Repository navigation
fix(deps): update dependencies - #24
Open
tannevaled wants to merge 1 commit into
Open
tannevaled wants to merge 1 commit into
tannevaled wants to merge 1 commit into
Conversation
tannevaled
force-pushed
the
renovate/deps
branch
from
November 6, 2026 02:57
f47830d to
1707b94
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 7, 2026 02:28
1707b94 to
f27ffc6
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 8, 2026 01:56
f27ffc6 to
c9bf8d1
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 9, 2026 02:11
c9bf8d1 to
0339aa4
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 10, 2026 04:25
0339aa4 to
f899356
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 11, 2026 02:21
f899356 to
846b268
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 12, 2026 03:29
846b268 to
225a21e
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 13, 2026 03:01
225a21e to
8d3f342
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 14, 2026 02:47
8d3f342 to
6c4d1ba
Compare
tannevaled
force-pushed
the
renovate/deps
branch
5 times, most recently
from
November 19, 2026 03:55
10422ba to
d9ab46e
Compare
tannevaled
force-pushed
the
renovate/deps
branch
from
November 20, 2026 03:47
d9ab46e to
ddf2e8d
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Note
This PR body was truncated due to platform limits.
This PR contains the following updates:
v0.72.0→v0.140.0v0.49.0→v0.56.2Release Notes
go-fleettools/fleettools (github.com/go-fleettools/fleettools)
v0.140.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.139.0...v0.140.0
v0.139.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.138.0...v0.139.0
v0.138.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.137.0...v0.138.0
v0.137.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.136.0...v0.137.0
v0.136.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.135.0...v0.136.0
v0.135.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.134.0...v0.135.0
v0.134.0Compare Source
Two commits since v0.133.0, and they are the two halves of one afternoon: a question three sessions could not answer, and the reason none of them noticed the answer was already half-written.
New:
cmd/testtime(#345)Which tests account for a package's wall clock, from
go test -json:It refuses the two arithmetic errors that make such a report lie. A parent's
Elapsedalready contains its subtests, so only top-level names are totalled and--subtestsanswers the other question in its own column. And parallel tests overlap, so when their sum exceeds the package's own elapsed it prints PARALLEL instead of a share — a percentage of a clock several tests occupied at once is a fraction of nothing. It reports and does not judge: no threshold, no "too slow". Exit 2 on input holding no test events, so a report over nothing cannot look like a clean one.⚠ The gap it fills is narrower than the claim that motivated it. That commit's message says the fleet's existing tools read no elapsed times;
cmd/testwhydoes, per PACKAGE, and uses-min-secondsas a witness that a package which finished implausibly fast ran nothing. What nobody had is the breakdown within a package, which is whattesttimeadds. The correction is intesttime's own doc comment and the README row; the commit message is immutable and carries the earlier wording.cmd/fleetverbs: a zero from--matchsays what it did not search (#346)--matchreads names and documentation. Printed alone, its zero reads as "nothing here does X" — and that is exactly how it was read on 2026-11-17, three times by one session and nearly a fourth by another:…while
cmd/testwhywas in that very tree, reading elapsed times.--textfinds it;--matchcannot, because a capability reached through a table or a doc line that never spells the word has no symbol named for it. The zero now carries its own surface and names the flag that widens it.This is the rule the command already followed for what it could not read — every unparseable file, skipped fixture and missing path is counted and named — extended to the case where everything was read and the search was narrow.
priorarthas always answered this way about its own surfaces ("2 surface(s) could not be read; that is not an absence of prior art").⭐ What exposed it was a positive control, not care:
--match ledger, a word certainly incmd/recur, also answered 0. Three repetitions of a narrow search are one measurement.The shape worth keeping from both
A tool whose search surface is narrower than its subject reports the gap as cleanliness.
stalerefgot the same treatment earlier the same day — it now prints a census of the extensions it walked past, after reporting 1 finding over "237 doc file(s)" where--ext hclfound 72. Three tools, one defect class, and in every case the number was right and the missing thing was the sentence saying what it covered.v0.133.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.132.0...v0.133.0
v0.132.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.131.0...v0.132.0
v0.131.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.130.0...v0.131.0
v0.130.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.129.0...v0.130.0
v0.129.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.128.0...v0.129.0
v0.128.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.127.1...v0.128.0
v0.127.1Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.127.0...v0.127.1
v0.127.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.126.0...v0.127.0
v0.126.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.125.0...v0.126.0
v0.125.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.124.1...v0.125.0
v0.124.1Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.124.0...v0.124.1
v0.124.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.123.0...v0.124.0
v0.123.0Compare Source
modtidynames its two non-verdicts instead of guessing them (#301, closing#294 and #300). Both were measured on 2026-11-15, and both were places where
the gate answered where it could not know.
A module with no Go source is
skipped, with the reason named.go mod tidycan only REMOVE requirements there, so its verdict carries noinformation — and the remedy it prescribed destroyed the repository: on a clean
go-fft/docsclone the hook returneddenywith text telling the next sessionto run
go mod tidyand commitgo.modandgo.sum, which deletes the Hugotheme from both. The signal is "no
.gofile of its own", decided beforegoruns, so it needs neither toolchain nor network. A Hugo configuration is
reported, never what decides: that site has no
[module]block, so a rulekeyed on one would have missed its own occurrence. A nested module's sources
never answer for its parent, and a file behind a build constraint still counts,
so a
//go:build toolsfile keeps a module judged.A
godirective above the installed toolchain is retried once with thetoolchain the directive names.
GOTOOLCHAIN=localon a host whose Go isolder cannot build such a module at all, and in the hook UNREAD means allow —
so the gate had gone silently off for every repository raised to
go 1.27.4,this one included:
modtidycould not readgo-fleettools/fleettools. Theretry reproduces CI rather than loosening the check, since CI's runner has the
version the directive requires. Not
GOTOOLCHAIN=auto, which would leave theversion unstated. The report names the toolchain that answered.
A broken
go.modis still UNREAD, never skipped: skipping is a clean answer,and "no
.gofile" does not read the file that is wrong. The repository's owncorpus caught that during review of the change.
Breaking: nothing for callers. The report gains a fourth class and a fourth
figure in its summary line, so anything parsing that line sees one more count.
v0.122.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.121.0...v0.122.0
v0.121.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.120.0...v0.121.0
v0.120.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.119.0...v0.120.0
v0.119.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.118.0...v0.119.0
v0.118.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.117.0...v0.118.0
v0.117.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.116.1...v0.117.0
v0.116.1Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.116.0...v0.116.1
v0.116.0Compare Source
What's Changed
uses:job, so a shared lane is not read as no lane by @tannevaled in #290Full Changelog: go-fleettools/fleettools@v0.115.0...v0.116.0
v0.115.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.114.0...v0.115.0
v0.114.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.113.0...v0.114.0
v0.113.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.112.0...v0.113.0
v0.112.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.111.0...v0.112.0
v0.111.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.110.0...v0.111.0
v0.110.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.109.0...v0.110.0
v0.109.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.108.0...v0.109.0
v0.108.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.107.0...v0.108.0
v0.107.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.106.0...v0.107.0
v0.106.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.105.0...v0.106.0
v0.105.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.104.0...v0.105.0
v0.104.0Compare Source
runson— which runner images the fleet's workflows ask forA new command, and the last fleet tool that was living outside a repository (fleettools#201 item 2).
Zero API calls: asking GitHub would be ~20 requests across ~1400 repositories, several times the hourly budget the Renovate runners share. It reads the checkouts already on disk, counting repositories rather than files — one repository with nine lanes is not nine exposures.
First pass: 1317 checkouts, 1247 with workflows, 3599 findings.
ubuntu-latest${{ matrix.os }}macos-latestwindows-latestubuntu-24.04-armubuntu-26.04-armWritten for the upstream migration of
ubuntu-latest(x64) to Ubuntu 26.04 between 2026-10-19 and 2026-11-19: inside that window a red CI can be the image moving while the repository did not.The staleness is measured, not disclaimed
Reading a disk costs you currency of evidence, and this says how much rather than warning you vaguely:
842 + 405 = 1247: not one checkout is confirmed current. The direction is stated too — a checkout predating a newly added workflow makes the count a floor, and an absence read off a stale disk looks exactly like an absence that is real.
Refusals
ubuntu-latest" and "nothing was read" are different answers.${{ matrix.os }}and[self-hosted, linux]are reported verbatim rather than resolved, because a wrong resolution is worse than an honest indirection.Full notes in #239.
v0.103.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.102.0...v0.103.0
v0.102.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.101.0...v0.102.0
v0.101.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.100.0...v0.101.0
v0.100.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.99.0...v0.100.0
v0.99.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.98.0...v0.99.0
v0.98.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.97.0...v0.98.0
v0.97.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.96.0...v0.97.0
v0.96.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.95.0...v0.96.0
v0.95.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.94.0...v0.95.0
v0.94.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.93.0...v0.94.0
v0.93.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.92.0...v0.93.0
v0.92.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.91.0...v0.92.0
v0.91.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.90.0...v0.91.0
v0.90.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.89.0...v0.90.0
v0.89.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.88.0...v0.89.0
v0.88.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.87.0...v0.88.0
v0.87.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.86.0...v0.87.0
v0.86.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.85.0...v0.86.0
v0.85.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.84.1...v0.85.0
v0.84.1Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.84.0...v0.84.1
v0.84.0Compare Source
What's Changed
Full Changelog: go-fleettools/fleettools@v0.83.0...v0.84.0
v0.83.0Compare Source
--verify-origin: the report described one diskpinscanwalksgo.modfiles in the working tree. About 25 sessions holdclones of very different ages on this machine, so a six-day-old clone writes
the same bytes as a stale pin — and nothing in the output said which tree was
read.
Re-reading each declaring repository's
go.modfrom GitHub:53 of 64 were already fixed upstream. The eleven real ones are
go-pdfkit/pdfkit,go-ruby-fsctl/fsctl, `go-slidConfiguration
📅 Schedule: (UTC)
🚦 Automerge: Enabled.
♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.
👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.
This PR has been generated by Mend Renovate CLI.