update_scholar_citations.py times out

[niebles]

Have you checked that your issue isn't already filed?

• I read through FAQ and searched through the past issues, none of which addressed my issue.
• Yes, I have checked that this issue isn't already filed.

Bug description

update_scholar_citations.py times out frequently

How to reproduce the bug

See

https://github.com/alshedivat/al-folio/actions/runs/26010226599
https://github.com/alshedivat/al-folio/actions/runs/25773194924

It looks like this script times out frequently. The log shows

❌ Citation update script failed with exit code 124.

which indicates the script times out after the 90s it is given.

The job log also shows:

🚀 Running citation update script (single attempt)...
❌ Citation update script failed with exit code 124 (started at Wed May 13 01:49:50 UTC 2026, ended at Wed May 13 01:51:20 UTC 2026).

Error messages and logs

See https://github.com/alshedivat/al-folio/actions/runs/25773194924
and https://github.com/alshedivat/al-folio/actions/runs/25773194924/job/75700408556

What operating system are you using?

Linux

Where are you seeing the problem on?

Deployed site

More info

No response

[niebles]

@dfuchss fyi

[dfuchss]

Hi @niebles , the issue here is that Google Scholar is blocking requests from the GitHub runners very frequently; if it is not blocked, the update takes around 20s ; see https://github.com/alshedivat/al-folio/actions/runs/25897366617/job/76113180335

The timeout is there to do not waste time , because it's unlikely that even waiting for 15 minutes or more would resolve this , because the IP is blocked.

So unfortunately I don't think that this is an issue that we can resolve ; For my instance of alfolio I simply increased the amount of tries to be one daily.

[niebles]

Got it.
Do you know why we are not seeing some of the messages such as
Fetching citations for Google Scholar ID
which happen before scholarly is invoked?

[dfuchss]

I don't know how the timeout command is handling stdout etc . That needs to be checked I guess :)

[dfuchss]

Probably python -u works better (disabling the buffers for stdX)

[dfuchss]

Python -u does the trick; then you see the output.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[alshedivat]

Closing — already labelled wontfix, and the root cause is confirmed to be outside this repo's control.

@dfuchss established in-thread that Google Scholar rate-blocks GitHub-hosted runner IPs. A longer timeout can't help a blocked IP; the request never completes regardless of how long it's given. Running the script locally, or from a runner on an unblocked IP, is the workaround.

One genuinely fixable thing was buried in here, so I've spun it out rather than let it close with this thread: the workflow runs timeout 90 python … without -u, so Python's buffered stdout is discarded when timeout kills it and the log comes back empty. Your finding that -u restores the output is now tracked as #3686.

Thanks for doing the diagnosis — the negative result is what made it clear this isn't fixable here.

---

Generated by Claude Code