Changeset does not load, until it does

[danieldegroot2]

Describe the bug

Ref #518. Possible bug occured when trying to reproduce this issue (wanted to check what the changeset was, even if the issue might be in old version of OSMCha)

• OSM / achavi shows bounding box, nothing else (okay, see also issue linked at end of this post)

OSMCha (adiffs) fails to load this changeset initially. It returns the following errors

adiffs.osmcha.org/changesets/99045675.adiff:1  Failed to load resource: the server responded with a status of 404 ()
console.js:35 Error: GET /changesets/99045675.adiff returned 404 
    at changeset.js:44:11
    at async changeset.js:34:5
    at async Cn (changeset.js:24:7)
e.<computed> @ console.js:35
/api/v1/changesets/99045675/:1  Failed to load resource: the server responded with a status of 404 ()
/api/v1/changesets/99045675/:1  Failed to load resource: the server responded with a status of 404 ()
console.js:35 Error: Resource not found.
    at aoi.js:27:15

However, eventually (after a minute) it seems to properly load this changeset. (see screenshots below)

Perhaps the website should 'soft fail' initially, showing notice that it will continue the attempt, only showing actual error when retrying / loading for longer does not help. (maybe it is also an additional detail to log if the issue actually persists) Or, perhaps there is (also) a bug with loading this changeset, not sure.

See also https://github.com/OSMCha/osmcha-frontend/issues/804#issuecomment-3103511724 (unrelated; fyi I was working on this comment when encountering this issue.)

Changeset URL

https://osmcha.org/changesets/99045675

Environment

OSMCha 1.2.6 with Chrome 138.0.7204.158 on Windows 10

Additional information

Initially showing error notices

Eventually showing changeset, error notices have been cleared from view.

Sidenote: Neither error notice mentions the user could themselves try reloading the page. This or another error might have done so in the past, see below linked issue. This could be good to know for the user, though of course it is somewhat optional detail.

---

See also #336 (same errors?)