fix(schemas): allow spec URLs without fragment

[caugner]

Summary

Fixes the spec_url_value pattern in the compat data schema to allow spec URLs without a fragment.

Test results and supporting details

See failing check: https://github.com/mdn/browser-compat-data/actions/runs/18875856784/job/53865336337?pr=28219#step:5:33

JSON Schema - 3 problems (3 errors, 0 warnings):
 ✖ http/headers/Idempotency-Key.json - Error → PATTERN must match pattern "(^https://[^#]+#.+)|(^https://github.com/WebAssembly/.+)|(^[https://registry.khronos.org/webgl/extensions/[^/]+/)](https://registry.khronos.org/webgl/extensions/[%5E/]+/))"

  4 |       "Idempotency-Key": {
  5 |         "__compat": {
> 6 |           "spec_url": "https://datatracker.ietf.org/doc/html/draft-ietf-httpapi-idempotency-key-header/",
    |                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 👈🏽  pattern must match pattern "(^https://[^#]+#.+)|(^https://github.com/WebAssembly/.+)|(^[https://registry.khronos.org/webgl/extensions/[^/]+/)](https://registry.khronos.org/webgl/extensions/[%5E/]+/))"
  7 |           "support": {
  8 |             "chrome": {
  9 |               "version_added": false
 ✖ http/headers/Idempotency-Key.json - Error → TYPE must be array

Related issues

Blocks #28219.

[github-actions]

Tip: Review these changes grouped by change (recommended for most PRs), or grouped by feature (for large PRs).

[Elchi3]

The idea behind forcing us to add fragments is that we always deep link to definitions. This has worked quite well in the past I think. Hesitant to move away from this practice.

The deep link for the HTTP header in question would be https://datatracker.ietf.org/doc/html/draft-ietf-httpapi-idempotency-key-header/#name-the-idempotency-key-http-re