Skip to content

fix: Bing Site Scan findings (report page, slide template, DSOC project pages, export link) - #101

Merged
ZeeshanAdilButt merged 4 commits into
mainfrom
fix/bing-site-scan
Nov 13, 2026
Merged

ZeeshanAdilButt merged 4 commits into
mainfrom
fix/bing-site-scan

Conversation

@ZeeshanAdilButt

Copy link
Copy Markdown
Member

Why

Bing's Site Scan of www.devweekends.com (37 pages) reported 3 errors, 6 warnings and 3 notices. The image alt findings were already fixed by earlier PRs. This fixes the four that were still true on the live site.

Findings and fixes

1. /reports/fellowship-2025-impact-report.html: no meta description. It is a public report, so it gets a description built from its own numbers, a canonical address and share tags.

2. /brand/session-template.html: no description, four h1 tags. It is a slide template speakers download, not a page for search. It now sends X-Robots-Tag: noindex. The file itself is unchanged.

3. /dsoc/projects/pathment and /dsoc/projects/voiceybill: no h1. Project pages were rendered entirely in the browser, so a crawler got an empty shell.

  • Project pages are now rendered on the server: the first HTML holds the title, description and the rest of the project. Each has its own title, description and canonical address. They are cached for an hour.
  • The two addresses Bing flagged are built-in samples that stand in when the database is unreachable, and they repeat real projects. They now have their h1 too, but are marked noindex, so search shows only the real project pages.
  • The real project pages (from the database) are added to the sitemap, which is rebuilt hourly.
  • If the database cannot be read on the server, the page loads the project in the browser as before, and is not marked noindex.
  • Dates on project pages are shown in UTC, the way they are stored. Before, a visitor west of UTC saw the previous day.
  • Mentor email addresses are not included in the server-rendered page. The page never showed them.

4. /api/feed/export: blocked by robots.txt. The block is deliberate. The link on /feed now has rel="nofollow" so crawlers are not sent to an address they may not fetch.

Checked on a local production build

  • The crawler view (no JavaScript) of every page above: h1, title, description, canonical, robots.
  • The two sample project pages in headless Chrome in two time zones: no hydration or console errors, same dates in both.
  • The database is not reachable from the build machine, so the database path is checked on the live site after deploy.

… search

Bing's Site Scan flagged both files for a missing meta description.

The 2025 impact report is a public page, so it gets a description built from
its own numbers, a canonical address and share tags.

The session slide template is a file speakers download, with four h1 tags by
design. It now answers with X-Robots-Tag: noindex instead of being polished
for search.
The export address is under /api/, which robots.txt blocks on purpose, so
Bing reported it as blocked. rel=nofollow tells crawlers not to go there.
A project page was rendered entirely in the browser, so crawlers got an empty
shell with no h1 (reported by Bing's Site Scan) and every project shared the
list page's title and canonical address.

The page now reads the project on the server and passes it to the same view,
so the first HTML holds the project. Each project has its own title,
description and canonical address, and is cached for an hour.

- The built-in samples repeat real projects and stand in when the database is
  unreachable, so they are marked noindex.
- If the database cannot be read, the browser loads the project as before and
  the page is not marked noindex.
- Dates are shown in UTC, as stored, so the server and the browser print the
  same day. Visitors west of UTC used to see the previous day.
- Mentor email addresses are not put in the page; it never showed them.
Now that project pages can be read by search engines, the active projects from
the database are listed with the date each last changed. The sitemap is rebuilt
hourly, so a new project appears without a deploy.
@ZeeshanAdilButt
ZeeshanAdilButt merged commit 2ccd985 into main Nov 13, 2026
2 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