Documentation change only. The binary will not be installed
using the "go tool" mechanism.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/133710046
(godoc is excluded from this CL since it will continue to use
/src/pkg in its URL namespace, making the necessary cleanup
more subtle.)
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/141770043
Missed comment from previous code review.
Next up: execution tests so this won't happen again
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/134480043
Improve the generated code by using a const instead of a var for the names string.
This requires some refactoring to get neat const() and var() blocks.
Also change the generate map code go use a single sliced string, to reduce the
size of the compiled representation (only one string value).
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/135450044
Refactor a little to make testing easier.
Add golden tests and a check fo splitIntoRuns, which is the subtlest piece.
Still to come: execution tests.
Also fix a few issues in the generated code.
LGTM=gri
R=gri
CC=golang-codereviews, josharian
https://golang.org/cl/134450044
This tool creates String methods from constant definitions.
It's a time-saver designed to be used from go generate.
The methods generated are efficient, more so than one
is likely to create by hand.
Given
package date
type Day int
const (
Monday Day = iota
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
)
the command
stringer -type Day
will create the file day_string.go containing
package date
import "fmt"
var (
_Day_indexes = []uint8{6, 13, 22, 30, 36, 44, 50}
_Day_names = "MondayTuesdayWednesdayThursdayFridaySaturdaySunday"
)
func (i Day) String() string {
if i < 0 || i >= Day(len(_Day_indexes)) {
return fmt.Sprintf("Day(%d)", i)
}
hi := _Day_indexes[i]
lo := uint8(0)
if i > 0 {
lo = _Day_indexes[i-1]
}
return _Day_names[lo:hi]
}
There are several strategies for the created method chosen according to
the structure of the sequence of constants.
Handles integer types only, both signed and unsigned. That's probably
all that is needed.
Tests to follow, but the test structure will be large so sending this out
separately. The code has been heavily hand-tested but there are
some bugs. Don't depend on this until the tests are installed.
LGTM=gri
R=gri, josharian
CC=golang-codereviews
https://golang.org/cl/136180043
LookupFieldOrMethod now also decides whether a found
method is actually in the method set. Simplifies call
sites. Added corresponding API tests.
TODO (separate CL): Decide what the correct value for
the indirect result should be (as required for code
generation). For now, the result value for indirect
is unchanged from before if a field/method is found.
Fixesgolang/go#8584.
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/132260043
e.g. chmod +w, checkout.
Also: add a TIPS section to the documentation.
LGTM=crawshaw
R=crawshaw, gri
CC=golang-codereviews
https://golang.org/cl/136780044
Be careful not to complain about math.Log and cmplx.Log.
Seems worthwhile since t.Log and t.Logf are often written but
rarely executed.
Nothing new turned up in the standard library.
Fixesgolang/go#8504.
LGTM=josharian, dsymonds
R=golang-codereviews, josharian, dsymonds
CC=golang-codereviews
https://golang.org/cl/130490043
s/enclosed by function/captured by func literal/
Users complained. They often do.
LGTM=josharian, adg
R=golang-codereviews, josharian, nightlyone, minux, adg
CC=golang-codereviews
https://golang.org/cl/132080043
This CL aims to fix the problem described here:
https://groups.google.com/forum/#!topic/golang-nuts/R6ms1n9KjiY
This makes it easier to parse via external tools such as editors. Editors can
show each function in a list and jump directly to each function with this
additional information. This pattern can be seen in other Go tools such as "go
test" in the form of:
--- FAIL: TestCover (0.52 seconds)
cover_test.go:43: example error
LGTM=adg
R=r, adg, josharian, dave
CC=golang-codereviews
https://golang.org/cl/131820043
We want to make an if look like two blocks and have the coverage
report for the else block decorate the "else" keyword with the right
color. To do this, we adjust the apparent starting point of the else
block to include the "else".
The previous code assumed the way to do this was to move the
width of "else " backwards from the else block's opening brace, but
that assumes there is a space there. Instead, we now just start the
else block exactly at the end of the if block. Simpler, cleaner, and
fixes a bug.
Fixesgolang/go#8557.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/127620043
As requested on Stack Overflow: http://goo.gl/ams9fY
(Kudos to sberry for his JavaScript solution, provided there.
This change does the same thing on the server side.)
LGTM=r
R=r
CC=golang-codereviews
https://golang.org/cl/127030043
And serialize the printing of each item with a mutex.
It is the formatted output of this tool, after all.
Also: minor doc tweaks.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/114620044
`go test` takes -run and -bench; the -test.run and -test.bench flags
are only for the test binary itself.
LGTM=adg
R=adg
CC=golang-codereviews
https://golang.org/cl/113390043
This change allows the directory front page to be more easily configurable.
Templates are now read only at start-up and stored in a map rather than re-parsed for each page rendering.
LGTM=adg
R=adg
CC=golang-codereviews
https://golang.org/cl/109080044
text/tabwriter cells are tab-terminated, not tab-separated. This creates trailing whitespace. However, with left-aligned columns, it is ok not to treat the final element as a cell. Demo: http://play.golang.org/p/m_ajG8SSZe
LGTM=gri
R=golang-codereviews, gobot, gri
CC=golang-codereviews, r
https://golang.org/cl/103650043
This CL introduces two vet checks. Statistics and code below are from a recent 50mb corpus of public code.
1. Check for redundant conjunctions and disjunctions. This check caught 26 instances, of which 20 were clearly copy/paste bugs and 6 appeared to be mere duplication. A typical example:
if xResolution < 0 || xResolution < 0 {
panic("SetSize(): width < 0 || height < 0")
}
2. Check for expressions of the form 'x != c1 || x != c2' or 'x == c1 && x == c2', with c1 and c2 constant expressions. This check caught 16 instances, of which all were bugs. A typical example:
if rf.uri.Scheme != "http" || rf.uri.Scheme != "ftp" {
rf.uri.Scheme = "file"
}
Fixesgolang/go#7622.
LGTM=rsc, r
R=golang-codereviews, jscrockett01, r, gri, rsc
CC=golang-codereviews
https://golang.org/cl/98120043
Really two fixes: Don't panic on bad instructions and don't complain about commented out instructions.
LGTM=r
R=r
CC=golang-codereviews
https://golang.org/cl/110070044
Clients such as compilers need this information in order
to correctly link against imported packages.
This also adds support for the condensed import data format
where the priority information is stored as a suffix of the
condensed import data, as well as support for archive files.
LGTM=gri
R=gri
CC=golang-codereviews, iant
https://golang.org/cl/78740043
Bare init functions omit calls to dependent init functions and the
use of an init guard. They are useful in cases where the client uses
a different calling convention for init functions, or cases where
it is easier for a client to analyze bare init functions.
LGTM=adonovan
R=adonovan
CC=golang-codereviews, iant
https://golang.org/cl/78780043
This is a common source of bugs, particularly for those new to Go. Running this on a corpus of public code flagged 114 instances.
This check may need to be updated once issue 7363 is resolved.
LGTM=r
R=golang-codereviews, r
CC=bradfitz, golang-codereviews
https://golang.org/cl/91010047
In command godoc, set IndexEnabled when the -write_index flag is set.
Previously you would need to (unintuitively) set the -http flag to
achieve this.
In package godoc, set up the FS tree before loading the index, and
then return before starting the index refresh loop. Previously the
index would be loaded and then immediately refreshed, negating the
benefits of the on-disk index.
TBR=bradfitz
R=golang-codereviews
CC=golang-codereviews
https://golang.org/cl/103370046
This removes much of the AST logic out of main.go,
and makes it easier to build custom vet binaries
The trade-off in this change is for flexibility.
There's very little change in the per-check files,
a lot less code in main.go (specifically the AST walking
logic has shrunk), and it makes it much easier to build
custom vet binaries simply by dropping new source files
in the directory.
LGTM=josharian, r
R=r, josharian, kamil.kisiel
CC=golang-codereviews
https://golang.org/cl/83400043
This will fix the images in Brad's GoCon presentation.
LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/100950043
Also fixes the following nits;
- literal IPv6 address handling
- URL host component handling in the case of a wildcard listen
- URL port component handling in the case of no port component in origin
Fixesgolang/go#8096.
LGTM=dan.kortschak, adg
R=adg, golang-codereviews, dan.kortschak
CC=golang-codereviews
https://golang.org/cl/102770046
If you have multiple runs in old.txt and new.txt
the default behavior is to match them up pairwise
and compare successive pairs (and if you have a
different number of runs in each file, benchcmp
refuses to do anything).
The new -best flag changes the behavior to instead
compare the fastest run of each benchmark from
the two files. This makes sense if you believe that
the fastest speed is the 'actual' speed and the slower
results are due to the computer spending time doing
non-benchmark work while the benchmark was
running.
LGTM=josharian
R=golang-codereviews, josharian
CC=golang-codereviews
https://golang.org/cl/102890047
They were disabled by mistake during the move to go.tools.
LGTM=dan.kortschak
R=golang-codereviews, dan.kortschak
CC=golang-codereviews
https://golang.org/cl/98440048
It was very ugly; a little tweaking helps godoc parse it better.
Also make unsafeptr.go not own the package doc (add a blank line)
and put one more sentence about that check into doc.go.
Fixesgolang/go#7925.
LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/98370044
Ignore calls to various flavours of atomic.AddInt with a wrong
number of arguments.
LGTM=r
R=golang-codereviews, r
CC=golang-codereviews
https://golang.org/cl/91370045
The old code printed the underlying type; e.g., the type
of time.Millisecond was reported to be int64 rather than
time.Duration.
Testsuite (and corresponding tests) in progress (another CL).
LGTM=adonovan
R=adonovan, dsymonds
CC=golang-codereviews
https://golang.org/cl/94770045
Assuming:
1) package declaration does not exist
2) the Fragment option is set
3) a main function exists
We will assume it is a main package and add the declaration.
This change also sets the Fragment option in goimports.
LGTM=crawshaw, bradfitz
R=bradfitz, crawshaw
CC=golang-codereviews
https://golang.org/cl/96850044
(The test thread is racing with the analysis thread, which
takes around 4ms on this input.)
LGTM=gri
R=gri, bradfitz
CC=golang-codereviews
https://golang.org/cl/89780044
The text and images are "baked in" to the godoc executable's
rodata section (~300KB) and are accessible from the godoc
server itself at /lib/godoc/analysis/help.html.
In due course, the page will become visible at
http://golang.org/lib/godoc/analysis/help.html, which will be
the canonical location for this doc (in announcements, etc).
The page is temporarily visible here, for those on the Google corp network:
http://172.26.104.127:7777/lib/godoc/analysis/help.html
Also:
- add link to new doc from source view pages.
- document -analysis flag in cmd/godoc/doc.go
- fix indentation of -analysis flag's help string
LGTM=gri
R=gri, bgarcia, r
CC=golang-codereviews
https://golang.org/cl/87110045
Details:
- auto-generate prefixes for std lib (e.g., "godex big" works now)
- apply filtering to package-level objects only
- nicer formatting of single-entry const, var, or type declaration
TBR=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/81360046
All of these work now as expected:
godex code.google.com/p/go.tools/go/types
godex go.tools/go/types
godex go/types
godex types
Also improved logging/verbose mode.
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/80930043
1) Split a path.name argument at the last '.' that
is not part of the path.
2) Try various importers always in the same order
for consistent results (use lists instead of maps).
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/80790043
See analysis.go for overview of new features.
See README for known bugs and issues.
Much UI polish, testing and optimization work remains, but
this is a starting point.
Flag: we add a new flag -analysis=type,pointer, default "",
for adventurous users only at this stage.
Type analysis takes ~10s for stdlib + go.tools;
Pointer analysis (currently) takes several minutes.
Dependencies: we now include jquery.treeview.js and its GIF
images among the resources. (bake.go now handles binary.)
LGTM=crawshaw, bgarcia
R=crawshaw, bgarcia
CC=bradfitz, golang-codereviews
https://golang.org/cl/60540044
It's pointless.
Also this fixes a crash, because the blank identifier no longer appears as a
defined object after CL 74190043 so we were getting nil pointer violations.
Even better, we get to re-enable a disabled test.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/75140043
Now we can say
vet -printf=false
to disable the printf test but run all others.
Implemented by creating a tri-state boolean flag that records whether it has been
set explicitly; before this, -printf=false was not distinguishable from not having
mentioned the printf flag at all.
Fixesgolang/go#7422.
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/72330043
Over time, a number of modules were added that used Warn instead of Bad
to report problems with the code, but the documentation states that
if there is a problem, the exit code must be 1, not 0. Warn does not set the
exit code and should be used only for internal errors and messages
triggered by the -v flag.
There's nothing substantive here except calling the other function in a few
places.
Fixesgolang/go#7017.
LGTM=crawshaw
R=golang-codereviews, crawshaw
CC=golang-codereviews
https://golang.org/cl/71860044
The old code was misleading in saying how many args were present.
Change the wording of the message to be unambiguous and change
the presentation of the format to include the full directive, making
it easier to correlate with the input (and fixing a silent bug).
Fixesgolang/go#6248.
LGTM=dsymonds
R=golang-codereviews, dsymonds
CC=golang-codereviews
https://golang.org/cl/69120044
An identifier X in anonymous struct field struct{X} is both a
definition of a field (*Var) and reference to a type
(*TypeName). Now that we have split the map, we can capture
both of these aspects.
Interestingly, every client but one was going to extra effort
to iterate over just the uses or just the defs; this
simplifies them.
Also, fix two bug related to tagless switches:
- An entry was being recorded in the Object map for a piece of
synthetic syntax.
- The "true" identifier was being looked up in the current scope,
which allowed perverse users to locally redefine it. Now
we use the bool (not untyped boolean) constant true, per the
consequent clarification of the spec (issue 7404).
+ tests.
Fixesgolang/go#7276
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/68270044
How it handles packages vs. directories vs. files was not explained.
LGTM=rsc
R=golang-codereviews, gobot, rsc
CC=golang-codereviews
https://golang.org/cl/67150043
benchcmp now preserves benchmark order. This restores the original
misc/benchcmp behavior. This also makes the output of benchcmp stable,
and groups together multiple -cpu results.
Magnitude-based sorting is still available via the -mag flag.
It is useful for surfacing items of note (particularly changes
in allocs) when making compiler changes and running broad
benchmarks.
Fixesgolang/go#7259.
LGTM=dave
R=dave, mtj
CC=bradfitz, dvyukov, golang-codereviews
https://golang.org/cl/60840045
These are the simplest possible descriptions of each command.
They may be fleshed out later.
Fixesgolang/go#7298.
LGTM=r
R=r
CC=golang-codereviews
https://golang.org/cl/61480044
Don't say the word "fork" (not accurate), and remove the
tab/comment flags that were removed from gofmt.
LGTM=adg
R=golang-codereviews, adg
CC=golang-codereviews
https://golang.org/cl/61410052
Previously, each word could be a package import path or a
comma-separated list of *.go file names. Now, if the
first word ends with ".go", all words are assumed to be
Go source files. This makes it impossible to specify
two ad-hoc packages from source files, but no-one needs that.
FromArgs also takes a boolean indicating whether tests
are wanted or not.
Also: ssadump: add -test flag to set that boolean.
For the oracle it's always true.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/61470047
Method-set caching is now performed externally using a MethodSetCache (if desired), not by the Types themselves.
This a minor deoptimization due to the extra maps, but avoids a situation in which method-sets are computed and frozen prematurely. (See b/7114)
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/61430045
This is intended to replace the awk-based misc/benchcmp.
It mostly matches the existing misc/benchcmp.
Notable changes:
* Written in Go.
* Minor whitespace changes in the output; the tabular nature of the
output is preserved, as is most number formatting and verbiage.
* Comparisons within each section are sorted from highest change to lowest.
* Proper handling of multiple benchmarks with the same name (issue 7016).
* Does not omit benchmark comparisons for which the new value is zero.
* Adds -changed flag to only show benchmarks whose value have changed.
Useful for memory-oriented, large-scale benchmark comparisons.
* Has tests.
* Formats small ns measurements with extra precision.
Updates golang/go#7016.
LGTM=r
R=golang-codereviews, dave, dvyukov, oleku.konko, bradfitz, gobot, r
CC=golang-codereviews
https://golang.org/cl/47980043
This results in significant improvement to type-checking time:
it reduces by 4% the entire running time of ssa/stdlib_test
(GOMAXPROCS=8, n=7).
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/57770043
The Importer type has been replaced with Config and Program.
Clients populate a Config, directly or more usually via
convenience functions. They then call its Load() method to do
all of the typechecking and transitive-closure computation.
ssa.NewProgram and ssa.CreatePackages have been fused into
ssa.Create, which now cannot fail, since (*Config).Load()
reports all type errors.
Also:
- The addition of an ssa.GlobalDebug builder mode flag
eliminates a loop-over-packages repeated in many clients.
- PackageInfo.Err flag unexported. Clients never see bad infos now.
- cmd/ssadump: now only looks for func "main" in package "main".
- importsOf deleted, was dead code.
STILL TODO:
- ParseFile seems like API creep (though it's convenient)
and CreateFromFiles is dangerous (w.r.t. FileSet identity).
Need to think more...
- the need for clients to rely on elementwise correspondence
of Config.CreatePkgs and Program.Created is a little sad.
- The command-line interface has not changed.
That will happen in a follow-up.
r recommends using a repeated flag: -package p -package q ...
R=gri
CC=axwalk, frederik.zipp, golang-codereviews
https://golang.org/cl/49530047
Packages were not being created for all types.Packages,
specifically, indirectly imported packages were missing.
(*Program).CreatePackages now iterates over the type-checker's
package map too.
Also: removed all concurrency from importer. I think it was
incorrect (and hard to fix).
Also: change LoadInitialPackages so that all named packages
are loaded from source. This happens regardless of whether
GCImporter is used to satisfy imports.
Details:
- importer.Config.SourceImports flag determines whether to
load all packages from *.go source.
(Before, this was indicated by Config.Build != nil.)
- importer.Config.Build field effectively defaults to
&go/build.Default. A zero importer.Config is now usable.
- importer.Importer.Config field is now exported.
- LoadPackage renamed to ImportPackage since the resulting
packages may come from GCImporter (and be incomplete).
- doImport and ImportPackage fused.
Fixesgolang/go#7028
R=gri, axwalk
CC=golang-codereviews
https://golang.org/cl/48770043
Also report an error for "cross-interpretation": not supported
due to the interpreter's assumption that host and target
{int,uint,uintptr} are the same. (Too tedious and messy to fix.)
Tested manually:
% cat d.go
package main
const m = ^uintptr(0)
const w = m>>8&1 + m>>16&1 + m>>32&1
func main() { println(m, w) }
% ./ssadump -build=P -run d.go
package main:
const m m = 18446744073709551615:uintptr
const w w = 3:uintptr
18446744073709551615 3
% GOARCH=386 ./ssadump -build=P -run d.go
package main:
const m m = 4294967295:uintptr
const w w = 2:uintptr
Error: Cross-interpretation is not yet supported (target has GOARCH 386, interpreter has amd64).
Fixesgolang/go#7080
R=gri
CC=golang-codereviews
https://golang.org/cl/49070043
The display of search results can now be changed. A slice of functions for displaying the results can now be provided. By default, three functions are provided to display documentation, source code, and textual results.
This makes it possible to replace them with equivalents that, say,
obtain search results from alternative source code search engines.
R=bradfitz, adg
CC=golang-codereviews
https://golang.org/cl/43470043
Not for Go 1.2. Still needs a flag.
Linux at least (and likely other OSes) don't like you doing
a few hundred readdirs at once and spawing as many threads.
R=golang-dev, adg, jeremyjackins
CC=golang-dev
https://golang.org/cl/30620043
Command set:
- what: an extremely fast query that parses a single
file and returns the AST stack, package name and the
set of query modes that apply to the current selection.
Intended for GUI tools that need to grey out UI elements.
- definition: shows the definition of an identifier.
- pointsto: the PTA features of 'describe' have been split
out into their own command.
- describe: with PTA stripped out, the cost is now bounded by
type checking.
Performance:
- The importer.Config.TypeCheckFuncBodies predicate supports
setting the 'IgnoreFuncBodies' typechecker flag on a
per-package basis. This means we can load dependencies from
source more quickly if we only need exported types.
(We avoid gcimport data because it may be absent or stale.)
This also means we can run type-based queries on packages
that aren't part of the pointer analysis scope. (Yay.)
- Modes that require only type analysis of the query package
run a "what" query first, and restrict their analysis scope
to just that package and its dependencies (sans func
bodies), making them much faster.
- We call newOracle not oracle.New in Query, so that the
'needs' bitset isn't ignored (oops!). This makes the
non-PTA queries faster.
Also:
- removed vestigial timers junk.
- pos.go: existing position utilties split out into own file.
Added parsePosFlag utility.
- numerous cosmetic tweaks.
+ very basic tests.
To do in follow-ups:
- sophisticated editor integration of "what".
- better tests.
- refactoring of control flow as described in comment.
- changes to "implements", "describe" commands.
- update design doc + user manual.
R=crawshaw, dominik.honnef
CC=golang-dev, gri
https://golang.org/cl/40630043
Using a type containing a sync type directly
in a function call (whether as a receiver,
a param, or a return value) is an easy way
to accidentally copy a lock or other sync primitive.
Check for it.
The test as implemented does not provide 100%
coverage; see the discussion near the bottom of
testdata/copylock.go for shortcomings.
Fixesgolang/go#6729.
R=adg, r, dsymonds
CC=golang-dev
https://golang.org/cl/23420043
Split the profile parsing code out of cmd/cover into
a reusable package, to support third-party coverage
tools.
R=golang-dev, r, adg
CC=golang-dev
https://golang.org/cl/36050045
Add explicit options to Corpus to control search indexing of documentation, Go source code, and full-text.
R=bradfitz, r
CC=golang-dev
https://golang.org/cl/24190043
- fixed a couple of TODOs
- various cleanups along the way
- adjusted clients
Once submitted, clients of go/types that don't explicitly
specify Config.Import will need to add the extra import:
import _ "code.google.com/p/go.tools/go/gcimporter"
to install the default (gc) importer in go/types.
R=adonovan, gri
CC=golang-dev
https://golang.org/cl/26390043
Performance increase is much smaller than expected;
need to investigate. Possibly, repeatedly (for each
line) accessing file sets from multiple goroutines
(in the scanner) is a bottle neck; or perhaps i/o.
R=adonovan
CC=golang-dev
https://golang.org/cl/23090043
The symbolic names such as NOSPLIT for annotations on the TEXT
directive appeared after vet started checking .s files. This CL tweaks
the regular expression to allow CAPITALS and the symbols | and +
as well as digits in that field, and interprets NOSPLIT as equivalent
to 7 in that field. All magic.
Fixesgolang/go#6695
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/18700044
This CL makes gotype usable again. Removed -r
(recursive) mode; use go/build to determine
the correct set of Go files when processing
a directory. The -v (verbose) mode now prints
some basic stats (duration, number of files,
lines, and lines/s).
Thoroughly restructured the code.
Applying gotype -v -a . to the go/types directory:
128.94141ms (40 files, 12008 lines, 93127 lines/s)
On a 2.8 GHz Quad-Core Intel Xeon, 800 MHz DDR2 FB-DIMM,
with go/types built with the (interal) debug flag set to
false. There's still quite a bit of room for performance
improvement in all parts of the code since no tuning has
been done.
R=golang-dev, adonovan
CC=golang-dev
https://golang.org/cl/19930043
This allows us to run/analyze multiple tests.
Also it causes the production code packages to be properly initialized.
Also:
- cmd/ssadump: improved usage message (add example;
incorporate LoadInitialPackages usage; explain how -run
finds main).
- pointer, oracle, ssa/interp: use CreateTestMainPackage.
- ssa/builder.go: remove 'rundefers' instruction from package init,
which no longer uses 'defer'.
R=gri
CC=golang-dev
https://golang.org/cl/15920047
A function such as this:
func one() (x int) {
defer func() { recover() }()
x = 1
panic("return")
}
that combines named return parameters (NRPs) with deferred calls
that call recover, may return non-zero values despite the
fact it doesn't even contain a return statement. (!)
This requires a change to the SSA API: all functions'
control-flow graphs now have a second entry point, called
Recover, which is the block at which control flow resumes
after a recovered panic. The Recover block simply loads the
NRPs and returns them.
As an optimization, most functions don't need a Recover block,
so it is omitted. In fact it is only needed for functions that
have NRPs and defer a call to another function that _may_ call
recover.
Dataflow analysis of SSA now requires extra work, since every
may-panic instruction has an implicit control-flow edge to
the Recover block. The only dataflow analysis so far implemented
is SSA renaming, for which we make the following simplifying
assumption: the Recover block only loads the NRPs and returns.
This means we don't really need to analyze it, we can just
skip the "lifting" of such NRPs. We also special-case the Recover
block in the dominance computation.
Rejected alternative approaches:
- Specifying a Recover block for every defer instruction (like a
traditional exception handler).
This seemed like excessive generality, since Go programs
only need the same degenerate form of Recover block.
- Adding an instruction to set the Recover block immediately
after the named return values are set up, so that dominance
can be computed without special-casing.
This didn't seem worth the effort.
Interpreter:
- This CL completely reimplements the panic/recover/
defer logic in the interpreter. It's clearer and simpler
and closer to the model in the spec.
- Some runtime panic messages have been changed to be closer
to gc's, since tests depend on it.
- The interpreter now requires that the runtime.runtimeError
type be part of the SSA program. This requires that clients
import this package prior to invoking the interpreter.
This in turn requires (Importer).ImportPackage(path string),
which this CL adds.
- All $GOROOT/test/recover{,1,2,3}.go tests are now passing.
NB, the bug described in coverage.go (defer/recover in a concatenated
init function) remains. Will be fixed in a follow-up.
Fixesgolang/go#6381
R=gri
CC=crawshaw, golang-dev
https://golang.org/cl/13844043
Break the basic block at the function literal. The code to do this analysis
was already there; this CL just factors it out more nicely and uses it in
one new place. Also adds a test.
Fixesgolang/go#6555.
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/14601043
- removed support for nil constants from go/exact
- instead define a singleton Nil Object (the nil _value_)
- in assignments, follow more closely spec wording
(pending spec CL 14415043)
- removed use of goto in checker.unary
- cleanup around handling of isRepresentable for
constants, with better error messages
- fix missing checks in checker.convertUntyped
- added isTyped (== !isUntyped) and isInterface predicates
- fixed hasNil predicate: unsafe.Pointer also has nil
- adjusted ssa per adonovan
- implememted types.Implements (wrapper arounfd types.MissingMethod)
- use types.Implements in vet (and fix a bug)
R=adonovan, r
CC=golang-dev
https://golang.org/cl/14438052
Present the files in lexical order so the output is reproducible
and easier to navigate. Do a little type rearrangement to simplify
things while we're there.
R=adg
CC=golang-dev
https://golang.org/cl/14357043
The redirect.PrefixHandler redirects "/foo/" to "/foo", which is what
we want in most cases ("/cl/", "/change/", etc) but not here.
So wrap it with a handler that handles "/blog/" explictly.
R=dsymonds
CC=golang-dev
https://golang.org/cl/14326043
