The more time I spend with GitHub Actions, the more convinced I become that workflow files belong in a repo but not in the commit history graph of the default ref; maybe under a different namespace, refs/ci/* or something. There's too much value in keeping them right next to the code, but too much negative value in requiring their changes to be intermingled with codebase commits.
That said I also firmly believe that workflow files should be STRICTLY orchestration; execution logic is app code.
Ben Zanin /
npub12c…dtlg9
2024-05-22 18:33:00
Author Public Key
npub12c2un6cf5j7e75mzxv5cjnyyuc8sek2asyd0h0v8j83fxzke60lqcdtlg9Published at
2024-05-22 18:33:00Event JSON
{
"id": "e6569384d0f9384e18c5ff508afdbb1174c5662c800a48ded880e9c23c95b19b",
"pubkey": "5615c9eb09a4bd9f53623329894c84e60f0cd95d811afbbd8791e2930ad9d3fe",
"created_at": 1716402780,
"kind": 1,
"tags": [
[
"proxy",
"https://mastodon.social/users/gnomon/statuses/112486172632635492",
"activitypub"
]
],
"content": "The more time I spend with GitHub Actions, the more convinced I become that workflow files belong in a repo but not in the commit history graph of the default ref; maybe under a different namespace, refs/ci/* or something. There's too much value in keeping them right next to the code, but too much negative value in requiring their changes to be intermingled with codebase commits.\n\nThat said I also firmly believe that workflow files should be STRICTLY orchestration; execution logic is app code.",
"sig": "7895899c08b56132f81ad275cd21c36f437b63406f4e77799f9258fd93087627183dc97f15701846fe3cd6ecaba84e52771322baef1adfa41cede78728b0da42"
}