- Compare the installed algod version before and after the upgrade and say "upgraded from X to Y", "already up to date" or "finished", instead of always claiming success.
- Read apt-get update output: fail on errors from the Algorand repository and on fatal apt errors (locks, malformed sources), warn on everything else. apt exits 0 on signature failures when lists are cached, so the exit code alone hides them.
- Before updating, compare the installed Algorand apt key with the published one. If it has expired, rotated or been revoked, show the installed key and the merged result and ask y/N; --accept-new-key answers yes for unattended runs. Only keyrings that hold just the pinned Algorand primary key, in the standard keyring directories, are touched.
- Fail with a clear message when no enabled apt repository provides the algorand package and nothing was upgraded.
- On macOS read the installed version from brew info --json=v2.

v1.7

fix: give --lines one meaning with and without --follow

Scan took the live log's size and offset by path, and Follow reopened
it by path and caught a swap only when the new file was shorter than
the offset. A rotation while the history was replaying lost what was
written to the old file in the meantime. If the new file had already
grown past the offset, its head was lost too, with no rotation notice.
Now that -f replays the whole history by default, that window lasts as
long as the replay.

Scan now opens the live log once, reads it only through that handle,
and returns the handle for Follow to carry on from. Follow reads the
old file to its end before it moves to the new one. An archive path
the rotation has pointed at the live log is skipped rather than read
twice.

--since 0s was described as a stream that starts at now, but Keep never
drops a line without a timestamp, and --since bounds the region read
rather than each line in it. Crash output near the end of the log is
still replayed. The help text, man page and decision record now say
that it skips the timestamped history, and a test pins the plain-text
case.

feat: identify nodekit via user-agent

fix: re-exec updated NodeKit before upgrading Algorand

Both genesis downloads reached for a package-level client, so a test had
no way to point them at a mock server. Take an api.HttpPkgInterface like
the release checks, the catchpoint lookup and the short links already do.
GetGenesis has no callers; handleDataDirMac gets the client from
algod.Install, which the install and bootstrap commands now supply.

Drop main_test.go with it. A test binary is not linked as package main,
so the linker never stamps main.version there and both sides of the
comparison were the same "dev" default: the test passed whether or not
init handed the version over. api/useragent_test.go covers SetVersion.

Every outbound request went out as Go's default "Go-http-client/1.1",
which tells a node operator reading their algod access log, or the
GitHub release API, nothing about what is calling them. Send
"Nodekit v<version>" instead, built from the version the linker stamps
into main so that releases can be told apart.

The shared HttpPkg wrapper covers the release checks, the catchpoint
lookup, the short links and the upgrade download; the generated algod
client gets the header through a request editor, since api/lf.go is
overwritten by `make generate`.

--follow substituted a backlog of 10 when --lines was not given, so -n
meant one thing alone and another beside -f: `-n 0 -f` replayed the
whole history, where `tail -n 0 -f` shows none of it. --follow now
replays the same set the command shows without it, --lines N shortens
that backlog, and a stream that starts at now is --since 0s.

The cost is that a bare -f reads the whole history before the first
new entry arrives, as the command without -f already does.

--follow substituted a backlog of 10 when --lines was not given, so -n
meant one thing alone and another beside -f: `-n 0 -f` replayed the
whole history, where `tail -n 0 -f` shows none of it. --follow now
replays the same set the command shows without it, --lines N shortens
that backlog, and a stream that starts at now is --since 0s.

The cost is that a bare -f reads the whole history before the first
new entry arrives, as the command without -f already does.

Both genesis downloads reached for a package-level client, so a test had
no way to point them at a mock server. Take an api.HttpPkgInterface like
the release checks, the catchpoint lookup and the short links already do.
GetGenesis has no callers; handleDataDirMac gets the client from
algod.Install, which the install and bootstrap commands now supply.

Drop main_test.go with it. A test binary is not linked as package main,
so the linker never stamps main.version there and both sides of the
comparison were the same "dev" default: the test passed whether or not
init handed the version over. api/useragent_test.go covers SetVersion.

Every outbound request went out as Go's default "Go-http-client/1.1",
which tells a node operator reading their algod access log, or the
GitHub release API, nothing about what is calling them. Send
"Nodekit v<version>" instead, built from the version the linker stamps
into main so that releases can be told apart.

The shared HttpPkg wrapper covers the release checks, the catchpoint
lookup, the short links and the upgrade download; the generated algod
client gets the header through a request editor, since api/lf.go is
overwritten by `make generate`.