fuzz: store the corpus

[tommady]

fulfill #5312

[tertsdiepraam]

Small tip: if you use one of the keywords that GitHub recognizes (i.e fixes or closes instead of fulfill) it will automatically link the issue to the PR: https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/using-keywords-in-issues-and-pull-requests

[tommady]

Note: You can only download artifacts in a workflow that were uploaded during the same workflow run.

hi @sylvestre
I am sorry the github artifacts way is not working.

specifically, the download can only be used in the same workflow as the doc said,
so maybe we should consider creating a new repo to store the corpus.

so the new plan:

1. a new repo for storing the corpus
2. github action upload or commit corpus in there
3. github action download or clone corpus from there

or I am sure the artifact has been uploaded into this repo,
but I don't have the permission to see it,
if I can we could consider using the github cli instead for the downloading.

WDYT?

[sylvestre]

i think you might want to try the cache storage? no?

[sylvestre]

new repo sounds great too :)

[tommady]

i think you might want to try the cache storage? no?

That sounds great, let me do some study and try on this
https:/actions/cache

thank you

[tommady]

Hi @sylvestre
the cache action working!

with and without cache the time compression

https:/uutils/coreutils/actions/runs/6468556171/job/17560815053
9m 36s
vs
https:/uutils/coreutils/actions/runs/6491029799/job/17627703803?pr=5363
8m 51s

[sylvestre]

and it seems that it works:
2023-10-12T03:56:14.2306567Z INFO: 854 files found in /home/runner/work/coreutils/coreutils/fuzz/corpus/fuzz_date

[sylvestre]

Bravo :)

[sylvestre]

Many thanks

[sylvestre]

and it works!

INFO: 93 files found in /home/runner/work/coreutils/coreutils/fuzz/corpus/fuzz_expr
well done