go(1) repsects GOROOT if set, but tool/go / gocross-wrapper.sh are explicitly intending to use our toolchain. We don't need to set GOROOT, just unset it, and then go(1) handles the rest.
Updates tailscale/corp#26717
Signed-off-by: James Tucker <james@tailscale.com>
This tweaks the just-added ./tool/go.{cmd,ps1} port of ./tool/go for
Windows.
Otherwise in Windows Terminal in Powershell, running just ".\tool\go"
picks up go.ps1 before go.cmd, which means execution gets denied
without the cmd script's -ExecutionPolicy Bypass part letting it work.
This makes it work in both cmd.exe and in Powershell.
Updates tailscale/corp#28679
Change-Id: Iaf628a9fd6cb95670633b2dbdb635dfb8afaa006
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
go.cmd lets you run just "./tool/go" on Windows the same as Linux/Darwin.
The batch script (go.md) then just invokes PowerShell which is more
powerful than batch.
I wanted this while debugging Windows CI performance by reproducing slow
tests on my local Windows laptop.
Updates tailscale/corp#28679
Change-Id: I6e520968da3cef3032091c1c4f4237f663cefcab
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We aim to make the tsgo directories be read-only mounts on builders.
But gocross was previously writing within the ~/.cache/tsgo/$HASH/
directories to make the synthetic GOROOT directories.
This moves them to ~/.cache/tsgoroot/$HASH/ instead.
Updates tailscale/corp#28679
Updates tailscale/corp#26717
Change-Id: I0d17730bbdce3d6374e79d49486826575d4690af
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Make the OS-specific staticcheck jobs only test stuff that's specialized
for that OS. Do that using a new ./tool/listpkgs program that's a fancy
'go list' with more filtering flags.
Updates tailscale/corp#28679
Change-Id: I790be2e3a0b42b105bd39f68c4b20e217a26de60
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
gocross is not needed like it used to be, now that Go does
version stamping itself.
We keep it for the xcode and Windows builds for now.
This simplifies things in the build, especially with upcoming build
system updates.
Updates tailscale/corp#28679
Updates tailscale/corp#26717
Change-Id: Ib4bebe6f50f3b9c3d6cd27323fca603e3dfb43cc
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Instead of using the version package (which depends on
tailcfg.CurrentCapabilityVersion) to get the git commit hash, do it
directly using debug.BuildInfo. This way, when changing struct fields in
tailcfg, we can successfully `go generate` it without compiler errors.
Updates #9634
Updates https://github.com/tailscale/corp/issues/26717
Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
Bump to latest 22.x LTS release for node as the 18.x line is going EOL this month.
Updates https://github.com/tailscale/corp/issues/27737
Signed-off-by: Mario Minardi <mario@tailscale.com>
CI / crossmin (amd64, plan9) (push) Waiting to run
CI / go_mod_tidy (push) Waiting to run
CI / crossmin (ppc64, aix) (push) Waiting to run
CI / android (push) Waiting to run
CI / wasm (push) Waiting to run
CI / tailscale_go (push) Waiting to run
CI / fuzz (push) Waiting to run
CI / depaware (push) Waiting to run
CI / licenses (push) Waiting to run
CI / staticcheck (386, windows) (push) Waiting to run
CI / staticcheck (amd64, darwin) (push) Waiting to run
CI / staticcheck (amd64, linux) (push) Waiting to run
CI / staticcheck (amd64, windows) (push) Waiting to run
CI / notify_slack (push) Blocked by required conditions
CI / check_mergeability (push) Blocked by required conditions
trimpath can be inconvenient for IDEs and LSPs that do not always
correctly handle module relative paths, and can also contribute to
caching bugs taking effect. We rarely have a real need for trimpath of
test produced binaries, so avoiding it should be a net win.
Updates #2988
Signed-off-by: James Tucker <james@tailscale.com>
To avoid warning:
find: warning: you have specified the global option -maxdepth after the argument -type, but global options are not positional, i.e., -maxdepth affects tests specified before it as well as those specified after it. Please specify global options before other arguments.
Fixestailscale/corp#23689
Change-Id: I91ee260b295c552c0a029883d5e406733e081478
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
So it doesn't delete and re-pull when switching between branches.
Updates tailscale/corp#17686
Change-Id: Iffb989781db42fcd673c5f03dbd0ce95972ede0f
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Bump node version to latest lts on the 18.x line which is 18.20.4 at the time of writing.
Updates https://github.com/tailscale/corp/issues/21741
Signed-off-by: Mario Minardi <mario@tailscale.com>
Fix regression from #8108 (Mar 2023). Since that change, gocross has
always been rebuilt on each run of ./tool/go (gocross-wrapper.sh),
adding ~100ms. (Well, not totally rebuilt; cmd/go's caching still
ends up working fine.)
The problem was $gocross_path was just "gocross", which isn't in my
path (and "." isn't in my $PATH, as it shouldn't be), so this line was
always evaluating to the empty string:
gotver="$($gocross_path gocross-version 2>/dev/null || echo '')"
The ./gocross is fine because of the earlier `cd "$repo_root"`
Updates tailscale/corp#21262
Updates tailscale/corp#21263
Change-Id: I80d25446097a3bb3423490c164352f0b569add5f
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We don't build a lot of tools with CGO, but we do build some, and it's
extremely valuable for production services in particular to have symbols
included - for perf and so on.
I tested various other builds that could be affected negatively, in
particular macOS/iOS, but those use split-dwarf already as part of their
build path, and Android which does not currently use gocross.
One binary which is normally 120mb only grew to 123mb, so the trade-off
is definitely worthwhile in context.
Updates tailscale/corp#20296
Signed-off-by: James Tucker <james@tailscale.com>
Tracking down the side effect can otherwise be a pain, for example on
Darwin an empty GOOS resulted in CGO being implicitly disabled. The user
intended for `export GOOS=` to act like unset, and while this is a
misunderstanding, the main toolchain would treat it this way.
Fixestailscale/corp#20059
Signed-off-by: James Tucker <james@tailscale.com>
Adds logic in gocross to detect environment variables and pass the right flags so that the backend can be built with the visionOS SDK.
Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
The new 'toolchain' directive in go.mod can sometimes force
the use of an upstream toolchain against our wishes. Concurrently,
some of our dependencies have added the 'toolchain' directive, which
transitively adds it to our own go.mod. Force all uses of gocross to
ignore that directive and stick to our customized toolchain.
Updates #cleanup
Signed-off-by: David Anderson <danderson@tailscale.com>
Run `staticcheck` with `U1000` to find unused code. This cleans up about
a half of it. I'll do the other half separately to keep PRs manageable.
Updates #cleanup
Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
In this commit, we have updated the build process for our Windows DLLs
to link statically with libgcc, ensuring our Windows DLLs are self-contained.
Updates #10617
Signed-off-by: Nick Khyl <nickk@tailscale.com>
This records test coverage for the amd64 no race tests and uploads the
results to coveralls.io.
Updates #cleanup
Signed-off-by: Ox Cart <ox.to.a.cart@gmail.com>
So we can experiment with disabling static linking for tests in CI to
make GitHub Actions output less spammy.
Updates tailscale/corp#13113
Signed-off-by: Brad Fitzpatrick <brad@danga.com>
We're going to need to build a DLL containing custom actions for the installer.
This patch adds the foundations of that capability to dist and gocross.
Updates https://github.com/tailscale/corp/issues/13998
Signed-off-by: Aaron Klotz <aaron@tailscale.com>
This builds the assets for the new web client as part of our release
process. The path to the web client source is specified by the
-web-client-root flag. This allows corp builds to first vendor the
tailscale.com module, and then build the web client assets in the vendor
directory.
The default value for the -web-client-root flag is empty, so no assets
are built by default.
This is an update of the previously reverted 0fb95ec
Updates tailscale/corp#13775
Signed-off-by: Will Norris <will@tailscale.com>
This caused breakages on the build server:
synology/dsm7/x86_64: chdir /home/ubuntu/builds/2023-08-21T21-47-38Z-unstable-main-tagged-devices/0/client/web: no such file or directory
synology/dsm7/i686: chdir /home/ubuntu/builds/2023-08-21T21-47-38Z-unstable-main-tagged-devices/0/client/web: no such file or directory
synology/dsm7/armv8: chdir /home/ubuntu/builds/2023-08-21T21-47-38Z-unstable-main-tagged-devices/0/client/web: no such file or directory
...
Reverting while I investigate.
This reverts commit 0fb95ec07daa81d2a30a44af7d969249cec5bdc8.
Signed-off-by: Will Norris <will@tailscale.com>
This builds the assets for the new web client as part of our release
process. These assets will soon be embedded into the cmd/tailscale
binary, but are not actually done so yet.
Updates tailscale/corp#13775
Signed-off-by: Will Norris <will@tailscale.com>
When running in our github CI environment, curl sometimes hangs while closing
the download from the nodejs.org server and fails with INTERNAL_ERROR. This is
likely caused by CI running behind some kind of load balancer or proxy that
handles HTTP/2 incorrectly in some minor way, so force curl to use HTTP 1.1.
Updates #8988
Signed-off-by: Val <valerie@tailscale.com>
Also switch the wrapper script to use bash not posix shell. We now
depend on bash elsewhere for saner behavior in esoteric areas, so
might as well use it everywhere for consistency.
Fixes#8425
Signed-off-by: David Anderson <danderson@tailscale.com>
This is needed in order to build our network extension on tvOS. First step for #8282
Signed-off-by: Andrea Gottardo <andrea@tailscale.com>
Co-authored-by: Andrea Gottardo <andrea@tailscale.com>
The subshell in which gocross gets built cd's to the corp checkout dir
near the top, so all future references to corp repository files should
be simple relative paths, and not reference $repo_root. When $repo_root
is an absolute path, it doesn't matter and everything works out, but on
some OSes and shells and invocations, $repo_root is a completely relative
path that is invalidated by the "cd".
Fixestailscale/corp#11183
Signed-off-by: David Anderson <danderson@tailscale.com>
This adds an initial and intentionally minimal configuration for
golang-ci, fixes the issues reported, and adds a GitHub Action to check
new pull requests against this linter configuration.
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I8f38fbc315836a19a094d0d3e986758b9313f163
It's used to control various opt-in functionality for the macOS and iOS
apps, and was lost in the migration to gocross.
Updates tailscale/tailscale#7769
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
Recent egrep builds produce a warning:
```
egrep: warning: egrep is obsolescent; using grep -E
```
Updates #cleanup
Signed-off-by: James Tucker <james@tailscale.com>
Previously, the build ended up embedding an empty string, which made
the shell wrapper rebuild gocross on every invocation. This is still
reasonably fast, but fixing the bypass shaves 80% off gocross's overhead
when no rebuild is needed.
Signed-off-by: David Anderson <danderson@tailscale.com>
A bunch of us invoke tool/go from outside the repo that hosts gocross,
as a way of accessing our version-controlled toolchain. This removes
assumptions from gocross that it's being invoked within the repository
that contains its source code and toolchain configuration.
Fixestailscale/corp#9627
Signed-off-by: David Anderson <danderson@tailscale.com>
This used to make sense, but after a refactor somewhere along the line
this results in trying to download from a malformed URL and generally
confusing failures.
Signed-off-by: David Anderson <danderson@tailscale.com>
Xcode changed how/what data it exports to build steps at some point
recently, so our old way of figuring out the minimum support version
for clang stopped working.
Updates tailscale/corp#4095
Signed-off-by: David Anderson <danderson@tailscale.com>
Sometimes, our cached toolchain ends up being an older version of
Go, older than our go.mod allows. In that scenario, gocross-wrapper.sh
would find a usable toolchain, but then fail to compile gocross.
This change makes the wrapper script check that the cached toolchain's
minor version is good enough to build tailscale.com, and re-bootstraps
in shell if not.
Signed-off-by: David Anderson <danderson@tailscale.com>
This makes gocross and its bootstrap script understand an absolute
path in go.toolchain.rev to mean "use the given toolchain directly".
Signed-off-by: David Anderson <danderson@tailscale.com>
