chadtmiller.com/knotless
git that lives in your Atmosphere account
1
10
Release 0.6.0
.changeset/admin-logs.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: minor | ||
--- | ||
| ||
# `knotless admin logs` reads the lines the appview logged | ||
| ||
An appview's operator can read the lines it logged, without SSH: `knotless admin logs --trace <id>` shows one request with the calls it made and how long each took, and `--message`, `--level` and `--min-ms` show the last of the rest. The MCP `logs` tool does the same. The appview keeps them across deploys, for 30 days at most, and logs each read with who asked. | ||
.changeset/branch-caches.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# A feature branch's pulls start from its own warm caches | ||
| ||
A branch whose pushes and pulls both run a workflow, as `on.push.refs` and `on.pull.base` both naming `refs/heads/feature/*` do, keeps that workflow's caches for itself: each passing push to it leaves what the pulls into it start from, and a pull landed there at the tree it was checked at is taken as the push's run, caches and all. A branch with none yet starts from the default branch's, and its caches go when it is deleted. | ||
.changeset/checks-bare-ref.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# `knotless checks --ref` takes a branch named bare | ||
| ||
`knotless checks --ref main` lists the checks of `refs/heads/main`, as `rerun` and `workflows` read a bare name, rather than none; a tag of that name, when there is no such branch. A branch deleted since, as a merged pull's, still names its checks. The MCP `checks` tool's `ref` takes one the same way. | ||
.changeset/logs-trace-whole.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# `knotless admin logs --trace` reads up to 500 lines of the request | ||
| ||
A read by trace now answers up to 500 lines unless `--limit` says fewer, instead of the last 100, so a long request's first lines are not cut off. Lines serve fails to write to its log store, as when another serve holds the file during a deploy, are kept and written with the next batch instead of lost. | ||
.changeset/org-admin-secrets.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# An organization's admins set its repositories' secrets | ||
| ||
`knotless secret <url> set`, `ls` and `rm`, and the MCP `secret` tool, now work on a repository an organization owns when you are logged in as an admin who joined it, not only as the organization's own account. The values are sealed as the organization's, so its runners open them as before. | ||
.changeset/org-admin-settings.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: minor | ||
--- | ||
| ||
# An organization's admins change its repositories' settings on the pages | ||
| ||
An admin who joined an organization now sees the Settings tab on its repositories, and changes everything there as the organization: description, branches and their rules, labels, collaborators, triagers, runners and secrets, and archiving. Transferring and deleting a repository stay with the organization's own login. | ||
.changeset/queue-advice.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# Queueing says to read an ejection's logs before queueing again | ||
| ||
`knotless merge --queue` ends by saying to read an ejected pull's check logs before queueing it again. The MCP `merge` tool's description says the same, says to merge the target branch in just before queueing, and says that a pull conflicting only with one ahead of it waits for that one rather than merging its branch in. | ||
.changeset/queue-record-head.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# A queued pull's record names the head it was queued at | ||
| ||
`knotless merge --queue`, and the MCP `merge` tool with `queue`, write the pull's head into its `dev.knotless.queue` record, as an ejection already does, and `knotless show --json` gives it. The queue's history then says whether a pull queued again after an ejection had changed. | ||
.changeset/queue-same-head.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# Queueing a pull again at the head it was ejected at says so | ||
| ||
`knotless merge --queue` says when the pull is going back into the queue at the very head the queue last ejected, and why it was ejected: unless that was not the pull's doing, it will be ejected again. The MCP `merge` tool answers it as `queued.ejectedBefore`, and `dev.knotless.queuePull` as `ejectedBefore`. | ||
.changeset/queue-stages.mddeleted+0−7
| @@ -1,7 +0,0 @@ | ||
--- | ||
default: patch | ||
--- | ||
| ||
# A queued pull says what it waits for | ||
| ||
`knotless show`, the MCP `show_item` tool and `knotless queue --json` say what a queued pull in no batch is waiting for: `Queued, 2nd of 3, running its checks`, `running its checks again on main as it is now` when one failed on an older tip, or `ready for the next batch`. Before, each just said `Queued`, except while a pre-flight was running. `--json` also gives each one's `stage`, its pre-flights' newest runs (`preflights`), and when the batch ahead of it started (`ahead`). | ||