Expand the scope of the TestAllDependenciesVendored test to check
that all modules in GOROOT are tidy, that packages are vendored,
the vendor content matches the upstream copy exactly, and that
bundled packages are re-generated (using x/tools/cmd/bundle at
the version selected in cmd module; this is deterministic and
guaranteed to be updated over time).
This is done in a conceptually simple way:
1. Make a temporary copy of the entire GOROOT tree (except .git),
one that is safe to modify.
2. Run a list of high-level commands, the same commands we expect
Go developers should be able to run in a normal complete GOROOT
tree to make it clean and tidy.
3. Diff the end result with the original GOROOT tree being tested
to catch any unexpected differences.
The current set of commands that are run require the cmd/go command,
and a functional compiler itself (because re-generating the syscall
package involves a directive like //go:generate go run [...]). As a
result, copying a large majority of the GOROOT tree is a requirement.
Instead of looking for the few files or directories that can we can
get away not copying (e.g., the testdata directories aren't strictly
needed at this time), we opt not to optimize and just do the simple
copy. This is motivated by these reasons:
• We end up having a complete, normal GOROOT tree, one that happens
to be located at another path. There's a very high likelihood that
module management/code generation commands, both the ones we run
today and any additional ones that we might want to add in the
future, will result in correct results even as the Go project
evolves over time.
• Having a completely stand-alone copy of the GOROOT tree without
symlinks minimizes the risk of some of the module management/code
generation commands, either now or in the future, from modifying
the user's original GOROOT tree, something that should not happen
during test execution. Overlays achieved with symlinks work well
when we can guarantee only new files are added, but that isn't
the case here.
• Copying the entire GOROOT (without .git), takes around 5 seconds
on a fairly modern computer with an SSD. The most we can save is
a couple of seconds.
(We make some minor exceptions: the GOROOT/.git directory isn't copied,
and GOROOT/{bin,pkg} are deemed safe to share and thus symlink instead
of copying. If these optimizations cease to be viable to make, we'll
need to remove them.)
Since this functionality is fairly expensive to execute and requires
network access, it runs only when the test is executed without -short
flag. The previous behavior of the TestAllDependenciesVendored test is
kept in -short test mode. all.bash runs package tests with -short flag,
so its behavior is unchanged. The expectation is that the new test will
run on some of the longtest builders to catch problems. Users can invoke
the test manually 'go test cmd/internal/moddeps' (and it's run as part
of 'go test cmd', again, only when -short flag isn't provided).
On a 2017 MacBook Pro, a successful long test takes under 15 seconds,
which should be within scope of all long tests that are selected by
'go test std cmd'. We may further adjust when and where the test runs
by default based on our experience.
Fixes#36852.
Fixes#41409.
Fixes#43687.
Updates #43440.
Change-Id: I9eb85205fec7ec62e3f867831a0a82e3c767f618
Reviewed-on: https://go-review.googlesource.com/c/go/+/283643
Run-TryBot: Dmitri Shuralyov <dmitshur@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Dmitri Shuralyov <dmitshur@golang.org>
If no files match the embed pattern, the Error field will be set on
the package output by go list. Also set the Incomplete field for
consistency.
Fixes#43727
Change-Id: I5b4bb2a03a751269641a9bc4ef1d0fa0e37d46aa
Reviewed-on: https://go-review.googlesource.com/c/go/+/284257
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
If a package vendored with go mod vendor depends on embedded
files contained in subdirectories, copy them into the the
corresponding place in the module's vendor tree. (Embeds in
parent directories are disallowed by the embed pattern rules, and
embeds in the same directory are copied because go mod vendor
already copies the non-go files in the package's own directory).
Export the vendor pattern expansion code in internal/load so
internal/modcmd's vendor code can use it.
Fixes#43077
Change-Id: I61edb344d73df590574a6498ffb6069e8d72a147
Reviewed-on: https://go-review.googlesource.com/c/go/+/283641
Trust: Michael Matloob <matloob@golang.org>
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
This will produce better errors when earlier versions of
Go compile code using //go:embed. (The import will cause
a compilation error but then the go command will add to
the output that the Go toolchain in use looks too old
and maybe that's the problem.)
This CL also adds a test for disallowing embed of a var inside a func.
It's a bit too difficult to rebase down into that CL.
The build system configuration check is delayed in order to
make it possible to use errorcheck for these tests.
Change-Id: I12ece4ff2d8d53380b63f54866e8f3497657d54c
Reviewed-on: https://go-review.googlesource.com/c/go/+/282718
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
If there are already errors emitted, don't run the Asmb2 pass
and just exit. At the point of Asmb2 relocations are already
resolved and errors should have been reported, if any. Asmb2 is
unlikely to emit additional useful users errors. Instead, the
invalid input may cause inconsistencies and crash the linker, or
it may emit some internal errors which are more confusing than
helpful. Exit on error before Asmb2.
Fixes#43748.
Change-Id: Icf6e27f2eef5b6259e921ec0e64bebad5dd805f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/284576
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
This switches openbsd/amd64 to thread creation via pthreads, rather than doing
direct system calls.
Update #36435
Change-Id: I1105d5c392aa3e4c445d99c8cb80b927712e3529
Reviewed-on: https://go-review.googlesource.com/c/go/+/250180
Trust: Joel Sing <joel@sing.id.au>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Tests should avoid writing to GOROOT when possible. Such writes
would fail if GOROOT is non-writeable, and it can interfere with
other tests that don't expect GOROOT to change during test execution.
Updates #28387.
Change-Id: I7d72614f218df3375540f5c2f9c9f8c11034f602
Reviewed-on: https://go-review.googlesource.com/c/go/+/284293
Trust: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The current implementation requires saying "string" or "[]byte"
and disallows aliases, defined types, and even "[]uint8".
This was not 100% intended and mostly just fell out of when
the checks were being done in the implementation (too early,
before typechecking).
After discussion on #43217 (forked into #43602),
the consensus was to allow all string and byte slice types,
same as we do for string conversions in the language itself.
This CL does that.
It's more code than you'd expect because the decision has
to be delayed until after typechecking.
But it also more closely aligns with the version that's
already on dev.regabi.
Fixes#43602.
Change-Id: Iba919cfadfbd5d7116f2bf47e2512fb1d5c36731
Reviewed-on: https://go-review.googlesource.com/c/go/+/282715
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Allowing embedding into []byte inside a func creates an
unfortunate problem: either all calls start with the same
underlying data and can see each other's changes to the
underlying data (surprising and racy!) or all calls start
by making their own copy of the underlying data
(surprising and expensive!).
After discussion on #43216, the consensus was to remove
support for all vars embedded inside functions.
Fixes#43216.
Change-Id: I01e62b5f0dcd9e8566c6d2286218e97803f54704
Reviewed-on: https://go-review.googlesource.com/c/go/+/282714
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
The docs were never updated for the change to the placement
of the DO NOT EDIT line.
Also, the description of the DO NOT EDIT line interrupted the
description of the //go:generate line, which made for some
confusing references in the text that followed. Move it lower.
Fixes#41196.
Change-Id: I6af2a199fa98d45f5ccac7cdf7e9e54257699e61
Reviewed-on: https://go-review.googlesource.com/c/go/+/283633
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
On Apple Silicon Mac, the C compiler has an annoying default
target selection, depending on the ancestor processes'
architecture. In particular, if the shell or IDE is x86, when
running "go build" even with a native ARM64 Go toolchain, the C
compiler defaults to x86, causing build failures. We pass "-arch"
flag explicitly to avoid this situation.
Fixes#43692.
Fixes#43476.
Updates golang/vscode-go#1087.
Change-Id: I80b6a116a114e11e273c6886e377a1cc969fa3f6
Reviewed-on: https://go-review.googlesource.com/c/go/+/283812
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The cgo header has an unnecessary space in the exported function
definition on non-windows goos.
This was introduced in go1.16 so it would be good to fix it before
release.
Example:
// Current behavior, notice there is an unecessary space
// between extern and void
extern void Foo();
// With this CL
extern void Foo();
Change-Id: Ic2c21f8d806fe35a7be7183dbfe35ac605b6e4f6
Reviewed-on: https://go-review.googlesource.com/c/go/+/283892
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Katie Hockman <katie@golang.org>
For #29062Fixes#43263
Change-Id: I160197c94cc4f936967cc22c82cec01663a14fe6
Reviewed-on: https://go-review.googlesource.com/c/go/+/283873
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Consider the following example,
func test(a, b float64, x uint64) uint64 {
if a < b {
x = 0
}
return x
}
func main() {
fmt.Println(test(1, math.NaN(), 123))
}
The output is 0, but the expectation is 123.
This is because the rewrite rule
(CSEL [cc] (MOVDconst [0]) y flag) => (CSEL0 [arm64Negate(cc)] y flag)
converts
FCMP NaN, 1
CSEL MI, 0, 123, R0 // if 1 < NaN then R0 = 0 else R0 = 123
to
FCMP NaN, 1
CSEL GE, 123, 0, R0 // if 1 >= NaN then R0 = 123 else R0 = 0
But both 1 < NaN and 1 >= NaN are false. So the output is 0, not 123.
The root cause is arm64Negate not handle negation of floating comparison
correctly. According to the ARM manual, the meaning of MI, GE, and PL
are
MI: Less than
GE: Greater than or equal to
PL: Greater than, equal to, or unordered
Because NaN cannot be compared with other numbers, the result of such
comparison is unordered. So when NaN is involved, unlike integer, the
result of !(a < b) is not a >= b, it is a >= b || a is NaN || b is NaN.
This is exactly what PL means. We add NotLessThanF to represent PL. Then
the negation of LessThanF is NotLessThanF rather than GreaterEqualF. The
same reason for the other floating comparison operations.
Fixes#43619
Change-Id: Ia511b0027ad067436bace9fbfd261dbeaae01bcd
Reviewed-on: https://go-review.googlesource.com/c/go/+/283572
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Keith Randall <khr@golang.org>
Renamed setErrorPos to setPos, made it a method of PackageError,
and removed its Package parameter and return value. This makes it
more clear that setPos modifies PackageError and does not create a new
Package.
Change-Id: I26c58d3d456c7c18a5c2598e1e8e158b1e6b4b36
Reviewed-on: https://go-review.googlesource.com/c/go/+/283637
Trust: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
When a command fails due to a module zip sum missing from go.sum,
if the module is in the build list, the go command will print a
'go mod download' command the user can run to fix it.
Previously, a hint was only printed if the module provided a package
in 'all'. We don't print a 'go get' hint, since we may not want to add
a new requirement to go.mod.
Fixes#43572
Change-Id: I88c61b1b42ad56c04e4482f6a1bb97ce758aaeff
Reviewed-on: https://go-review.googlesource.com/c/go/+/282712
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
For #36460
Updates #36465
Change-Id: Id818dce21d39a48cf5fc9c015b30497dce9cd1ef
Reviewed-on: https://go-review.googlesource.com/c/go/+/278596
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
For some reason (that I didn't look into), externally linked
AIX binaries don't have runtime.symtab symbol. Since recent Go
releases (Go 1.3 maybe?), that symbol is empty and not necessary
anyway. Don't require it.
Fixes#40972.
Change-Id: I73a1f0142195ea6debdba8a4f6e12cadc3980dc5
Reviewed-on: https://go-review.googlesource.com/c/go/+/279995
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
Module-related help pages now contain a brief summary and point to the
reference documentation at golang.org/ref/mod for details.
Help pages for commands like 'go get' still describe the basic usage
and summarize flags but don't provide as much background detail.
Fixes#41427Fixes#43419
Change-Id: Icacd38e0f33c352c447cc5a496c99674493abde2
Reviewed-on: https://go-review.googlesource.com/c/go/+/282615
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
* All commands the user can run to fix the problem now appear alone on
a separate line after a tab.
* Removed -d from 'go get' commands.
* Replaced 'go mod tidy' with 'go mod download $modpath' when a
package might be provided by a module missing a sum.
* Errors about 'path@version' syntax are more explicit.
Fixes#29415Fixes#42087Fixes#43430Fixes#43523
Change-Id: I4427c2c4506a727a2c727d652fd2d506bb134d3b
Reviewed-on: https://go-review.googlesource.com/c/go/+/282121
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Previously, commands that wrote go.sum (except 'go mod tidy') would
retain sums for zip files of directly required modules. Sums of
indirect dependencies wouldn't be retained unless they were used to
load packages.
With this change, sums for indirect dependencies will be retained if
they're available. This allows users to add missing sums with
'go mod download example.com/mod', which previously only worked for
directly required modules.
Note that 'go mod download' without arguments now adds sums for every
module in the build list. That matches 1.15 behavior.
For #41103
Change-Id: I4cce2bf1c73578dae836bdb5adb32da071554f1a
Reviewed-on: https://go-review.googlesource.com/c/go/+/282692
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
I have a real 7,000-line Go program (not so big)
that took over two minutes to report a trivial init cycle.
I thought the compiler was in an infinite loop but
it was actually just very slow.
CL 170062 rewrote init cycle reporting but replaced
a linear-time algorithm with an exponential one:
it explores all paths through the call graph of functions
involved in the cycle.
The net effect was that Go 1.12 took 0.25 seconds to load,
typecheck, and then diagnose the cycle in my program,
while Go 1.13 takes 600X longer.
This CL makes the new reporting code run in linear time,
restoring the speed of Go 1.12 but preserving the semantic
fixes from CL 170062.
Change-Id: I7d6dc95676d577d9b96f5953b516a64db93249bf
Reviewed-on: https://go-review.googlesource.com/c/go/+/282314
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This way, if a SIGINT is sent to the go command,
it is forwarded on to the underlying tool.
Otherwise trying to use os.Process.Signal to kill
"go tool compile" only kills the "go tool" not the "compile".
Change-Id: Iac7cd4f06096469f5e76164df813a379c0da3822
Reviewed-on: https://go-review.googlesource.com/c/go/+/282312
Trust: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Current optimization: When we copy a->b and then b->c, we might as well
copy a->c instead of b->c (then b might be dead and go away).
*Except* if a is a volatile location (might be clobbered by a call).
In that case, we really do want to copy a immediately, because there
might be a call before we can do the a->c copy.
User calls can't happen in between, because the rule matches up the
memory states. But calls inserted for memory barriers, particularly
runtime.typedmemmove, can.
(I guess we could introduce a register-calling-convention version
of runtime.typedmemmove, but that seems a bigger change than this one.)
Fixes#43570
Change-Id: Ifa518bb1a6f3a8dd46c352d4fd54ea9713b3eb1a
Reviewed-on: https://go-review.googlesource.com/c/go/+/282492
Trust: Keith Randall <khr@golang.org>
Trust: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
For #41127
Change-Id: I794a082299c6dce4202223197ece1864bed36810
Reviewed-on: https://go-review.googlesource.com/c/go/+/282555
Trust: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
In particular bring in CL 201973, which reverts support for multiple
keys in a struct tag.
For #40281
For #43083
For #43226
Change-Id: I66e76639cbbca55bdbff6956acdb0a97650fdd31
Reviewed-on: https://go-review.googlesource.com/c/go/+/282412
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
For the example in #43551, before late call expansion, the OpArg type is
decomposed to int64. But the late call expansion is currently decompose
it to "x.Key" instead.
This CL make expand_calls decompose further for struct { 1-field type }
and array [1]elem.
This matches the previous rules for early decompose args:
(StructSelect (StructMake1 x)) => x
(ArraySelect (ArrayMake1 x)) => x
Fixes#43551
Change-Id: I2f1ebe18cb81cb967f494331c3d237535d2859e7
Reviewed-on: https://go-review.googlesource.com/c/go/+/282332
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
This change the struct fields for EmbedPatterns and EmbedFiles
to the Package struct listed in the go list documentation that
specifies the fields available to the go list template.
Fixes#43081
Change-Id: I89c325a9d6292a6ce484ee588b172d2f84e2333a
Reviewed-on: https://go-review.googlesource.com/c/go/+/282195
Trust: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Jay Conrod <jayconrod@google.com>
We are trying to avoid by not automatically updating go.mod. The
suggestion should be that users actually add the dependencies they
need, and the command in an easily copy-pastable form now.
Fixes: #43430
Change-Id: I2227dab498fcd8d66184c94ebe9e776629ccadfd
Reviewed-on: https://go-review.googlesource.com/c/go/+/280713
Run-TryBot: Baokun Lee <bk@golangcn.org>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Trust: Jay Conrod <jayconrod@google.com>
Trust: Bryan C. Mills <bcmills@google.com>
modload.Init now sets the default value for -mod if it wasn't set
explicitly. This happens before go.mod is loaded, so
modload.LoadModFile sets the default value again in order to enable
automatic vendoring.
Previously, cfg.BuildMod wasn't set at all if LoadModFile wasn't
called, as is the case for commands that run outside of a module
root. This problem only affected 'go install pkg@version' since other
commands are either forbidden in module mode or run with -mod=mod
(like 'go get' and 'go mod' subcommands).
This change also suppresses "missing sum" errors when -mod=readonly is
enabled and there is no module root.
Fixes#43278
Related #40278
Change-Id: I6071cc42bc5e24d0d7e84556e5bfd8e368e0019d
Reviewed-on: https://go-review.googlesource.com/c/go/+/279490
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
For #41191
Change-Id: I75d327759c3d9ef061c19a80b9b2619038dedf68
Reviewed-on: https://go-review.googlesource.com/c/go/+/281492
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
In a gccgo installation the standard library sources are not available.
Change-Id: I929f3645e3ac95a1fa7047d6a3d243159a86ba66
Reviewed-on: https://go-review.googlesource.com/c/go/+/281493
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
'go get pkg@vers' will now add an explicit requirement for the module
providing pkg if that version was already indirectly required.
'go get mod@vers' will do the same if mod is a module path but not a
package.
Requirements promoted this way will be marked "// indirect" because
'go get' doesn't know whether they're needed to build packages in the
main module. So users should prefer to run 'go get ./pkg' (where ./pkg
is a package in the main module) to promote requirements.
Fixes#43131
Change-Id: Ifbb65b71274b3cc752a7a593d6ddd875f7de23b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/278812
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Jay Conrod <jayconrod@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
By default (and with -mod=readonly), the go command imports an error
if a package provided by an implicitly required module is
imported by a package in the main module. This import requires an
update to go.mod: the module must be required explicitly.
The package loader now provides a hint that 'go get' should be run on
the importing package. This is preferred to 'go get' on the imported
package, since that would add an "// indirect" requirement.
For #43131
Change-Id: I0b353ce8ac8c4ddf1a9863544dfaf6c1964daf42
Reviewed-on: https://go-review.googlesource.com/c/go/+/279528
Trust: Jay Conrod <jayconrod@google.com>
Run-TryBot: Jay Conrod <jayconrod@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
The Go PE linker does not support enough generalized PE logic to
properly handle .rsrc sections gracefully. Instead a few things are
special cased for these. The linker also does not support PE's "grouped
sections" features, in which input objects have several named sections
that are sorted, merged, and renamed in the output file. In the past,
more sophisticated support for resources or for PE features like grouped
sections have not been necessary, as Go's own object formats are pretty
vanilla, and GNU binutils also produces pretty vanilla objects where all
sections are already merged.
However, GNU binutils is lagging with arm support, and here LLVM has
picked up the slack. In particular, LLVM has its own rc/cvtres combo,
which are glued together in mingw LLVM distributions as windres, a
command line compatible tool with binutils' windres, which supports arm
and arm64. But there's a key difference between binutils' windres and
LLVM's windres: the LLVM one uses proper grouped sections.
So, this commit adds grouped sections support for resource sections to
the linker. We don't attempt to plumb generic support for grouped
sections, just as there isn't generic support already for what resources
require. Instead we augment the resource handling logic to deal with
standard two-section resource objects.
We also add a test for this, akin to the current test for more vanilla
binutils resource objects, and make sure that the rsrc tests are always
performed.
Fixes#42866.
Fixes#43182.
Change-Id: I059450021405cdf2ef1c195ddbab3960764ad711
Reviewed-on: https://go-review.googlesource.com/c/go/+/268337
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Go 1.15 pack's r command creates the output file if it does not
exist. The system "ar" command does this as well. Do the same.
For bazelbuild/rules_go#2762.
Change-Id: Icd88396b5c714b735c859a29ab29851e4301f4d2
Reviewed-on: https://go-review.googlesource.com/c/go/+/279516
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Treat the compiler's -linkobj output as "compiler object, which
means "pack c" will "see through" the file and add individual
entry to the new archive, instead of the object as a whole.
This is somewhat peculiar. But Go 1.15's cmd/pack does this,
although seemingly accidental. We just do the same. FWIW, it
does make things more consistent with/without -linkobj flag.
Fixes#43271.
Change-Id: I6b2d99256db7ebf0fa430f85afa7464e334f6bcb
Reviewed-on: https://go-review.googlesource.com/c/go/+/279483
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Previously, reassigned was failing to detect reassignments due to
channel receives in select statements (OSELRECV, OSELRECV2), or due to
standalone 2-value receive assignments (OAS2RECV). This was reported
as a devirtualization panic, but could have caused mis-inlining as
well.
Fixes#43292.
Change-Id: Ic8079c20c0587aeacff9596697fdeba80a697b12
Reviewed-on: https://go-review.googlesource.com/c/go/+/279352
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
When doing external linking on Windows, auto-detect the linker flavor
(bfd vs gold vs lld) and when linking with "lld", avoid the use of
"-T" (linker script), since this option is not supported by lld.
[Note: the Go linker currently employs -T to ensure proper placement
of the .debug_gdb_scripts section, to work around issues in older
versions of binutils; LLD recognizes this section and does place it
properly].
Updates #39326.
Change-Id: I3ea79cdceef2316bf86eccdb60188ac3655264ed
Reviewed-on: https://go-review.googlesource.com/c/go/+/278932
Trust: Than McIntosh <thanm@google.com>
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Mach-O relocation addend is signed 24-bit. When external linking,
if the addend is larger, we cannot put it directly into a Mach-O
relocation. This CL handles large addend by creating "label"
symbols at sym+0x800000, sym+(0x800000*2), etc., and emitting
Mach-O relocations that target the label symbols with a smaller
addend. The label symbols are generated late (similar to what
we do for RISC-V64).
One complexity comes from handling of carrier symbols, which does
not track its size or its inner symbols. But relocations can
target them. We track them in a side table (similar to what we
do for XCOFF, xcoffUpdateOuterSize).
Fixes#42738.
Change-Id: I8c53ab2397f8b88870d26f00e9026285e5ff5584
Reviewed-on: https://go-review.googlesource.com/c/go/+/278332
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
When testing if a flag (e.g. "-no-pie") is supported by the
external linker, pass arch-specific flags (like "-marm").
In particular, on the ARM builder, if CGO_LDFLAGS=-march=armv6
is set, the C toolchain fails to build if -marm is not passed.
# cc -march=armv6 1.c
1.c: In function 'main':
1.c:3:1: sorry, unimplemented: Thumb-1 hard-float VFP ABI
int main() {
^~~
This makes the Go linker think "-no-pie" is not supported when it
actually is.
Passing -marm makes it work.
Fixes#43202.
Change-Id: I4e8b71f08818993cbbcb2494b310c68d812d6b50
Reviewed-on: https://go-review.googlesource.com/c/go/+/278592
Trust: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
The code in the new (introduced in 1.15) Go object file reader was
casting a pointer-mmaped-memory into a large array prior to performing
a read of the relocations section:
return (*[1<<20]Reloc)(unsafe.Pointer(&r.b[off]))[:n:n]
For very large object files, this artificial array isn't large enough
(that is, there are more than 1048576 relocs to read), so update the
code to use a larger artifical array size.
Fixes#41621.
Change-Id: Ic047c8aef4f8a3839f2e7e3594bce652ebd6bd5b
Reviewed-on: https://go-review.googlesource.com/c/go/+/278492
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jeremy Faller <jeremy@golang.org>
Trust: Than McIntosh <thanm@google.com>
Dmitri Shuralyov
Michael Matloob
Jay Conrod
Constantin Konstantinidis
Russ Cox
Cherry Zhang
Joel Sing
Quim Muntal
Ian Lance Taylor
Junchen Li
Bryan C. Mills
Jakub Warczarek
Keith Randall
Cuong Manh Le
Ikko Ashimine
Baokun Lee
markruler
Jason A. Donenfeld
Matthew Dempsky
Than McIntosh