Age | Commit message (Collapse) | Author |
|
Fix suggested by Minux.
Change-Id: Ia7aa8ccccc16535af4ec3ad23830ef0aa5d776ac
Reviewed-on: https://go-review.googlesource.com/27193
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Minux Ma <minux@golang.org>
|
|
This makes a bunch of changes to package syntax to tweak line numbers
for AST nodes. For example, short variable declaration statements are
now associated with the location of the ":=" token, and function calls
are associated with the location of the final ")" token. These help
satisfy many unit tests that assume the old parser's behavior.
Because many of these changes are questionable, they're guarded behind
a new "gcCompat" const to make them easy to identify and revisit in
the future.
A handful of remaining tests are too difficult to make behave
identically. These have been updated to execute with -newparser=0 and
comments explaining why they need to be fixed.
all.bash now passes with both the old and new parsers.
Change-Id: Iab834b71ca8698d39269f261eb5c92a0d55a3bf4
|
|
Introduce a new noder type to transform package syntax's AST into gc's
Node tree. Hidden behind a new -newparser flag.
Change-Id: Id0e862ef6196c41533876afc4ec289e21d422d18
|
|
Change-Id: Ie6dd2318e031be445c0b1ae65d4c78723d5a1167
|
|
Change-Id: I64baf334a35c72336d26fa6755c67eb9d6f4e93c
|
|
No behavior change.
Change-Id: I93c7cb7ab76b5afa74e6a1f092b7045fb96467b1
|
|
Fixes cmd/go's TestIssue7108.
Change-Id: I8436b0e3e5a9b36649d46a9c2c741d820115a5d1
|
|
If we the caller provided an error handler, still return our best
effort parsed AST tree.
|
|
pos and line already refer to the position immediately before the
returned newline character, so no need to decrement them further.
Change-Id: I1d7a32cde55a5b8a1c6fd882ec8ba532869cc858
|
|
For string/comment not terminated errors, report error at the
start of the string/comment rather than the end of the file.
|
|
- various minor cleanups related to literal scanning
- report newlines in character and regular string literals (plus tests)
|
|
Change-Id: I5e2522d34720d7687e1e9ff0831936702976a1c7
|
|
Fixes typeswitch3.go and at least recognizes the error in typesw.go
(albeit wrong line number and non-pretty-printed syntax.Expr output).
Change-Id: I38b8e923265b0e7b3c301aea2f4a901bbabc24b5
|
|
+ minor documentation improvements
|
|
|
|
Fixes fixedbugs/bug031.go.
Change-Id: Icc1846368802d71ade14e1a6b79cb6ee651b663e
|
|
+ many more test cases
|
|
The source's rune reader logic was overly clever and tried to
determine if an invalid rune was due to an actual source error
or due to not enough bytes in the buffer.
Replaced with slightly optimized version of bufio Reader's
ReadRune method for now. Performance may be a tiny bit lower
but we can always optimize later.
Fixes endless loops when parsing certain source files.
Added test case to ScanError test.
|
|
If an ErrorHandler is provided with a syntax.ReadXXX
function, it is invoked for each error encountered.
Will need to be refined, but should enable progress
with all.bash.
Also:
- added tests for lexical errors
- fixed endless loops when encountering non-terminated
strings and comments
|
|
|
|
Fixes test/eof.go.
Change-Id: Idaa2713bb1669ec165f3a2687fcfa2f8d8c70c74
|
|
Previously, we failed to recognize "**/" as the end of a general
comment.
Fixes ken/embed.go.
Change-Id: Ibed746da105453420ec2d184e7be1a5de15927a6
|
|
|
|
Change-Id: Iba8c9d9c01516706a2fc6ca5502401cfc063f210
|
|
Change-Id: Id4d9b40900b97708d2b4586f5a745dcb8b0eb8bd
|
|
Change IfStmt's Else field from []Stmt to Stmt like in go/ast.
Change-Id: I835577beaf12b6e5895bc93041c13403143b4d2d
|
|
Export token constants needed by users (Break, Continue, Defer,
Fallthrough, Go, and Goto).
Fix parsing of Continue statements: previously they would be parsed as
a BranchStmt with Break as the token.
Change aNode, aDecl, aStmt, aExpr, etc.'s tag methods to take a
pointer receiver to prevent accidents like trying to type assert from
syntax.Expr to syntax.ListExpr instead of *syntax.ListExpr.
Relocate pointer fields to the end of AST nodes so the GC can
short-circuit scanning objects.
Change-Id: Ib7505e75726816e260b9b29a2f726f76bf1a38b4
|
|
$ go test -run StdLib -fast
parsed 1074061 lines (2828 files) in 571.1019ms (1880681 lines/s)
allocated 263.676Mb (461.696Mb/s)
PASS
|
|
No performance impact:
$ go test -run StdLib -fast
parsed 1073074 lines (2823 files) in 575.606804ms (1864248 lines/s)
allocated 263.956Mb (458.570Mb/s)
PASS
|
|
Syntax tree nodes, scanner, parser, basic printers.
Builds syntax trees for entire Go std lib at a rate of ~1.8M lines/s
in warmed up state (MacMini, 2.3 GHz Intel Core i7, 8GB RAM):
$ go test -run StdLib -fast
parsed 1074617 lines (2832 files) in 579.66364ms (1853863 lines/s)
allocated 282.212Mb (486.854Mb/s)
PASS
|
|
This only triggers a few places in the stdlib,
but it helps a lot when it does.
Before:
runtime.mapassign1 t=1 size=2400 args=0x20 locals=0xe0
After:
runtime.mapassign1 t=1 size=2352 args=0x20 locals=0xd8
name old time/op new time/op delta
MapPop100-8 19.8µs ±11% 18.4µs ± 9% -7.16% (p=0.000 n=20+19)
MapPop1000-8 367µs ±17% 335µs ±11% -8.63% (p=0.000 n=19+19)
MapPop10000-8 7.29ms ±15% 6.86ms ±12% -5.84% (p=0.020 n=20+20)
Change-Id: I9faf32f95a6ba6a6d5d0818eab32cc271e01d10a
Reviewed-on: https://go-review.googlesource.com/26666
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
|
|
The frontend rewriting lowers them to runtime calls on 386. It
matches explicitly uint32, but missed uint.
Fixes #16738.
Change-Id: Iece7a45edf74615baca052a53273c208f057636d
Reviewed-on: https://go-review.googlesource.com/27085
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
|
|
Might as well tell the runtime how large the map is going to be.
This avoids grow work and allocations while the map is being built.
Will wait for 1.8.
Fixes #15880
Fixes #16279
Change-Id: I377e3e5ec1e2e76ea2a50cc00810adda20ad0e79
Reviewed-on: https://go-review.googlesource.com/23558
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
|
|
Update #16415
Change-Id: I83e0966931ada2f1ed02304685bb45effdd71268
Reviewed-on: https://go-review.googlesource.com/26665
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
This CL teaches SSA to recognize code of the form
// b is a boolean value, i is an int of some flavor
if b {
i = 1
} else {
i = 0
}
and use b's underlying 0/1 representation for i
instead of generating jumps.
Unfortunately, it does not work on the obvious code:
func bool2int(b bool) int {
if b {
return 1
}
return 0
}
This is left for future work.
Note that the existing phiopt optimizations also don't work for:
func neg(b bool) bool {
if b {
return false
}
return true
}
In the meantime, runtime authors and the like can use:
func bool2int(b bool) int {
var i int
if b {
i = 1
} else {
i = 0
}
return i
}
This compiles to:
"".bool2int t=1 size=16 args=0x10 locals=0x0
0x0000 00000 (x.go:25) TEXT "".bool2int(SB), $0-16
0x0000 00000 (x.go:25) FUNCDATA $0, gclocals·23e8278e2b69a3a75fa59b23c49ed6ad(SB)
0x0000 00000 (x.go:25) FUNCDATA $1, gclocals·33cdeccccebe80329f1fdbee7f5874cb(SB)
0x0000 00000 (x.go:32) MOVBLZX "".b+8(FP), AX
0x0005 00005 (x.go:32) MOVBQZX AL, AX
0x0008 00008 (x.go:32) MOVQ AX, "".~r1+16(FP)
0x000d 00013 (x.go:32) RET
The extraneous MOVBQZX is #15300.
This optimization also helps range and slice.
The compiler must protect against pointers pointing
to the end of a slice/string. It does this by increasing
a pointer by either 0 or 1 * elemsize, based on a condition.
This CL optimizes away a jump in that code.
This CL triggers 382 times while compiling the standard library.
Updating code to utilize this optimization is left for future CLs.
Updates #6011
Change-Id: Ia7c1185f8aa223c543f91a3cd6d4a2a09c691c70
Reviewed-on: https://go-review.googlesource.com/22711
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
|
|
For #16369.
Change-Id: I4c9f5a66b95558adcc1bcface164b9b2b4382d2f
Reviewed-on: https://go-review.googlesource.com/24979
Reviewed-by: Alan Donovan <adonovan@google.com>
|
|
Fixes #16722
Change-Id: I50a0e69d3e79d13bc1860cd983267c3db087a4b8
Reviewed-on: https://go-review.googlesource.com/27119
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
|
|
Don't fold constant factors into a multiply
beyond the capacity of a MULQ instruction (32 bits).
Fixes #16733
Change-Id: Idc213c6cb06f7c94008a8cf9e60a9e77d085fd89
Reviewed-on: https://go-review.googlesource.com/27160
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
|
|
We already inlined
_, ok = e.(T)
_, ok = i.(E)
_, ok = e.(E)
The only ok-only variants not inlined are now
_, ok = i.(I)
_, ok = e.(I)
These call getitab, so are non-trivial.
Change-Id: Ie45fd8933ee179a679b92ce925079b94cff0ee12
Reviewed-on: https://go-review.googlesource.com/26658
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
|
|
Now that assembler opcodes have their own type, they can have a true
stringer, rather than explicit calls to Aconv, which makes for nicer
format strings.
Change-Id: Ic77f5f8ac38b4e519dcaa08c93e7b732226f7bfe
Reviewed-on: https://go-review.googlesource.com/25045
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
|
|
Pick up a bunch of changes and fixes.
Change-Id: If4101f7185d433a4c89096bc786ee5de8eeabac0
Reviewed-on: https://go-review.googlesource.com/27123
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
No functional changes, makes vet happy.
Updates #11041
Change-Id: I59f3aba46d19b86d605508978652d76a1fe7ac7b
Reviewed-on: https://go-review.googlesource.com/27125
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
Makes vet happy.
Updates #11041
Change-Id: I23ca413c03ff387359440af8114786cd7880a048
Reviewed-on: https://go-review.googlesource.com/27124
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
Fixes (legit) vet warnings.
Fix some verb tenses while we're here.
Updates #11041
Change-Id: I27e995f55b38f4cf584e97a67b8545e8247e83d6
Reviewed-on: https://go-review.googlesource.com/27122
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
|
|
Give *recordingConn the correct WriteTo signature
to be an io.WriterTo. This makes vet happy.
It also means that it'll report errors,
which were previously being ignored.
Updates #11041
Change-Id: I13f171407d63f4b62427679bff362eb74faddca5
Reviewed-on: https://go-review.googlesource.com/27121
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
Caught by vet.
Updates #11041
Change-Id: I4dbb2eeaf633eea5976074840064edc2349e01d8
Reviewed-on: https://go-review.googlesource.com/27120
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
|
|
container/list/list_test.go:274: self-assignment of e1 to e1
container/list/list_test.go:274: self-assignment of e4 to e4
container/list/list_test.go:282: self-assignment of e1 to e1
container/list/list_test.go:286: self-assignment of e1 to e1
container/list/list_test.go:286: self-assignment of e4 to e4
Updates #11041
Change-Id: Ibd90cf6a924e93497908f437b814c3fc82937f4a
Reviewed-on: https://go-review.googlesource.com/27114
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
CL 23393 introduces PAUTOHEAP, and access of PAUTOHEAP variable is
rewritten to indirection of a PAUTO variable. Mark this variable
non-nil, so this indirection does not introduce extra nil checks.
Change-Id: I31853eed5e60238b6c5bc0546e2e9ab340dcddd9
Reviewed-on: https://go-review.googlesource.com/26831
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
|
|
For tests, assign to _.
For benchmarks, assign to a sink.
Updates #11041
Change-Id: I87c5543245c7bc74dceb38902f4551768dd37948
Reviewed-on: https://go-review.googlesource.com/27116
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
|
|
We were generating format strings containing
a lone %. Vet legitimately complains:
cmd/compile/internal/gc/constFold_test.go:339: unrecognized printf verb ' '
The fix doesn't make for very readable code,
but it is simple and obviously correct.
Updates #11041
Change-Id: I90bd2d1d140887f5229752a279f7e46921472fbb
Reviewed-on: https://go-review.googlesource.com/27115
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|