Restarting pipelines with errors is returning 500

[anbraten]

If I try to restart a pipeline with errors I am getting a 500

[qwerty287]

@anbraten What's the error? It works locally for me

[anbraten]

Try to restart https://ci.woodpecker-ci.org/repos/3780/pipeline/9507

[qwerty287]

Ah you mean failed pipeline... Can reproduce locally 👍

[anbraten]

It prints those in the logs instead of sending it to the UI:

{"level":"error","error":"no secret found for \"QUAY_IO_TOKEN\"","time":"2023-11-05T14:42:54Z","message":"failure to createPipelineItems for woodpecker-ci/woodpecker"}
{"level":"error","error":"Error #01: failure to createPipelineItems for woodpecker-ci/woodpecker\n","ip":"","latency":50.612691,"method":"POST","path":"/api/repos/3780/pipelines/9507","status":500,"time":"2023-11-05T14:42:54Z","user-agent":"","time":"2023-11-05T14:42:54Z"}

[qwerty287]

Hmm so I couldn't reproduce it, it didn't work me for a different reason.

[qwerty287]

So the issue seems that the secret QUAY_IO_TOKEN is not available for PRs (see https://ci.woodpecker-ci.org/orgs/2/settings).

This can be fixed by either allowing this or rewriting our workflows to not load this token on PRs.

[anbraten]

Main issue is actually in case we get an error ie for a missing secret, we do not store the config. So when restarting a throws an error now. Its however kinda hard to decided if we should "always" store the config even in case of an error as it could also be an error unable to parse the config at all