRANSLATING.mddescribes the #269 ordering.Fixed in
bb9d0c07con main.