This extracts some of the test code from packages, and adds the ability to run
the same test with multiple drivers.
This should generally be useful for all tools that run on top of go/packages as
well when writing tests for them.
Change-Id: I88c596ad07c0782270c5798d92ae29f7549943cf
Reviewed-on: https://go-review.googlesource.com/c/140118
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Some editors combine stdout and stderr when running tools, so
printing to stderr without failing confuses them. Stop printing go
list's warnings.
They're still useful for debugging, so users can set
GOPACKAGESPRINTGOLISTERRORS to see them. This isn't part of the API and
has no compatibility guarantees.
Change-Id: I9cf091cf59d082123149888468fa0d126dc972c7
Reviewed-on: https://go-review.googlesource.com/c/142997
Run-TryBot: Heschi Kreinick <heschi@google.com>
Reviewed-by: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Can't do module tests on a pre-modules version of Go.
Change-Id: I8e3e22d472f37ebc9f930a68c9a6c0f4b7ba7ea0
Reviewed-on: https://go-review.googlesource.com/c/142700
Run-TryBot: Heschi Kreinick <heschi@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Add an implementation of name= for go list. It will be used to
implement goimports and godoc-like lookups by package name.
Imported a copy of the semver package from the stdlib to do version
comparison, and tweaked the gopathwalk API to include a hint about what
kind of source directory is being traversed.
Note that the tests, despite my best efforts, are not hermetic: go list
insists on doing version lookups in situations where it seems to me like
it shouldn't need to.
I think this implementation is ready for serious use. The one thing I'm
nervous about is that it currently does a substring match when looking
for a package name, so if you look up a package named "a" you will get
a huge number of results. This matches goimports' behavior but I don't
know if it's suitable for general use.
Change-Id: I2b7f823b74571fe30d3bd9c7dfafb4e6a40df5d3
Reviewed-on: https://go-review.googlesource.com/c/138878
Run-TryBot: Heschi Kreinick <heschi@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
Get the value of GOARCH from the config instead.
It should have been set to the same as the one from the environment by
default, but may have been overidden by the caller.
Change-Id: If9a6c0ae998c1c72ad2a68fe83c8bb9f5614a189
Reviewed-on: https://go-review.googlesource.com/c/142361
Run-TryBot: Ian Cottrell <iancottrell@google.com>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
We agreed on a different set of invalid queries than what the test
tests. The change got lost in a merge. Fix it.
Change-Id: I812e561d924f5dbd0c29e3a6ec5fb53022d09487
Reviewed-on: https://go-review.googlesource.com/c/142359
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Add a guarantee for breaking changes.
Change-Id: I7c176eb0c3a309ad187e3a33a645996e397d09be
Reviewed-on: https://go-review.googlesource.com/c/141684
Reviewed-by: Alan Donovan <adonovan@google.com>
This change updates the language accepted through the arguments
to packages.Load to make it more consistent. There are now two layers.
A pattern containing an "=" is considered to be a special query, and the
part of the pattern up to the first "=" is considered the query type. All
other patterns are to be interpreted as the build system interprets it.
For now two special queries exist. file= has the behavior that contains:
did: finding packages containing the given file. The query type pattern=
is used to pass through a pattern to be interpreted by the build system.
pattern= is a type of escaping to allow passing through patterns that
contain an "=" to be interpreted by the underlying buildsystem. To allow
for new query types to be introduced, packages.Load will report an
error if the qury type is not understood. We expect name= to be added in
an upcoming change.
"contains:" changes to "file=". A new
Change-Id: I1b208d1c998c67d5556cdc872d7694273cedb7e4
Reviewed-on: https://go-review.googlesource.com/c/141681
Reviewed-by: Alan Donovan <adonovan@google.com>
The typechecker's Sizes function is not currently set correctly.
The correct answer requires information known only to the build
system's query tool (go list, blaze query, etc), and we need to
add a mechanism for it to return this information.
In the meantime, we use this simple workaround: if the GOARCH
environment variable is set, we use that to determine sizes according
to the conventions of gc. Otherwise, we use the architecture for which
the application was compiled. Both could easily be incorrect, but this
is nonetheless progress.
This change should fix the tests of go/analysis/passes/shift,
which are currently broken for GOARCH=386 because the analysistest
driver uses go/packages, which ignores GOARCH.
Change-Id: Iabe3211ad513a9a94eadd6d8f4b2068f7abdd053
Reviewed-on: https://go-review.googlesource.com/c/141757
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This allows users of go/packages to replace the contents of already
existing files, to support use-cases such as unsaved files in editors.
BREAKING CHANGE: This CL changes the signature of the function provided
to Config.ParseFile.
Change-Id: I6ce50336060832679e9f64f8d201b44651772e0b
Reviewed-on: https://go-review.googlesource.com/c/139798
Reviewed-by: Alan Donovan <adonovan@google.com>
This test is unsurprisingly failing on the nocgo builder because
the cgo packages don't exist on those builders.
Updates golang/go#28040
Change-Id: I633b73bb48e76824645e4e8dd141fb42c9adc19f
Reviewed-on: https://go-review.googlesource.com/c/140121
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Some applications (for example, diagnostics shown to a user in an
editor) may want to distinguish between errors generated by
the driver, parser, and type-checker. The Error struct did not have any
mechanism for doing this, so add an ErrorKind field and set it in
appendError.
Change-Id: If347163225d1e3a567e98610e9ba8a0930e4659c
Reviewed-on: https://go-review.googlesource.com/139317
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
TestLoadImportsC just works on Go 1.10. And there's no good reason
that TestLoadAllSyntaxImportErrors shouldn't work, even though
it's currently always skipped.
Change-Id: Icd8d311f12c5731cc635937a00251eab0a3077ec
Reviewed-on: https://go-review.googlesource.com/139117
Reviewed-by: Alan Donovan <adonovan@google.com>
A previous change (golang.org/cl/137096) that made absolute package paths relative to
GOROOT or GOPATH entries also accidentally applied to .go filesnames.
Filter those out of the list of paths considered to make relative to
search path.
(package paths that don't start with './' or '/' are relative to GOROOT or GOPATH,
but filenames are not.)
Change-Id: I67fbd0e5caa7e53f3ab5b77f55d6841fe2132578
Reviewed-on: https://go-review.googlesource.com/c/138880
Reviewed-by: Alan Donovan <adonovan@google.com>
This applies a part of golang.org/cl/137075 to the fallback. If go list
doesn't find a package, it returns an almost empty package structure
with an error set on it. That change passed through those packages
from the 1.11+ go list, so users could determine there wasn't a match.
This change does the same from the go 1.10 fallback go list code.
Change-Id: I98acc186c0a9eeef0416e9fec0e1fe0e29ddc51c
Reviewed-on: https://go-review.googlesource.com/c/139158
Reviewed-by: Alan Donovan <adonovan@google.com>
Another change to bring the go1.10 fallback's functionality in line with
the go1.11 go list implementation. The fallback now uses go env to
determine GOPATH and GOROOT and searches them to see if any match an
absolute path, and if so trims the GOPATH/GOROOT entry off the start
of the path.
Fixesgolang/go#27734
Change-Id: Ibd2313fc4301d42fd8c0cd98f1f3e7a313d65eb7
Reviewed-on: https://go-review.googlesource.com/137096
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
The documentation for Load says:
“Load returns an error if any of the patterns was invalid as defined
by the underlying build system. It may return an empty list of
packages without an error, for instance for an empty expansion of a
valid wildcard. Errors associated with a particular package are
recorded in the corresponding Package's Errors list, and do not cause
Load to return an error.”
Therefore, it should not be an error for a pattern to match no
packages. If the pattern is a literal package path that does not
exist, we should prefer to return a *Package for it with an error in
the Errors field.
Change-Id: Iaecfb920097e3b520e763bd52c0e326d2e7a4861
Reviewed-on: https://go-review.googlesource.com/137075
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
For all packages p, q where there's an dependency path of
p.test -> ... -> q -> ... -> p, Go creates test variants
q [p.test] of each q and replaces the dependency on q with
a dependency on q [p.test], and replaces p with the expanded
test variant p[p.test]. Fix the fallback logic to add
these missing test variants. (Before this change, it was
only producing the variant of p: p [p.test].)
Fixesgolang/go#27670
Change-Id: Ic56ba35fadcdf8c5928ec76f5a7b0ebe650c9f02
Reviewed-on: https://go-review.googlesource.com/136176
Reviewed-by: Ian Cottrell <iancottrell@google.com>
Part of the testmain generation logic (sans coverage) has been taken
from go build and put into golist_fallback_testmain.go.
golist_fallback invokes this logic and builds up the package metadata
for the testmain.
The tests checking for testmain are now no longer skipped from
packages_test.go.
Change-Id: I487a947f087f3ad4161ea6c2bed06ebb2f833422
Reviewed-on: https://go-review.googlesource.com/134119
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Errors relating to a particular package (as opposed to the Load
operation as a whole) are now only recorded in the Errors
field of the corresponding Package.
Clients are responsible for printing or otherwise handling errors.
This is a breaking API change.
The PrintErrors function prints all accumulated errors,
dependencies first, and is provided for convenience.
PrintErrors is based on Visit, another helper function, which
visits each Package in an import graph.
Also:
- add an Example documenting typical use.
- update clients to handle errors explicitly.
Depends on https://go-review.googlesource.com/c/tools/+/130576
Change-Id: I39407ab7f46dae2f0dd0fdde21080e172e1258aa
Reviewed-on: https://go-review.googlesource.com/131015
Reviewed-by: Michael Matloob <matloob@golang.org>
This change fixes a bug that occured in the golist fallback logic when
an x_test package imported its own package under test. For a package
"a", if "a_test" imported "a", we'd populate "a_test"'s import map
with an entry "a [a.test]" pointing to the test variant of the package
with id "a [a.test]". This change fixes the key to be "a", the correct
import path of the package, not "a [a.test], which is the ID".
Change-Id: If798f2675b01aa537c6ccc129dc35d042d967337
Reviewed-on: https://go-review.googlesource.com/133356
Reviewed-by: Ian Cottrell <iancottrell@google.com>
Rather than document the range of possible error types, requiring
clumsy client code to extract position information, we now expose a
single concrete type for all errors. Position information (in
standardized string form) is technically optional, but we should
strive for 100%, fixing gaps as they arise.
This change enables us to unify the Package and JSON structs in a
follow-up.
Question: should we eliminate the Config.Error hook and be silent by default?
Pro:
+ most clients suppress it.
Con:
- clients that want to print errors (e.g. vet-like tools) would have
to traverse the entire import graph to find them.
- silence is not the most fail-safe behavior.
Change-Id: Ie92b9fb7641ceda429f00928474b650d1dfadedd
Reviewed-on: https://go-review.googlesource.com/130576
Reviewed-by: Michael Matloob <matloob@golang.org>
This makes Load use the value of GOPACKAGESDRIVER in the current
process's environment if the loader's environment wasn't explicitly
set.
Note that if GOPACKAGESDRIVER is not set, we'll still search PATH in
the current process, regardless of whether PATH is set in the loader's
environment. This has not changed.
Change-Id: I3081a6cb0876a21ccd4d87b4b8d082aa47911b75
Reviewed-on: https://go-review.googlesource.com/131016
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Visit packages in deterministic order for predictable JSON.
(I don't know why the test only appears to fail on darwin;
I would expect it to randomly fail in linux 50% of the time.)
Change-Id: I5270804077fc9ca8f529a1f4657e1c35f0586579
Reviewed-on: https://go-review.googlesource.com/130755
Reviewed-by: Michael Matloob <matloob@golang.org>
This change adds the -buildflag argument to go/packages/gopackages,
which may be passed repeatedly to set config.BuildFlags. I felt -flag
was too vague.
config.Flags is renamed to config.BuildFlags for consistency, and
packages.findExternalDriver now passes -buildflag instead of -flag to
drivers. This will break existing drivers.
Change-Id: Iaed58026373a46e137a236ee9a652eb3a9433ee3
Reviewed-on: https://go-review.googlesource.com/130136
Reviewed-by: Alan Donovan <adonovan@google.com>
The legacy implementation currently always loads from source code, so
packages are more complete than expected.
Change-Id: Ib8c9f7ac590038108dba05c1f47d22e70734945c
Reviewed-on: https://go-review.googlesource.com/130095
Reviewed-by: Michael Matloob <matloob@golang.org>
In ModeTypes, we only need the roots, in ModeSyntax we need the direct
dependancies.
We now do exactly what we need rather than the entire set of transitive
dependancies.
Without this change tools operating on source of a deep dependancy tree are far
to slow.
Change-Id: I3ddbe660709184758ad33ab7cdea860244433ce6
Reviewed-on: https://go-review.googlesource.com/129355
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
This saves 35% of wall time in a command such as
$ gopackages -mode=syntax golang.org/x/tools/go/packages/gopackages fmt
that would otherwise typecheck every function in each package that
depends on fmt. It also applies in -mode=types when source was loaded
as a fall back for missing export data due to an error in a lower package.
Also, remove Config.TypeChecker field now that not a single of its
subfields is passed through go/packages to the type checker.
(The Sizes function is logically a result, not an input, of loading,
though we have yet to implement it properly.)
Change-Id: I472d21b34fc5e2832f7353e82992a67a06e4e4cc
Reviewed-on: https://go-review.googlesource.com/129497
Reviewed-by: Ian Cottrell <iancottrell@google.com>
This is fixes for review comments made on cl/128120
Change-Id: Ie6382d6ec91db2bbdece564a83ae028be4e2e269
Reviewed-on: https://go-review.googlesource.com/129000
Run-TryBot: Ian Cottrell <iancottrell@google.com>
Reviewed-by: Alan Donovan <adonovan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
For any identity imports to be added to the import map of a package,
make sure these are not already in there. If the package imports "C",
be sure not to overwrite the import map entries for "runtime/cgo", "unsafe",
and "syscall".
Test the import map when test variants are loaded. This tests the fix
in 'go list' made by https://golang.org/cl/128836Fixesgolang/go#26847
Change-Id: I05dff4d3a75fab03f333f6d88128de6a6bf169e6
Reviewed-on: https://go-review.googlesource.com/128876
Run-TryBot: Suzy Mueller <suzmue@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Copy imports.VendorlessPath into go/packages to avoid an unnecessary
dependency on imports.
Change-Id: Ie77c2fb87f2199f139ece9f3d1b707f065fc1a79
Reviewed-on: https://go-review.googlesource.com/128996
Reviewed-by: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This deletes the raw and golist packatges, instead using the main
Package structure as the serialzied form between the build tool and the
loader.
This requires a few new fields on Package for information that we used
to hide, and some extra json tags on Package.
It also required Package to have custom JSON marshalling methods.
Change-Id: Ib72171bc93e2b494b97f4b0266504acaa9b3f647
Reviewed-on: https://go-review.googlesource.com/128120
Reviewed-by: Michael Matloob <matloob@golang.org>
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This change addresses some style issues pointed out by rsc in
golang.org/cl/126177. It doesn't fix the comment on listfunc
or the comment on the processing of the containing directories.
Those will be handled in a separate cl.
Change-Id: I2094ad1e5e2ee181a646466675f4d993bd69ff9b
Reviewed-on: https://go-review.googlesource.com/128897
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
This change addresses adonovan's comments in golang.org/cl/125939.
Make sure to include CXXFiles, MFiles, HFiles, FFiles, SwigFiles,
SwigCXXFiles, and SysoFiles from the go list output into the Package
struct's OtherFiles field.
When computing packages from the output of fallback golist,
make sure to create the test variant of a package whenever creating
an x_test for a package because the x_test always depends on
the test variant even when there are no non-x_test files.
Sort dependency packages from the fallback golist to ensure
output is deterministic.
Change-Id: I3a942898c7edbe0ad62dbdf3d521775ffd9b9594
Reviewed-on: https://go-review.googlesource.com/128838
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Also:
- make a nil *Config equivalent to new(Config).
- add a test for nil *Config.
- document that Load may return an empty list of packages.
Change-Id: I642133abe18553ca8c7f46b7bd2709a03eda0b28
Reviewed-on: https://go-review.googlesource.com/128875
Reviewed-by: Michael Matloob <matloob@golang.org>
Use export data in LoadTypes mode, when it is available to get the type information.
The lock can be moved back to loadFromExportData, as each package is
loaded in topological order, and all calls to loadFromExportData for a
particular package occur when that package is being loaded.
Fixesgolang/go#26834
Change-Id: Ib6c28eb8642a473cc100d54d0aac7b90644d5d22
Reviewed-on: https://go-review.googlesource.com/128365
Reviewed-by: Michael Matloob <matloob@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Also collect errors so we can set them on the Package (once
golang.org/cl/128120 is in).
Change-Id: I2950405404f060312813e4aa27393496078a3b7e
Reviewed-on: https://go-review.googlesource.com/128357
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
...as it is required not just for Syntax processing but also
when loading export data.
Clarify comments surrounding Types, Fset.
Remove bogus comment re: Dir.
Change-Id: I8d234bc311d73cbed4fd271ef883ebba22ba2708
Reviewed-on: https://go-review.googlesource.com/128595
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
Check for the go-build subdirectory to identify a filepath in the go
build cache. The go command uses a subdirectory of the default os cache
called "go-build" for the build cache.
Additionally, this responds to the comments about function names and tests in the following
CLs:
https://golang.org/cl/125536https://golang.org/cl/125302
Fixes https://github.com/golang/go/issues/26387
Change-Id: I4ec65822e1e178d3de215a773e42f09fcc5bdb0d
Reviewed-on: https://go-review.googlesource.com/128360
Run-TryBot: Suzy Mueller <suzmue@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
Reviewed-by: Michael Matloob <matloob@golang.org>
The fallback was being reset to the incorrect value
after it was set. This change fixes it.
Also add a test that triggers the fallback (a query with both
a contains: line and a regular package path).
Change-Id: I49a9aeb3a0c7d7cc308ac56f4985545315a5bfd2
Reviewed-on: https://go-review.googlesource.com/128356
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
This allows an external binary (not go list) to be the source of package
information.
It uses a binary called gopackagesraw if present in the PATH.
The binary can be overriden by specifying the GOPACKAGESRAW
environment variable.
The command must accept the -test -deps -export and -flags, and take a
list of package patterns to match. It then returns a raw.Results followed by
the matching raw.Package structs in json format on stdout.
Change-Id: I836014d837a284999ded0c1157b8e6dac058ba69
Reviewed-on: https://go-review.googlesource.com/125938
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Matloob <matloob@golang.org>
This change moves the cgo support for go loader to a common
internal package that can be reused by go/packages. That
support is used to invoke cgo to process the files returned
by the fallback pre-go1.11 Go list.
The cgo processing does not propagate all of the config correctly.
We're building a go/build.Package to get the cgo command line
invocation and it might be incorrectly configured if the user's
configuration is different from go/build.Default.
Change-Id: I0de3c65cb723587e1551edeb77219614ba9650bf
Reviewed-on: https://go-review.googlesource.com/127816
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Cottrell <iancottrell@google.com>
Diamonds in the dependency graph could cause the same package
to be processed by two separate gorotines, causing a potential
race between checking if the package's Types field is set, and
setting the field.
adonovan's comment suggested this might be unnecessary. We'll
ask him about this once he returns from his vacation.
Change-Id: I27511b6e44bb1b6a04ea294206fd8ed4ea60f56f
Reviewed-on: https://go-review.googlesource.com/127636
Run-TryBot: Michael Matloob <matloob@golang.org>
Reviewed-by: Ian Cottrell <iancottrell@google.com>
Reviewed-by: Suzy Mueller <suzmue@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ian Cottrell
Heschi Kreinick
Michael Matloob
Alan Donovan
Rebecca Stambler
Bryan C. Mills
Koichi Shiraishi
Todd Neal
Jay Conrod
Suzy Mueller
Kazuhiro Sera