Several fixes related to GOPRIVATE handling and links:
+ In Go source, fix links matching GOPRIVATE for external modules.
Previously, in these cases we'd try to match <mod>@v1.2.3/<suffix>,
which wasn't the correct input into the GOPRIVATE matching algorithm.
+ Similarly check GOPRIVATE for go.mod require statement hovers.
+ Likewise, for documentLink requests (both mod and source).
+ Move the existing hover regtest to link_test.go, and expand to cover
all these cases.
Along the way, I encountered a couple apparent bugs, which I fixed:
+ Correctly handle the case where there is only one require in a go.mod
file. This was exercised by the regtest, so took some debugging.
+ Only format links [like](this) if the requested format is actually
markdown.
Fixesgolang/go#36998
Change-Id: I92011821f646f2a7449dcca619483f83bdeb54b0
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238029
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
To prevent misleading errors from outstanding go command invocations at
test completion, properly shutdown the LSP connection before cleaning up
exported files.
Change-Id: I9ad175060fefc5b914e544c5f58b9b6658405edc
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238546
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
While working in this code, I was confused by the fact that modHandle
and modData were used for both ModHandle and ModTidyHandle. Separate out
the two so it's a little more clear. This is mostly a copy-paste, with a
few fields renamed to simplify them.
Change-Id: I20ea11c32a624fd250eabf901d09ebb05fbad062
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238337
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
Refactor internal/lsp/cache to use a new temporary go.mod file for each
go command invocation. This cleans up the abstraction in the source
package, as we no longer are aware of temporary go.mod files.
This will also fix the raciness of reusing the same temporary go.mod
file for each invocation.
Updates golang/go#37318.
Fixesgolang/go#39504.
Change-Id: I90bc17a678b5df222ab598c8f7dbf6c6fdd393f6
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237517
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
Revert https://golang.org/cl/234480, which was unnecessary, and move it
to a more appropriate file.
Change-Id: I3f5a3eccaf0ffe324fee8e27945a2e5ece2ff12c
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238597
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
When a workspace package is deleted, we need to stop trying to load it.
Check that every workspace package still has at least one .go file when
we copy them between snapshots.
I think this is correct, but even if it's not, orphaned file loading
should patch it up.
Fixesgolang/go#38977.
Change-Id: I0b11010a40aac09f619f54b5ba02e2467b15a36c
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238028
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Bad go/packages results were being cached, causing unexpected behavior
when new x tests are created. This fix also allows us to enable a
regression test that had been previously disabled.
Change-Id: I93c05df2f4d32c6c8a43e9f5aaeeb30bc4a32f3a
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238058
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
In GOPATH mode, you can import example.com/foo even if it has a go.mod
that says its path is example.com/foo/v2. Adjust our import path
resolution to support that case.
Fixesgolang/go#39560.
Change-Id: I758a2220c579c4374084365c3a78a3a2bbd14b01
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238260
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
In a case where the type info for an ast.CompositeLit isn't found in
forEachLexicalRef(...) gopls will panic.
e.g. trying to rename "foo" where it is declared in this case:
func fn() {
var foo bool
make(map[string]bool
if foo {
}
}
Note the missing ')' after make.
This change will skip ast.CompositeLits if the type can't be found, and
that fixes the panic. But it also do rename the identifier as long as it
is possible, and I'm not convinced that we should allow rename at all if
the source can't be compiled.
Updates golang/go#39614
Change-Id: Ibb50b15ce4b31056f2f1da52a4dcab7b8b91a320
Reviewed-on: https://go-review.googlesource.com/c/tools/+/238042
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
Add a regtest to verify that GOPRIVATE identifiers are not given a link
to pkg.go.dev. For efficiency, as well as to exercise dynamic
configuration, do all this in a single regtest.
Updates golang/go#36998
Change-Id: I9102a11312db5c334fdbd30cce9ca2d2e19e9ac2
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237938
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
The gopls workspace environment defaults to the process environment in
which gopls was started. This means that when switching environments,
gopls can potentially get a different environment when connecting as an
editor sidecar from when forwarding requests to the daemon.
To (hopefully mostly) mitigate this pain point, inject the Go
environment when forwarding the 'initialize' request, which contains
InitializationOptions containing the 'env' configuration. We could go
further and send the entire os.Environ(), but that seems problematic
both in its unbounded nature, and because in many cases the user may not
actually want to send their process env over the wire. Gopls behavior
should *mostly* be parameterized by gopls binary and Go env, and after
this change these should match for forwarder and daemon.
For go1.15, Explicitly set the GOMODCACHE environment variable in the
regtest sandbox. Without this, regtests were failing in the forwarded
environment because they implicitly shared a module cache.
Fixesgolang/go#37830
Change-Id: Ic1b335506f8b481505eac9f74c0df6293dc07158
Reviewed-on: https://go-review.googlesource.com/c/tools/+/234109
Run-TryBot: Robert Findley <rfindley@google.com>
Reviewed-by: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Limiting the Config to the view seems reasonable, considering that it is
only used to run the `go` command. I prefer just having the cache run
go commands, so that source doesn't have to deal with the environment.
This also enables CL 237517.
Change-Id: I639c082592de30e9682dc25cdd12c7751ddb4f97
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237600
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
We should always go through the runner to avoid load serialization
errors. This change also updates the comments and uses pointers to
Invocations consistently. This causes a few updates throughout x/tools.
This change also results in a few modifications to the go:generate code
lenses, as they operate on directories rather than file URIs.
Change-Id: I6306e761e68dfdd23f1b410e44aab0ffa85d234c
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237685
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
Now that fillstruct is an analyzer, we can simplify the code that calls
it in code_action.go. We introduce a new class of analyzer --
convenience analyzers, which are closer to commands. These represent
suggestions that won't necessarily improve the quality or correctness of
your code, but they offer small helper functions for the user.
This CL also combines the refactor rewrite tests with the suggested fix
tests, since they are effectively the same.
For now, we only support convenience analyzers when a code action was
requested on the same line as the fix. I'm not sure how to otherwise
handle this without bothering the user with unnecessary diagnostics.
Change-Id: I7f0aa198b5ee9964a907d709bae6380093d4ef21
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237687
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
We were marking normal tests as workspace packages, but pruning x_tests.
Don't do that.
Fixesgolang/go#39578.
Change-Id: Ia8eceea38e63ac17e40f50d6837de9c7a668469b
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237943
Run-TryBot: Heschi Kreinick <heschi@google.com>
Reviewed-by: zikaeroh <zikaeroh@gmail.com>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
The go command ignores go files in specially-named directories like
testdata. We probably still want to try to offer services like
autocomplete and jump-to-definition there, but diagnostics are less
likely to be helpful. As a compromise, just squash diagnostics.
Note that the rules we use are slightly wrong; see the comment on
ignoreFiles for details.
Fixesgolang/go#39563.
Change-Id: I0bc00639e68bd71ea55d867af36e07ef4ec780a1
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237638
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
The current implementation of the fill struct tool is not a part of
the analysis framework. This commit moves the functionality from the
source directory to the analysis directory.
Change-Id: Ibe37b57f3e6680c8729932dbbe010a4642600e4a
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237258
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Calculate and display very crude memory usage statistics. This is
complicated by various levels of sharing and indirection, so the numbers
should be taken with *large* grains of salt and interpreted mostly by
experts.
Still, the results are interesting and helpful.
Change-Id: Ia9aff974c7d5fddd63df0cfd5cecc08ead33cf84
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236163
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
ModHandle races with the initial workspace load if the go.mod file does
not yet exist. We should await for the initial workspace load to
complete before proceeding with update codelenses, etc.
Part of trying to figure out the flakes in golang/go#39504.
Also a few staticcheck fixes, and fix the Windows line endings in
fill_struct.go, because `git gofmt` complains.
Change-Id: Ide21a47137390792d1afb924740cff0bb6f0b764
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237419
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
The error messages from view cancellation clutter up the logs when
testing, especially if you're running a single subtest.
A few quick staticcheck fixes in the CL also.
Change-Id: Ia1ed5360ac754023c589ed526ec0ed3e94a79b2f
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237637
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
The cache does not know the sessions associated with it, so the
debug template cannot display them.
Change-Id: I9b043592351ddce097f4e74f3e9aa7a6ad24e613
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237618
Run-TryBot: Ian Cottrell <iancottrell@google.com>
Reviewed-by: Robert Findley <rfindley@google.com>
Ignore ignored the builtin package and files that start with _. The
latter should already be ignored by "go list". The former seems
like too much effort to me. People shouldn't edit random parts of the
stdlib, and ignoring changes to (e.g.) the Error interface seems like
the least of the trouble they can get themselves into.
Remove it for now. If we get complains I'll re-add it, probably by
rejecting the write entirely somewhere.
We incidentally relied on this in the identifier functions; change those
to treat the builtin package slightly more specially.
Change-Id: I005b02a66b1a987c50a3074d53a2d28ff07d3324
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237597
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
When you write a test in (say) the fmt package, you get a test variant
augmented with the test files. In many cases you also get test variants
of the things the fmt package depends on. The primary test variant,
(fmt [fmt.test]) is interesting to us, because it contains the tests.
But the intermediate variants (testing [fmt.test]) aren't -- the user
can only get to them indirectly. We certainly don't need to fully parse
them.
Treat intermediate test variants as non-workspace packages. This doesn't
accomplish much yet but paves the way for later optimizations.
Updates golang/go#36943.
Change-Id: I1a20abcd2d67767f07132a75a20f098be6f19a76
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236397
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
We use file identities pervasively throughout gopls. Prior to this
change, the identity is the modification date of an unopened file, or
the hash of an opened file. That means that opening a file changes its
identity, which causes unnecessary churn in the cache.
Unfortunately, there isn't an easy way to fix this. Changing the
cache key to something else, such as the modification time, means that
we won't unify cache entries if a change is made and then undone. The
approach here is to read files eagerly in GetFile, so that we know their
hashes immediately. That resolves the churn, but means that we do a ton
of file IO at startup.
Incidental changes:
Remove the FileSystem interface; there was only one implementation and
it added a fair amount of cruft. We have many other places that assume
os.Stat and such work.
Add direct accessors to FileHandle for URI, Kind, and Version. Most uses
of (FileHandle).Identity were for stuff that we derive solely from the
URI, and this helped me disentangle them. It is a *ton* of churn,
though. I can revert it if you want.
Change-Id: Ia2133bc527f71daf81c9d674951726a232ca5bc9
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237037
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
The shared server was executing on ctx.Background(), which meant it
didn't have a debug.Instance. This resulted in logs being printed to
stderr, due to the fallback behavior of the global exporter.
Fixesgolang/go#39130
Change-Id: Ibb968f78d69752452bec71a7abeff808b1cccb04
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237583
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Cottrell <iancottrell@google.com>
Use the -v flag to control whether RPC logs are always printed, rather
than a regtest specific flag.
Updates golang/go#39130
Change-Id: Ie6293815adee4b59defd80cfc015838cfbf2b3e8
Reviewed-on: https://go-review.googlesource.com/c/tools/+/235920
Run-TryBot: Robert Findley <rfindley@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Peter Weinberger <pjw@google.com>
The only substantial change is data types for CallHeirarchy.
util.ts has changed the hash, and adapted to a new source layout, plus
the usual pointless whitespace changes. code.ts has learned a little more
about typescript ASTs.
Change-Id: I9cb3a9a9034d46f4a479123779da3bb3474e4a42
Reviewed-on: https://go-review.googlesource.com/c/tools/+/237377
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
The existing implementation looks for matching types in the declaration
and the leftmost value in the return statement. Instead, the analyzer
now searches across all the values in the return statement to find one
that matches the type in the declaration. Additionally, if a value in
the return statement does not match any type in the declaration, it is
not overriden.
Change-Id: I4d4aed0ef67e59bfd886b44055d523a8c478255c
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236962
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
This uses log messages to convey information to the debug system, which
has the benefit of logging even if the debug pages are not active and
also not requiring systems to reach into the debug system or require
extra lifetime tracking Not all things are decoupled yet as there are a
couple of places (notably the handshaker) that read information out of
the debug system.
Change-Id: Iec1f81c34ab3b11b3e3d6e6eb39b98ee5ed0d849
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236337
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
When document.Highlight is called with the cursor on a loop statement or
branch statement, gopls doesn't look for labels. Placing the cursor at
the break statement below highlights the inner for loop:
Outer:
for {
for {
break Outer
}
}
By making highlight label aware, and ensure that unlabeled "break" in
"switch"/"select" doesn't highlight the outer loop, this change fixes
loop highlighting.
Adding support for highlight of "switch" and "select" will be handled in
a separate CL.
Updates golang/go#39275
Change-Id: I7014aa7b0dfb1da871863ced611623be995f3944
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236524
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This moves as much code outside the protocol generator
as possible making it easier to maintain both the code
and the generator.
Change-Id: I7fe932a58facece5bb0bd5a9c158e5cc7d5a277b
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236838
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
regtests can use Await to wait for diagnostic expectations. But sometimes
it is useful (or more robust) to then look at the specific diagnostics.
This change introduces env.DiagnosticsFor, which returns the current
diagnostics for a file.
Change-Id: Iea35d28f6679289795bc853f156aae351279b205
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236837
Run-TryBot: Peter Weinberger <pjw@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
This removes the interfaces and the debug structs in the lsprpc package
in favour of just having the debug structs in the debug package.
Change-Id: I67541688444f4ef367007740c5446dcd7be6575f
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236198
Reviewed-by: Robert Findley <rfindley@google.com>
This removes all the cache/session/view hooks from the lsp and instead
uses normal introspection from the client to find all the entities
rather than keeping shadow copies of the object graph in the debug page
handler.
This required the addition only of the ability to view the sessions open
on a cache and exposing the unique identifier for all the objects, both
of which are useful and reasonable things to have in the API anyway.
Change-Id: Ic19903e7b19832ca79290954ec373d4177789742
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236197
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
This removes a TODO in the tests and has them check for shutdown
correctly.
This also required adding a Close to editor that does the Shutdown/Exit
sequence rather than just Shutdown.
Change-Id: I6433b76eb2dd1fe40b2332685fdf25ebc4fd4577
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236717
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
The existing implementation contains strict equality rules for
comparing the return type in the function declaration against
the return types in the function body. This causes the code in the
return statement to be incorrectly overriden with its corresponding
zero value type by the fillreturns analyzer.
Fixesgolang/go#38707
Change-Id: I07522aaf85f2a5f811a86cbc0951ae61035ff55e
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236617
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
In addition to adding a `go mod vendor` command option, which can be
exposed via an editor client frontend, we show a suggestion to users who
experience the "inconsistent vendoring" error message.
The main change made here is that we save the view initialization error,
and we return it if the view has absolutely no metadata. This seems
reasonable enough, but my fear is that it may lead to us showing
outdated error messages. I will spend some time improving the handling
of initialization errors in follow-up CLs.
Updates golang/go#39100
Change-Id: Iba21fb3fbfa4bca956fdf63736b397c47fc7ae44
Reviewed-on: https://go-review.googlesource.com/c/tools/+/235619
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
It was directly generating messages and sending them on the conn, now it
just uses an editor method like all the other tests.
It was also broken because it never opened the file it was hovering in, so I am
not sure it was testing anything useful before.
Change-Id: I7a1b444015c95c82a0a137d3bb1da661ed9331af
Reviewed-on: https://go-review.googlesource.com/c/tools/+/232983
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
Really the name is wrong now, but this is just a stepping stone towards removing
it entirely in favour of a new listener/dialer/server/client pattern, so I am
minimizing the churn by leaving the names alone for now.
Change-Id: I771d117490763ebe05ed2a8c52d490deeb4d5333
Reviewed-on: https://go-review.googlesource.com/c/tools/+/232878
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
This will allow varying implementations and wrappers, and more
closely matches the concepts used in the net library.
Change-Id: I4be4c6efb3def0eda2693f482cbb0c6f776e5642
Reviewed-on: https://go-review.googlesource.com/c/tools/+/232877
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Findley <rfindley@google.com>
When testing Highlight the highlight count is checked against expected
number of highlights. If it doesn't match t.Errorf(...) is called and
the test continues.
A few lines below the test ranges over results using the index for both
result and expected result leading to a panic if there are less then
expected highlights.
This change fails fast with t.Fatalf(...) instead to avoid the panic.
Change-Id: I33d0973f3145c307d9084d037ffbb73b244a3acb
Reviewed-on: https://go-review.googlesource.com/c/tools/+/236099
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Run-TryBot: Rebecca Stambler <rstambler@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
