Several of my students were confused by trying to use both the error
return and a reply return, so I figured it was worth explicitly clarifying
that returning an error overrides the reply.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6327051
Flame motivated me to get around to adding extended key usage support
so that code signing certificates can't be used for TLS server
authentication and vice versa.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6304065
if we were to use sizeof(sa.sa_mask) instead of 8 as the last argument
to rt_sigaction, we would have already fixed this bug, so also updated
Linux/386 and Linux/amd64 files to use that; also test the return value
of rt_sigaction.
R=dave, rsc
CC=golang-dev
https://golang.org/cl/6297087
If the server replies with an HTTP response before we're done
writing our body (for instance "401 Unauthorized" response), we
were previously ignoring that, since we returned our write
error ("broken pipe", etc) before ever reading the response.
Now we read and write at the same time.
Fixes#3595
R=rsc, adg
CC=golang-dev
https://golang.org/cl/6238043
Also, fixes one violation found during testing where both
response and error could be non-nil when a CheckRedirect test
failed. This is arguably a minor API (behavior, not
signature) change, but it wasn't documented either way and was
inconsistent & non-Go like. Any code depending on the old
behavior was wrong anyway.
R=adg, rsc
CC=golang-dev
https://golang.org/cl/6307088
A comment to that effect was introduced
with rev d332f4b9cef5 but the respective
code wasn't deleted.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/6304086
Fixes a situation where a nested bad type would still
permit the outer type to install a working engine, leading
to inconsistent behavior.
Fixes#3273.
R=bsiegert, rsc
CC=golang-dev
https://golang.org/cl/6294067
A comment map associates comments with AST nodes
and permits correct updating of the AST's comment
list when the AST is manipulated.
R=rsc
CC=golang-dev
https://golang.org/cl/6281044
This slipped in with the implementation of Getpid in CL 5909043.
I'd exclude that CL entirely but it is tangled up in the Win32finddata changes.
R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/6297065
Preserve old API by using correct struct in system call
and then copying the results, as we did for SetsockoptLinger.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6307065
This new error is the only API change in the current draft of
Go 1.0.2 CLs. I'd like to include the CL that introduced it,
because it replaces a mysterious 'internal error' with a
useful error message, but I don't want any API changes,
so unexport the error constant for now. It can be
re-exported for Go 1.1.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6294055
Fixes#3708.
The fix to allow 5{c,g,l} to compile under clang 3.1 broke cross
compilation on darwin using the Apple default compiler on 10.7.3.
This failure was introduced in 9b455eb64690.
This has been tested by cross compiling on darwin/amd64 to linux/arm using
* gcc version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00)
* clang version 3.1 (branches/release_31)
As well as on linux/arm using
* gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5)
* Ubuntu clang version 3.0-6ubuntu3 (tags/RELEASE_30/final) (based on LLVM 3.0)
* Debian clang version 3.1-4 (branches/release_31) (based on LLVM 3.1)
R=consalus, rsc
CC=golang-dev
https://golang.org/cl/6307058
The type declarations were being generated using
a range over a map, which meant that successive
runs produced different orders. This will make sure
successive runs produce the same files.
Fixes#3707.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6300062
If there are mutually recursive functions, there is a cycle in
the dependency graph, so the order is actually dependency order
among the strongly connected components: mutually recursive
functions get put into the same batch and analyzed together.
(Until now the entire package was put in one batch.)
The non-recursive case (single function, maybe with some
closures inside) will be able to be more precise about inputs
that escape only back to outputs, but that is not implemented yet.
R=ken2
CC=golang-dev, lvd
https://golang.org/cl/6304050
CL 4313064 fixed its test case but did not address a
general enough problem:
type T1 struct { F *T2 }
type T2 T1
type T3 T2
could still end up copying the definition of T1 for T2
before T1 was done being evaluated, or T3 before T2
was done.
In order to propagate the updates correctly,
record a copy of an incomplete type for re-execution
once the type is completed. Roll back CL 4313064.
Fixes#3709.
R=ken2
CC=golang-dev, lstoakes
https://golang.org/cl/6301059
This is part 1 of a 2 part changelist. Part 2 contains the mechanical
change to parse.go to compare atoms (ints) instead of strings.
The overall effect of the two changes are:
benchmark old ns/op new ns/op delta
BenchmarkParser 4462274 4058254 -9.05%
BenchmarkRawLevelTokenizer 913202 912917 -0.03%
BenchmarkLowLevelTokenizer 1268626 1267836 -0.06%
BenchmarkHighLevelTokenizer 1947305 1968944 +1.11%
R=rsc
CC=andybalholm, golang-dev, r
https://golang.org/cl/6305053
The reordering speedup in CL 6245068 changed the semantics
of %#v by delaying the clearing of some flags. Restore the old
semantics and add a test.
Fixes#3706.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6302048
CL 6250075 removed AI_MASK mask on all BSD variants,
however FreeBSD's AI_MASK does not include AI_V4MAPPED
and AI_ALL, and its libc is strict about the ai_flags.
This will fix the FreeBSD builder.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6305054
On netbsd/386, tv_sec is a 64-bit integer for both timeval and timespec.
Fix the time handling code so that it works correctly.
R=golang-dev, rsc, m4dh4tt3r
CC=golang-dev
https://golang.org/cl/6256056
Thanks to Dave Cheney for the magic words "comm page".
benchmark old ns/op new ns/op delta
BenchmarkNow 197 33 -83.05%
This should make profiling a little better on OS X.
The raw time saved is unlikely to matter: what likely matters
more is that it seems like OS X sends profiling signals on the
way out of system calls more often than it should; avoiding
the system call should increase the accuracy of cpu profiles.
The 386 version would be similar but needs to do different
math for CPU speeds less than 1 GHz. (Apparently Apple has
never shipped a 64-bit CPU with such a slow clock.)
R=golang-dev, bradfitz, dave, minux.ma, r
CC=golang-dev
https://golang.org/cl/6275056
amd64 was done in CL 6275056.
We don't attempt to handle machines with clock speeds
less than 1 GHz. Those will fall back to the system call.
benchmark old ns/op new ns/op delta
BenchmarkNow 364 38 -89.53%
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6307045
Using an int64 for a block size doesn't make
sense on 32bit platforms but extracts a performance
penalty dealing with double word quantities on Arm.
linux/arm
benchmark old ns/op new ns/op delta
BenchmarkGobDecode 155401600 144589300 -6.96%
BenchmarkGobEncode 72772220 62460940 -14.17%
BenchmarkGzip 5822632 2604797 -55.26%
BenchmarkGunzip 326321 151721 -53.51%
benchmark old MB/s new MB/s speedup
BenchmarkGobDecode 4.94 5.31 1.07x
BenchmarkGobEncode 10.55 12.29 1.16x
R=golang-dev, rsc, bradfitz
CC=golang-dev
https://golang.org/cl/6272047
The original implementation of closures created the
underlying top-level function during walk, which is fairly
late in the compilation process and caused ordering-based
complications due to earlier stages that had to be repeated
any number of times.
Create the underlying function during typecheck, much
earlier, so that later stages can be run just once.
The result is a simpler compilation sequence.
R=ken2
CC=golang-dev
https://golang.org/cl/6279049
although the comment says it uses libc's getenv, without NOPLAN9DEFINES
it actually uses p9getenv which strdups.
R=golang-dev, dave, rsc
CC=golang-dev
https://golang.org/cl/6285046
The recent shuffle in parsing formats exposed probably unintentional
behavior in time.Parse, namely that it was mostly ignoring ".99999"
in the format, producing the following behavior:
fmt.Println(time.Parse("03:04:05.999 MST", "12:00:00.888 PDT")) // error (.888 unexpected)
fmt.Println(time.Parse("03:04:05.999", "12:00:00")) // error (input too short)
fmt.Println(time.Parse("03:04:05.999 MST", "12:00:00 PDT")) // ok (extra bytes on input make it ok)
http://play.golang.org/p/ESJ1UYXzq2
API CHANGE:
This CL makes all three examples valid: ".999" can match an
empty string or else a fractional second with at most nine digits.
Fixes#3701.
R=r, r
CC=golang-dev
https://golang.org/cl/6267045
An attempt to profit from CL 6176043 (fix to superpolinomial
runtime of karatsuba multiplication) and determine a better
karatsuba threshold. The result indicates that 32 is still
a reasonable value. Left the threshold as is (== 32), but
made some minor changes to the calibrate code which are
worthwhile saving (use of existing benchmarking code for
better results, better use of package time).
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6260062
pipe2 is equivalent to pipe with flags set to 0.
However, pipe2 was only added recently. Using pipe
instead improves compatibility with NetBSD 5.
R=jsing
CC=golang-dev
https://golang.org/cl/6268045
The cleanup also makes it ~5% faster, but that's
not the point of this CL.
Optimizations can come in future CLs.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6286043
It's very unfortunate that the type of Data field of struct
RawSockaddr is [14]uint8 on Linux/ARM instead of [14]int8
on all the others.
btw, it should be [14]int8 according to my header files.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6275050
Move address info flags to per-platform files. This is needed to
enable cgo on NetBSD (and later OpenBSD), as some of the currently
used AI_* defines do not exist on these platforms.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6250075
Use perfect cuckoo hash, to avoid binary search.
Define Atom bits as offset+len in long string instead
of enumeration, to avoid string headers.
Before: 1909 string bytes + 6060 tables = 7969 total data
After: 1406 string bytes + 2048 tables = 3454 total data
benchmark old ns/op new ns/op delta
BenchmarkLookup 83878 64681 -22.89%
R=nigeltao, r
CC=golang-dev
https://golang.org/cl/6262051
Ceil to 4.81 from 20.6 ns/op
Floor to 4.37 from 13.5 ns/op
Trunc to 3.97 from 14.3 ns/op
Also changed three MOVSDs to MOVAPDs in log_amd64.s
R=rsc, golang-dev
CC=golang-dev
https://golang.org/cl/6262048
Currently walk() doesn't check for err == SkipDir when iterating
a directory list, but such promise is made in the docs for WalkFunc.
Fixes#3486.
R=rsc, r
CC=golang-dev
https://golang.org/cl/6257059
Now that gri has made go/parser 15% faster, I offer this
change to slow back down cmd/api ~proportionately, adding
FreeBSD to the go1-checked set of platforms.
Really we should have done this earlier. This will prevent us
from breaking FreeBSD compatibility accidentally in the
future.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6279044
To avoid goroutines during init, the nextItem function was a
clever workaround. Now that init goroutines are permitted,
restore the original, simpler design.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6282043
- only compute current line position if needed
(i.e., if a comment is present)
- added benchmark
benchmark old ns/op new ns/op delta
BenchmarkParse 10902990 9313330 -14.58%
benchmark old MB/s new MB/s speedup
BenchmarkParse 5.31 6.22 1.17x
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6270043
Saving the code in case we improve things enough that
it matters later, but at least right now it is not worth doing.
R=ken2
CC=golang-dev
https://golang.org/cl/6248071
The previous code was preparing arrays of entries that would be
filled if there was one entry every 128 bytes. Moving to a 4096
byte interval reduces the overhead per megabyte of address space
to 2kB from 64kB (on 64-bit systems).
The performance impact will be negative for very small MemProfileRate.
test/bench/garbage/tree2 -heapsize 800000000 (default memprofilerate)
Before: mprof 65993056 bytes (1664 bucketmem + 65991392 addrmem)
After: mprof 1989984 bytes (1680 bucketmem + 1988304 addrmem)
R=golang-dev, rsc
CC=golang-dev, remy
https://golang.org/cl/6257069
The previous heap profile format did not include buckets with
zero used bytes. Also add several missing MemStats fields in
debug mode.
R=golang-dev, rsc
CC=golang-dev, remy
https://golang.org/cl/6249068
Drop expecttaken function in favor of extra argument
to gbranch and bgen. Mark loop condition as likely to
be true, so that loops are generated inline.
The main benefit here is contiguous code when trying
to read the generated assembly. It has only minor effects
on the timing, and they mostly cancel the minor effects
that aligning function entry points had. One exception:
both changes made Fannkuch faster.
Compared to before CL 6244066 (before aligned functions)
benchmark old ns/op new ns/op delta
BenchmarkBinaryTree17 4222117400 4201958800 -0.48%
BenchmarkFannkuch11 3462631800 3215908600 -7.13%
BenchmarkGobDecode 20887622 20899164 +0.06%
BenchmarkGobEncode 9548772 9439083 -1.15%
BenchmarkGzip 151687 152060 +0.25%
BenchmarkGunzip 8742 8711 -0.35%
BenchmarkJSONEncode 62730560 62686700 -0.07%
BenchmarkJSONDecode 252569180 252368960 -0.08%
BenchmarkMandelbrot200 5267599 5252531 -0.29%
BenchmarkRevcomp25M 980813500 985248400 +0.45%
BenchmarkTemplate 361259100 357414680 -1.06%
Compared to tip (aligned functions):
benchmark old ns/op new ns/op delta
BenchmarkBinaryTree17 4140739800 4201958800 +1.48%
BenchmarkFannkuch11 3259914400 3215908600 -1.35%
BenchmarkGobDecode 20620222 20899164 +1.35%
BenchmarkGobEncode 9384886 9439083 +0.58%
BenchmarkGzip 150333 152060 +1.15%
BenchmarkGunzip 8741 8711 -0.34%
BenchmarkJSONEncode 65210990 62686700 -3.87%
BenchmarkJSONDecode 249394860 252368960 +1.19%
BenchmarkMandelbrot200 5273394 5252531 -0.40%
BenchmarkRevcomp25M 996013800 985248400 -1.08%
BenchmarkTemplate 360620840 357414680 -0.89%
R=ken2
CC=golang-dev
https://golang.org/cl/6245069
On 6l and 8l, this is a real instruction, guaranteed to
cause an 'undefined instruction' exception.
On 5l, we simulate it as BL to address 0.
The plan is to use it as a signal to the linker that this
point in the instruction stream cannot be reached
(hence the changes to nofollow). This will help the
compiler explain that panicindex and friends do not
return without having to put a list of these functions
in the linker.
R=ken2
CC=golang-dev
https://golang.org/cl/6255064
16 seems pretty standard on x86 for function entry.
I don't know if ARM would benefit, so I used just 4
(single instruction alignment).
This has a minor absolute effect on the current timings.
The main hope is that it will make them more consistent from
run to run.
benchmark old ns/op new ns/op delta
BenchmarkBinaryTree17 4222117400 4140739800 -1.93%
BenchmarkFannkuch11 3462631800 3259914400 -5.85%
BenchmarkGobDecode 20887622 20620222 -1.28%
BenchmarkGobEncode 9548772 9384886 -1.72%
BenchmarkGzip 151687 150333 -0.89%
BenchmarkGunzip 8742 8741 -0.01%
BenchmarkJSONEncode 62730560 65210990 +3.95%
BenchmarkJSONDecode 252569180 249394860 -1.26%
BenchmarkMandelbrot200 5267599 5273394 +0.11%
BenchmarkRevcomp25M 980813500 996013800 +1.55%
BenchmarkTemplate 361259100 360620840 -0.18%
R=ken2
CC=golang-dev
https://golang.org/cl/6244066
The code was inconsistent about when it used
brchain(x) and when it used x directly, with the result
that you could end up emitting code for brchain(x) but
leave the jump pointing at an unemitted x.
R=ken2
CC=golang-dev
https://golang.org/cl/6250077
This bug has been introduced in the following revision:
changeset: 11404:26dceba5c610
user: Ivan Krasin <krasin@golang.org>
date: Mon Jan 23 09:19:39 2012 -0500
summary: compress/flate: reduce memory pressure at cost of additional arithmetic operation.
This is the review page for that CL: https://golang.org/cl/5555070/
R=rsc, imkrasin
CC=golang-dev
https://golang.org/cl/6249067
The correct procid is needed for unparking LWPs on NetBSD - always
initialise procid in minit() so that cgo works correctly. The non-cgo
case already works correctly since procid is initialised via
lwp_create().
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6257071
On NetBSD a cgo enabled binary has more than 32 sections - bump NSECTS
so that we can actually link them successfully.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6261052
filterPaeth takes []byte arguments instead of byte arguments,
which avoids some redudant computation of the previous pixel
in the inner loop.
Also eliminate a bounds check in decoding the up filter.
benchmark old ns/op new ns/op delta
BenchmarkDecodeGray 3139636 2812531 -10.42%
BenchmarkDecodeNRGBAGradient 12341520 10971680 -11.10%
BenchmarkDecodeNRGBAOpaque 10740780 9612455 -10.51%
BenchmarkDecodePaletted 1819535 1818913 -0.03%
BenchmarkDecodeRGB 8974695 8178070 -8.88%
R=rsc
CC=golang-dev
https://golang.org/cl/6243061
A block with finalizer might also be profiled. The special bit
is needed to unregister the block from the profile. It will be
unset only when the block is freed.
Fixes#3668.
R=golang-dev, rsc
CC=golang-dev, remy
https://golang.org/cl/6249066
The check for Stringer etc. can only fire if the test is not a builtin, so avoid
the expensive check if we know there's no chance.
Also put in a fast path for pad, which saves a more modest amount.
benchmark old ns/op new ns/op delta
BenchmarkSprintfEmpty 148 152 +2.70%
BenchmarkSprintfString 585 497 -15.04%
BenchmarkSprintfInt 441 396 -10.20%
BenchmarkSprintfIntInt 718 603 -16.02%
BenchmarkSprintfPrefixedInt 676 621 -8.14%
BenchmarkSprintfFloat 1003 953 -4.99%
BenchmarkManyArgs 2945 2312 -21.49%
BenchmarkScanInts 1704152 1734441 +1.78%
BenchmarkScanRecursiveInt 1837397 1828920 -0.46%
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6245068
This prevents clients from seeing RSTs and missing the response
body.
TCP stacks vary. The included test failed on Darwin before but
passed on Linux.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6256066
It was only being used for (*Stmt).Exec, not Query, and not for
the same two methods on *DB.
This unifies (*Stmt).Exec's old inline code into the old
subsetArgs function, renaming it in the process (changing the
old word "subset" to "driver", mostly converted earlier)
Fixes#3640
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6258045
It's sad to introduce a new macro, but rnd shows up consistently
in profiles, and the function call overwhelms the two arithmetic
instructions it performs.
R=r
CC=golang-dev
https://golang.org/cl/6260051
Plan 9 versions for amd64 have 2 megabyte pages.
This also fixes the logic for 32-bit vs 64-bit Plan 9,
making 64-bit the default, and adds logic to generate
a symbols table.
R=golang-dev, rsc, rminnich, ality, 0intro
CC=golang-dev, john
https://golang.org/cl/6218046
The old code generated for a bounds check was
CMP
JLT ok
CALL panicindex
ok:
...
The new code is (once the linker finishes with it):
CMP
JGE panic
...
panic:
CALL panicindex
which moves the calls out of line, putting more useful
code in each cache line. This matters especially in tight
loops, such as in Fannkuch. The benefit is more modest
elsewhere, but real.
From test/bench/go1, amd64:
benchmark old ns/op new ns/op delta
BenchmarkBinaryTree17 6096092000 6088808000 -0.12%
BenchmarkFannkuch11 6151404000 4020463000 -34.64%
BenchmarkGobDecode 28990050 28894630 -0.33%
BenchmarkGobEncode 12406310 12136730 -2.17%
BenchmarkGzip 179923 179903 -0.01%
BenchmarkGunzip 11219 11130 -0.79%
BenchmarkJSONEncode 86429350 86515900 +0.10%
BenchmarkJSONDecode 334593800 315728400 -5.64%
BenchmarkRevcomp25M 1219763000 1180767000 -3.20%
BenchmarkTemplate 492947600 483646800 -1.89%
And 386:
benchmark old ns/op new ns/op delta
BenchmarkBinaryTree17 6354902000 6243000000 -1.76%
BenchmarkFannkuch11 8043769000 7326965000 -8.91%
BenchmarkGobDecode 19010800 18941230 -0.37%
BenchmarkGobEncode 14077500 13792460 -2.02%
BenchmarkGzip 194087 193619 -0.24%
BenchmarkGunzip 12495 12457 -0.30%
BenchmarkJSONEncode 125636400 125451400 -0.15%
BenchmarkJSONDecode 696648600 685032800 -1.67%
BenchmarkRevcomp25M 2058088000 2052545000 -0.27%
BenchmarkTemplate 602140000 589876800 -2.04%
To implement this, two new instruction forms:
JLT target // same as always
JLT $0, target // branch expected not taken
JLT $1, target // branch expected taken
The linker could also emit the prediction prefixes, but it
does not: expected taken branches are reversed so that the
expected case is not taken (as in example above), and
the default expectaton for such a jump is not taken
already.
R=golang-dev, gri, r, dave
CC=golang-dev
https://golang.org/cl/6248049
Implement the (3-per-family) Noah's Ark clause (i.e. don't put
more than three identical elements on the list of active formatting
elements.
Also, when running tests, sort attributes by name before dumping
them.
Pass 4 additional tests with Noah's Ark clause (including one
that needs attributes to be sorted).
Pass 5 additional, unrelated tests because of sorting attributes.
R=nigeltao, rsc
CC=golang-dev
https://golang.org/cl/6247056
CanonicalHeaderKey didn't allocate, but it did use unnecessary
CPU in the hot path, deciding it didn't need to allocate.
I considered using constants for all these common header keys
but I didn't think it would be prettier. "Content-Length" looks
better than contentLength or hdrContentLength, etc.
R=golang-dev, dave
CC=golang-dev
https://golang.org/cl/6255053
Comment on cache keys above connectMethod says "http to proxy, http
anywhere after that", however in reality target address was always
included, which prevented http requests to different target
addresses to reuse the same http proxy connection.
R=golang-dev, r, rsc, bradfitz
CC=golang-dev
https://golang.org/cl/5901064
CL 5956051 introduced too many call != nil checks, so
attempt to improve this by splitting logic into three
distinct parts.
R=r
CC=golang-dev
https://golang.org/cl/6248048
for expr1, expr2 = range slice
was assigning to expr1 and expr2 in sequence
instead of in parallel. Now it assigns in parallel,
as it should. This matters for things like
for i, x[i] = range slice.
Fixes#3464.
R=ken2
CC=golang-dev
https://golang.org/cl/6252048
This is from CL 5451105 but was dropped from that CL.
See also CL 6137051.
The only change compared to 5451105 is to check for
h != nil in reflect·mapiterinit; allowing use of nil maps
must have happened after that original CL.
Fixes#3573.
R=golang-dev, dave, r
CC=golang-dev
https://golang.org/cl/6215078
Remove redundant checks for integration points.
Ignore null bytes in text.
Don't break out of foreign content for a <font> tag unless it
has a color, face, or size attribute.
Check for MathML text integration points when breaking out of
foreign content.
Pass two new tests.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6256045
The bulk of the gains come from hoisting the modulo ops outside of
the inner loop.
Reducing the digest type from 8 bytes to 4 bytes gains another 1% on
the hash/adler32 micro-benchmark.
Benchmarks for $GOOS,$GOARCH = linux,amd64 below.
hash/adler32 benchmark:
benchmark old ns/op new ns/op delta
BenchmarkAdler32KB 1660 1364 -17.83%
image/png benchmark:
benchmark old ns/op new ns/op delta
BenchmarkDecodeGray 2466909 2425539 -1.68%
BenchmarkDecodeNRGBAGradient 9884500 9751705 -1.34%
BenchmarkDecodeNRGBAOpaque 8511615 8379800 -1.55%
BenchmarkDecodePaletted 1366683 1330677 -2.63%
BenchmarkDecodeRGB 6987496 6884974 -1.47%
BenchmarkEncodePaletted 6292408 6040052 -4.01%
BenchmarkEncodeRGBOpaque 19780680 19178440 -3.04%
BenchmarkEncodeRGBA 80738600 79076800 -2.06%
Wall time for Denis Cheremisov's PNG-decoding program given in
https://groups.google.com/group/golang-nuts/browse_thread/thread/22aa8a05040fdd49
Before: 2.44s
After: 2.26s
Delta: -7%
R=rsc
CC=golang-dev
https://golang.org/cl/6251044
When client fails to write a request is sends caller that error,
however server might have failed to read that request in the mean
time and replied with that error. When client then reads the
response the call would no longer be pending, so call will be nil
Handle this gracefully by discarding such server responses
R=golang-dev, r
CC=golang-dev, rsc
https://golang.org/cl/5956051
* Eliminate bounds check on known small shifts.
* Rewrite x<<s | x>>(32-s) as a rotate (constant s).
* More aggressive (but still minimal) range analysis.
R=ken, dave, iant
CC=golang-dev
https://golang.org/cl/6209077
The previous attempt to explain this got it backwards (all the more reason to be
sad we couldn't make the two functions behave the same).
Fixes#3669.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6249051
There's no need for the 16-bit arithmetic here,
and it tickles a long-standing compiler bug.
Fix the exp code not to use 16-bit math and
create an explicit test for the compiler bug.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6256048
- interface methods appeared under VarDecl in search results
(long-standing TODO)
- don't walk parts of AST which contain no indexable material
(minor performance tuning)
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6228047
The documentation says so, but in the case of a normalized
integral Rat, the denominator was a new value. Changed the
internal representation to use an Int to represent the
denominator (with the sign ignored), so a reference to it
can always be returned.
Clarified documentation and added test cases.
Fixes#3521.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6237045
* Shift/rotate by constant doesn't have to stop subprop. (also in 8g)
* Remove redundant MOVLQZX instructions.
* An attempt at issuing loads early.
Good for 0.5% on a good day, might not be worth keeping.
Need to understand more about whether the x86
looks ahead to what loads might be coming up.
R=ken2, ken
CC=golang-dev
https://golang.org/cl/6203091
Detect HTML integration points and MathML text integration points.
At these points, process tokens as HTML, not as foreign content.
Pass 33 more tests.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6249044
Import updated test data from the WebKit Subversion repository (SVN revision 118111).
Some of the old tests were failing because we were HTML5 compliant, but the tests weren't.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6228049
I needed this to explore per-GOOS/GOARCH differences in pkg
syscall for a recent CL. Others may find it useful too.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6236046
- there is no label scope at package level
- open/close all scopes symmetrically now
that there is only one parse entry point
(parseFile)
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6230047
CL 5823055 removed a line introduced in Linux/ARM cgo support.
Because readsym() now returns nil for "$a", "$d" mapping symbols,
no matter the settings of `needSym', we still have to guard against
them in ldelf().
R=golang-dev, dave, rsc
CC=golang-dev
https://golang.org/cl/6220073
This quiets all.bash noise for upcoming features we know about.
The all.bash warnings will now only print for things not in next.txt
(or in next.txt but not in the API).
Once an API is frozen, we rename next.txt to a new frozen file
(like go1.txt)
Fixes#3651
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6218069
Handle text, comment, and doctype tokens in afterBodyIM, afterAfterBodyIM,
and afterAfterFramesetIM.
Pass three more tests.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6231043
Also: simplified some existing tests.
No support for Rats for now because the precision-preserving
default notation (fractions of the form a/b) is not a valid
JSON value.
Fixes#3657.
R=golang-dev, bradfitz, rsc
CC=golang-dev
https://golang.org/cl/6211079
Currently, if you pass some data to a template as an interface (e.g. interface{})
and extract that value that value as a parameter for a function, it fails, saying
wrong type.
This is because it is only looking at the interface type, not the interface content.
This CL uses the underlying content as the parameter to the func.
Fixes#3642.
R=golang-dev, r, r
CC=golang-dev
https://golang.org/cl/6218052
Introduce a newsym() to cmd/lib.c to add a symbol but don't add
them to hash table.
Introduce a new bit flag SHIDDEN and bit mask SMASK to handle hidden
and/or local symbols in ELF symbol tables. Though we still need to order
the symbol table entries correctly.
Fix for issue 3261 comment #9.
For CL 5822049.
R=iant, rsc
CC=golang-dev
https://golang.org/cl/5823055
ld -r could generate multiple section symbols for the same section,
but with different values, we have to take that into account.
Fixes#3322.
Part of issue 3261.
For CL 5822049.
R=golang-dev, iant, rsc, iant
CC=golang-dev
https://golang.org/cl/5823059
Now that we've fixed the Expect: test, this CL should be okay.
««« original CL description
net/http: revert 97d027b3aa68
Revert the following change set:
changeset: 13018:97d027b3aa68
user: Gustavo Niemeyer <gustavo@niemeyer.net>
date: Mon Apr 23 22:00:16 2012 -0300
summary: net/http: allow clients to disable keep-alive
This broke a test on Windows 64 and somebody else
will have to check.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6112054
»»»
Fixes#3540.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6228046
In both the web and command line tool,
the comment is shown after the declaration.
But in the code the comment is obviously before.
Make the text not refer to a specific order.
R=r, dsymonds
CC=golang-dev
https://golang.org/cl/6206094
The speedup is a combination of unrolling/specializing
the actual code and also making the compiler generate better code.
Go 1.0.1 (size: 1239 code + 320 data = 1559 total)
md5.BenchmarkHash1K 1000000 7178 ns/op 142.64 MB/s
md5.BenchmarkHash8K 200000 56834 ns/op 144.14 MB/s
Partial unroll (size: 1115 code + 256 data = 1371 total)
md5.BenchmarkHash1K 5000000 2513 ns/op 407.37 MB/s
md5.BenchmarkHash8K 500000 19406 ns/op 422.13 MB/s
Complete unroll (size: 1900 code + 0 data = 1900 code)
md5.BenchmarkHash1K 5000000 2442 ns/op 419.18 MB/s
md5.BenchmarkHash8K 500000 18957 ns/op 432.13 MB/s
Comparing Go 1.0.1 and the complete unroll (this CL):
benchmark old MB/s new MB/s speedup
md5.BenchmarkHash1K 142.64 419.18 2.94x
md5.BenchmarkHash8K 144.14 432.13 3.00x
On the same machine, 'openssl speed md5' reports 441 MB/s
and 531 MB/s for our two cases, so this CL is at 90% and 80% of
those speeds, which is at least in the right ballpark.
OpenSSL is using carefully engineered assembly, so we are
unlikely to catch up completely.
Measurements on a Mid-2010 MacPro5,1.
R=golang-dev, bradfitz, agl
CC=golang-dev
https://golang.org/cl/6220046
This just eliminates some duplication.
Also add a pointer to RFC 1122, in case
this comes up again.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6229044
Fail more usefully, and Logf in one place instead of Errorf where
an error is acceptable.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6221059
Comment groups must end at the end of a line (or the
next non-comment token) if the group started on a line
with non-comment tokens.
This is important for correct computation of "lead"
and "line" comments (Doc and Comment fields in AST nodes).
Without this fix, the "line" comment for F1 in the
following example:
type T struct {
F1 int // comment1
// comment2
F2 int
}
is "// comment1// comment2" rather than just "// comment1".
This bug was present from Day 1 but only visible when
looking at export-filtered ASTs where only comments
associated with AST nodes are printed, and only in rare
cases (e.g, in the case above, if F2 where not exported,
godoc would show "// comment2" anyway because it was
considered part of the "line" comment for F1).
The bug fix is very small (parser.go). The bulk of the
changes are additional test cases (parser_test.go).
The fix exposed a caching bug in go/printer via one of the
existing tests, hence the changes to printer.go.
As an aside, the fix removes the the need for empty lines
before an "// Output" comment for some special cases of
code examples (e.g.: src/pkg/strings/example_test.go, Count
example).
No impact on gofmt formatting of src, misc.
Fixes#3139.
R=rsc
CC=golang-dev
https://golang.org/cl/6209080
This includes the NIST test suite for ECDSA and alters the test to
parse and evaluate it.
R=golang-dev, bradfitz, rsc, b
CC=golang-dev
https://golang.org/cl/6219058
This fixes occasional 64-bit failures.
Maybe it will fix the 32-bit failures too,
so re-enable on 32-bit for now.
R=golang-dev, bradfitz, r, dvyukov
CC=golang-dev
https://golang.org/cl/6218050
Clean up flow of control.
Ignore </table>, </tbody>, </tfoot>, </thead>, </tr> if there is not
an appropriate element in table scope.
Pass 3 more tests.
R=golang-dev, nigeltao
CC=golang-dev
https://golang.org/cl/6206093
http://www.rsa.com/rsalabs/node.asp?id=2125:
NOTE: A new OID has been defined for the combination
of the v1.5 signature scheme and the SHA-224 hash function:
sha224WithRSAEncryption OBJECT IDENTIFIER ::=
Like the other sha*WithRSAEncryption OIDs in PKCS #1 v2.1,
this OID has NULL parameters.
The DigestInfo encoding for SHA-224 (see Section 9.2, Note 1) is:
(0x)30 2d 30 0d 06 09 60 86 48 01 65 03 04 02 04 05 00 04 1c || H
R=golang-dev, agl
CC=golang-dev
https://golang.org/cl/6208076
The main content of this CL is a test case checking the reported
issue 3511 and a tiny fix for it. A subsequent CL will refactor
the fix as proposed issue 3511.
Fixes#3511.
R=golang-dev, steven.hartland, bradfitz
CC=golang-dev
https://golang.org/cl/6013049
Two tests added in 820ffde8c are expected to fail until the fix
for Issue 3540 goes back in (pending Windows net fixes), so
make those tests just Logf for now, with a TODO to re-enable.
Add a new client test.
Rearrange the transport code to be more readable, and fix the
bug from 820ffde8c where the persistConn was being closed before
the body was fully ready.
Fixes#3644
Updates #1967 (not yet fixed, but should be after Issue 3540)
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/6211069
RawSockaddrDatalink and SockaddrDatalink need to match - make Data
have length 12 for both.
R=golang-dev, mikioh.mikioh
CC=golang-dev
https://golang.org/cl/6223051
Delete cases that just fall down to "anything else" action.
Handle </tbody>, </tfoot>, and </thead>.
R=golang-dev, nigeltao
CC=golang-dev
https://golang.org/cl/6203061
Connections did not close if Request.Close or Response.Close was true. This meant that if the user wanted the connection to close, or if the server requested it via "Connection: close", the connection would not be closed.
Fixes#1967.
R=golang-dev, rsc, bradfitz
CC=golang-dev
https://golang.org/cl/6201044
key and simple comparisson. Search is not yet implemented in this CL.
Changed some of the types of table_test.go to allow reuse in the new test.
Also reduced number of primary values for illegal runes to 1 (both map to
the same).
R=r
CC=golang-dev
https://golang.org/cl/6202062
This CL makes
type T struct { *U }
behave in a similar way to:
type T struct { U }
Fixes#3108.
R=golang-dev, rsc, gustavo
CC=golang-dev
https://golang.org/cl/5694044
With the timed semacquire patch
(kernel-tsemacquire) for Plan 9,
we can now properly do a timed
wait for the semaphore, in
semasleep.
R=golang-dev, rsc, rminnich, ality, r
CC=0intro, golang-dev, john, mirtchovski
https://golang.org/cl/6197046
Implement getcontext and sigprocmask for NetBSD - these will soon be
used by the thread handling code.
Also fix netbsd/386 signal handling - there is no sigreturn, just
return so that we hit the trampoline.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6215049
Before:
./x.go:6: first argument to append must be slice; have nil
After:
./x.go:6: first argument to append must be typed slice; have untyped nil
Fixes#3616.
R=ken2
CC=golang-dev
https://golang.org/cl/6209067
Update/correct NetBSD signal handling - most of this is needed due to
the correctly generated runtime definitions.
R=golang-dev, m4dh4tt3r, rsc
CC=golang-dev
https://golang.org/cl/6195079
Fix and regenerate runtime defs for NetBSD.
Whilst the mcontext struct can be handled across architectures,
the registers are provided as defines that index an array, rather
than as members of the struct. Since these are architecture
dependent, include them via a defs_netbsd_<arch>.go file.
R=golang-dev, m4dh4tt3r, rsc
CC=golang-dev
https://golang.org/cl/6190070
Parallel GC needs to know in advance how many helper threads will be there.
Hopefully it's the last patch before I can tackle parallel sweep phase.
The benchmarks are unaffected.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6200064
Set the TLS base using the _lwp_setprivate() syscall, instead of via
sysarch(). NetBSD tracks the pointer passed to _lwp_setprivate() and
restores this value when restoring mcontext. If sysarch() is used
directly, restoring an mcontext trashes the FS/GS value, resulting
in a segfault when we next try to access the TLS.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6206062
Empty parts can be either of the form:
a) "--separator\r\n", header (w/ trailing 2xCRLF), \r\n "--separator"...
or
b) "--separator\r\n", header (w/ trailing 2xCRLF), "--separator"...
We never handled case b). In fact the RFC seems kinda vague about
it, but browsers seem to do a), and App Engine's synthetic POST
bodies after blob uploads is of form b).
So handle them both, and add a bunch of tests.
(I can't promise these are the last fixes to multipart, especially
considering its history, but I'm growing increasingly confident at
least, and I've never submitted a multipart CL with known bugs
outstanding, including this time.)
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/6212046
Unbalanced extra right parenthesis produced an internal error instead of
a more descriptive one.
Fixes#3406.
R=r, rsc
CC=golang-dev
https://golang.org/cl/6201063
Fix mkerrors.sh so that it works on NetBSD.
Remove directory mode bits from types - this already appears in errors.
Regenerate the z* files now that cgo is working.
R=golang-dev, m4dh4tt3r, r
CC=golang-dev
https://golang.org/cl/6201077
No perf/semantic changes, merely improves code health.
There were several questions as to why Once.Do uses
atomic.CompareAndSwap to do a store.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6208057
we can't add the division result to n during iteration, because it might
turn n into NaN or Inf.
R=golang-dev, rsc, iant, iant
CC=golang-dev
https://golang.org/cl/6197045
Use correct syscall numbers and arguments for NetBSD.
Provide a trampoline for signal returns (using signal API 3).
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6209048
Manually repair/update z* files for netbsd - this allows Go to
compile again on 386.
R=golang-dev, rsc, minux.ma
CC=golang-dev
https://golang.org/cl/6194064
Also set maxContractLen automatically.
Note that the table size is much bigger than it needs to be.
Optimization is best done, though, when the language specific
tables are added.
R=r
CC=golang-dev
https://golang.org/cl/6167044
go help remote used to reference "go help importpath", which has
changed to "go help packages".
Fixes#3598.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6201065
Tested manually.
Fixes#3554.
Before:
$ cd $GOROOT/src/pkg
$ go list io
io
$ go list io/...
io
io/ioutil
$ cd $GOROOT/src/pkg/io
$ go list .
io
$ go list ./...
io/ioutil
After:
$ cd $GOROOT/src/pkg
$ go list io
io
$ go list io/...
io
io/ioutil
$ cd $GOROOT/src/pkg/io
$ go list .
io
$ go list ./...
io
io/ioutil
$ go list ././...
io
io/ioutil
$ go list ././.././io/...
io
io/ioutil
$ go list ../image
image
$ go list ../image/...
image
image/color
image/draw
image/gif
image/jpeg
image/png
$ go list ../.../template
html/template
text/template
$ cd $GOROOT/src/pkg
$ go list ./io
io
$ go list ./io/...
io
io/ioutil
$ go list ./.../pprof
net/http/pprof
runtime/pprof
$ go list ./compress
can't load package: package compress: no Go source files in /home/nigeltao/go/src/pkg/compress
$ go list ./compress/...
compress/bzip2
compress/flate
compress/gzip
compress/lzw
compress/zlib
$ cd $GOROOT/src/pkg/code.google.com
$ go list ./p/leveldb-go/...
code.google.com/p/leveldb-go/leveldb
code.google.com/p/leveldb-go/leveldb/crc
code.google.com/p/leveldb-go/leveldb/db
code.google.com/p/leveldb-go/leveldb/memdb
code.google.com/p/leveldb-go/leveldb/memfs
code.google.com/p/leveldb-go/leveldb/record
code.google.com/p/leveldb-go/leveldb/table
code.google.com/p/leveldb-go/manualtest/filelock
$ go list ./p/.../truetype
code.google.com/p/freetype-go/example/truetype
code.google.com/p/freetype-go/freetype/truetype
$ go list ./p/.../example
warning: "./p/.../example" matched no packages
$ go list ./p/.../example/...
code.google.com/p/freetype-go/example/freetype
code.google.com/p/freetype-go/example/gamma
code.google.com/p/freetype-go/example/raster
code.google.com/p/freetype-go/example/round
code.google.com/p/freetype-go/example/truetype
code.google.com/p/x-go-binding/example/imgview
code.google.com/p/x-go-binding/example/xgb
R=rsc
CC=golang-dev
https://golang.org/cl/6194056
These files change from exactly 10003 bytes long to 100003: a digit,
a '.', 100k digits, and a '\n'.
The magic constants in compress/flate/deflate_test.go change since
deflateInflateStringTests checks that the compressed form of e.txt
is not 'too large'. I'm not exactly sure how these numbers were
originally calculated (they were introduced in codereview 5554066
"make lazy matching work"); perhaps krasin@golang.org can comment.
My change was to increase the first one (no compression) to a tight
bound, and multiply all the others by 10.
Benchcmp numbers for compress/flate and compress/lzw below. LZW's
window size of 4096 is less than 10k, so shows no significant change.
Flate's window size is 32768, between 10k and 100k, and so the .*1e5
and .*1e6 benchmarks show a dramatic drop, since the compressed forms
are no longer a trivial forward copy of 10k digits repeated over and
over, but should now be more representative of real world usage.
compress/flate:
benchmark old MB/s new MB/s speedup
BenchmarkDecodeDigitsSpeed1e4 16.58 16.52 1.00x
BenchmarkDecodeDigitsSpeed1e5 68.09 18.10 0.27x
BenchmarkDecodeDigitsSpeed1e6 124.63 18.35 0.15x
BenchmarkDecodeDigitsDefault1e4 17.21 17.12 0.99x
BenchmarkDecodeDigitsDefault1e5 118.28 19.19 0.16x
BenchmarkDecodeDigitsDefault1e6 295.62 20.52 0.07x
BenchmarkDecodeDigitsCompress1e4 17.22 17.17 1.00x
BenchmarkDecodeDigitsCompress1e5 118.19 19.21 0.16x
BenchmarkDecodeDigitsCompress1e6 295.59 20.55 0.07x
BenchmarkEncodeDigitsSpeed1e4 8.18 8.19 1.00x
BenchmarkEncodeDigitsSpeed1e5 43.22 12.84 0.30x
BenchmarkEncodeDigitsSpeed1e6 80.76 13.48 0.17x
BenchmarkEncodeDigitsDefault1e4 6.29 6.19 0.98x
BenchmarkEncodeDigitsDefault1e5 31.63 3.60 0.11x
BenchmarkEncodeDigitsDefault1e6 52.97 3.24 0.06x
BenchmarkEncodeDigitsCompress1e4 6.20 6.19 1.00x
BenchmarkEncodeDigitsCompress1e5 31.59 3.59 0.11x
BenchmarkEncodeDigitsCompress1e6 53.18 3.25 0.06x
compress/lzw:
benchmark old MB/s new MB/s speedup
BenchmarkDecoder1e4 21.99 22.09 1.00x
BenchmarkDecoder1e5 22.77 22.71 1.00x
BenchmarkDecoder1e6 22.90 22.90 1.00x
BenchmarkEncoder1e4 21.04 21.19 1.01x
BenchmarkEncoder1e5 22.06 22.06 1.00x
BenchmarkEncoder1e6 22.16 22.28 1.01x
R=rsc
CC=golang-dev, krasin
https://golang.org/cl/6207043
I'm not sure where the BOM came from, originally.
http://www.gutenberg.org/files/74/74.txt doesn't have it, although
a fresh download of that URL gives me "\r\n"s instead of plain "\n"s,
and the extra line "Character set encoding: ASCII". Maybe Project
Gutenberg has changed their server configuration since we added that
file to the Go repo.
Anyway, this change is just manually excising the BOM from the start
of the file, leaving pure ASCII.
R=r, bradfitz
CC=golang-dev, krasin, rsc
https://golang.org/cl/6197061
dictates a CJK rune is only part of a certain specified range if it
is explicitly defined in the Unicode Codepoint Database.
Fixed the code and some of the tests accordingly.
R=r
CC=golang-dev
https://golang.org/cl/6160044
This is a follow-up to CL 5978051.
Use kernel cas64 helper if we can, fallback to LDREXD/STREXD if
we are on ARMv6 or higher, and to lock-emulated cas64 if on ARMv5.
A future CL will fix {Add,Load,Store}{Int,Uint}64 and issue 3331.
R=bradfitz, rsc
CC=golang-dev
https://golang.org/cl/6034048
Decode AT_RANDOM, AT_HWCAP, and AT_PLATFORM.
This CL only make use of AT_RANDOM, but future CLs will make use of the others.
R=dave, rsc
CC=golang-dev
https://golang.org/cl/5978051
Some newer Linux distributions (Ubuntu ARM at least) use a new multiarch
directory organization, where dynamic linker is no longer in the hardcoded
path in our linker.
For example, Ubuntu 12.04 ARM hardfloat places its dynamic linker at
/lib/arm-linux-gnueabihf/ld-linux.so.3
Ref: http://lackof.org/taggart/hacking/multiarch/
Also, to support Debian GNU/kFreeBSD as a FreeBSD variant, we need this capability, so it's part of issue 3533.
This CL add a new pragma (#pragma dynlinker "path") to cc.
R=iant, rsc
CC=golang-dev
https://golang.org/cl/6086043
1. In CL 5989057, I made a mistake in the last minute change.
"MOVW.W R4, -4(SP)" should really be "MOVW.W R4, -4(R13)",
as 5l will rewrite offset for SP.
2. misc/cgo/test/issue1560.go tests for parallel sleep of 1s,
but on ARM, the deadline is frequently missed, so change sleep
time to 2s on ARM.
R=golang-dev, dave, rsc
CC=golang-dev
https://golang.org/cl/6202043
This adds proper note handling for Plan 9,
and fixes the issue of properly killing go procs.
Without this change, the first go proc that dies
(using runtime·exit()) would kill all the running
go procs. Proper signal handling is needed.
R=golang-dev, ality, rminnich, rsc
CC=golang-dev, john, mirtchovski
https://golang.org/cl/5617048
Instead use a new type, "Note", whose underlying
type is just a string. This change allows us to
remove the exported os.Plan9Note type.
R=bradfitz, seed, rsc
CC=golang-dev
https://golang.org/cl/6015046
Part 1 of CL 5601044 (cgo: Linux/ARM support)
Limitation: doesn't support thumb library yet.
R=golang-dev, dave, rsc
CC=golang-dev
https://golang.org/cl/5991065
crypto.Hash.New() changed to panicking when the hash function isn't
linked in, but crypto/x509 still expects it to return nil.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6175047
This makes it possible to inline the prefetch of upcoming
memory addresses during garbage collection, instead of
needing to flush registers, make a function call, and
reload registers. On garbage collection-heavy workloads,
this results in a 5% speedup.
Fixes#3493.
R=dvyukov, ken, r, dave
CC=golang-dev
https://golang.org/cl/5990066
The first bug was that tertiary ignorables had the same colElem as
implicit colElems, yielding unexpected results. The current encoding
ensures that a non-implicit colElem is never 0. This fix uncovered
another bug of the trie that indexed incorrectly into the null block.
This was caused by an unfinished optimization that would avoid the
need to max out the most-significant bits of continuation bytes.
This bug was also present in the trie used in exp/norm and has been
fixed there as well. The appearence of the bug was rare, as the lower
blocks happened to be nearly nil.
R=r
CC=golang-dev
https://golang.org/cl/6127070
In CL 6127051, nigeltao suggested that further gains
were possible by improving the performance of flate.
This CL adds a set of benchmarks (based on compress/lzw)
that can be used to judge any future improvements.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6128049
This closes any internal descriptors (pipes, etc) that Cmd.Start() had
opened before it failed.
Fixes#3468.
R=golang-dev, iant, bradfitz
CC=golang-dev
https://golang.org/cl/5986044
In resolving 3507, the fix had to be applied individually to
the four *Conn types, tcp, udp, rawip and unix, due to the
duplicate code in each Conn type.
This CL consolidates the common net.Conn methods that all four
*Conn types implement into a base conn type.
Pros:
* The fix for 3507 would have only needed to be applied to one
method. Further improvements, such as possibly removing the
c.fd != nil check in c.ok(), would benefit from this CL.
* Nearly 300 lines removed from the net package.
* The public interface and documentation are not changed.
* I think this is an excellent example of the power of embedding.
Cons:
* The net package is already distributed over many files, this
CL adds another place to look.
* The fix for 3507 was a total of 16 lines changed, this follow
up CL could be considered to be an overreaction as new Conn types
are unlikely to be added in the near future.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6098047
The new logic is "use go1 if it's there, otherwise no tag."
Nothing needs to say "I require go1.0.1", and I want to
preserve some flexibility in defining what tags mean.
Right now (before go1.0.1) there is only one possible tag,
"go1", and I'd like to keep it that way.
R=golang-dev, bradfitz, r, adg
CC=golang-dev
https://golang.org/cl/6112060
Now that Go will no longer work on OpenBSD versions prior to 5.1,
remove the sysctl workaround that was needed for 5.0 and earlier.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6116050
1. Readdirnames was erroneously returning an
empty slice on every invocation.
2. The logic for determining which files to
close before exec was incorrect. If the
set of files to be kept open (provided by
the caller) did not include the files
opened at startup, those files would be
accidentally closed.
I also cleaned up readdupdevice while I was
in the vicinity.
R=golang-dev, seed, rsc
CC=golang-dev
https://golang.org/cl/6016044
Switch from using the rfork() syscall on OpenBSD, to the __tfork()
syscall. The __tfork() syscall is the preferred way of creating
system threads and the rfork() syscall has recently been removed.
Note: this will break compatibility with OpenBSD releases prior to 5.1.
R=golang-dev, bradfitz, devon.odell, rsc
CC=golang-dev
https://golang.org/cl/6037048
context for change lists of lower-level types. The public APIs are defined
in builder.go and collate.go. Type table is the glue between the lower and
higher level code and might be a good starting point for understanding the
collation code.
R=r, r
CC=golang-dev
https://golang.org/cl/5999053
The trie code looks a lot like the trie in exp/norm. It uses different
types, however. Also, there is only a lookup for []byte and the unsafe
lookup methods have been dropped, as well as sparse mode.
There is now a method for generating a trie. To output Go code, one now needs
to first generate a trie and then call print() on it.
R=r, r, mpvl
CC=golang-dev
https://golang.org/cl/5966064
Revert the following change set:
changeset: 13018:97d027b3aa68
user: Gustavo Niemeyer <gustavo@niemeyer.net>
date: Mon Apr 23 22:00:16 2012 -0300
summary: net/http: allow clients to disable keep-alive
This broke a test on Windows 64 and somebody else
will have to check.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6112054
On Solaris versions before Solaris 11, the kernel will not
return more than 1040 on a single read from /dev/urandom.
R=golang-dev, agl, bradfitz, rsc, iant, dchest
CC=golang-dev
https://golang.org/cl/6113046
As in:
const format = "%s"
fmt.Printf(format, "hi")
Also fix a couple of bugs by rewriting the routine.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6099057
Don't foster-parent text nodes that consist only of whitespace.
(I implemented this entirely in inTableIM instead of creating an
inTableTextIM, because the sole purpose of inTableTextIM seems to be
to combine character tokens into a string, which our tokenizer does
already.)
Use parseImpliedToken to clarify a couple of cases.
Handle <style>, <script>, <input>, and <form>.
Ignore doctype tokens.
Pass 20 additional tests.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6117048
This CL corrects the remaining differences that I could find between the
implementation of inBodyIM and the spec:
Handle <rp> and <rt>.
Adjust SVG and MathML attributes.
Reconstruct active formatting elements in the "any other start tag" case.
Pass 7 additional tests.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6101055
If a part ends with "--boundary--", without a final "\r\n",
that's also a graceful EOF, and we should return io.EOF instead
of the fmt-wrapped io.EOF from bufio.Reader.ReadSlice.
I found this bug parsing an image attachment from gmail.
Minimal test case stripped down from the original
gmail-generated attachment included.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6118043
When I increased the number of bits that gccgo uses for
untyped floats, the expression 0.52*1e9 was no longer
integral. This patch fixes that.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6113043
Moves the error detection back into execution, where it used to be,
and improves the error message.
Rolls back most of 6009048, which broke lower-case keys in maps.
If it weren't for maps we could detect this at compile time rather than
execution time.
Fixes#3542.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/6098051
Fixes#3507.
Applied the suggested fix from rsc. If the connection
is in closing state then errClosing will bubble up to
the caller.
The fix has been applied to udp, ip and unix as well as
their code path include nil'ing c.fd on close. Func
tests are available in the linked issue that verified
the bug existed there as well.
R=rsc, fullung, alex.brainman, mikioh.mikioh
CC=golang-dev
https://golang.org/cl/6002053
This lets the test pass on PPC64 GNU/Linux, which uses a much
larger page size and thus uses more memory to hold blocks
allocated for memory profiling.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6048054
The reverseproxy test depended on the behavior of
runtime.NumGoroutines(), which makes no guarantee when
goroutines are reaped. Instead, modify the flushLoop()
to invoke a callback when it returns, so the exit
from the loop can be tested, instead of the number
of gorountines running.
R=bradfitz
CC=golang-dev
https://golang.org/cl/6068046
Clean up the flow of control.
Fix the TODO for handling <html> tags.
Add a case to ignore doctype declarations.
Pass one additional test.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6072047
syscall.Exit would originally kill only the calling
Go proc, leaving behind other procs in the
same group. This change makes syscall.Exit call
runtime·exit, which due to CL
https://golang.org/cl/5617048
will cleanly exit all the Go procs in the group.
R=golang-dev, rsc, rminnich, remyoudompheng, ality, john
CC=golang-dev, mirtchovski
https://golang.org/cl/6036051
This method will allow us to be explicit about what we're doing when
we insert an implied token, and avoid repeating the logic involved in
multiple places.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6060048
When FlushInterval is specified on ReverseProxy, the ResponseWriter is
wrapped with a maxLatencyWriter that periodically flushes in a
goroutine. That goroutine was not being cleaned up at the end of the
request. This resulted in a panic when Flush() was being called on a
ResponseWriter that was closed.
The code was updated to always send the done message to the flushLoop()
goroutine after copying the body. Futhermore, the code was refactored to
allow the test to verify the maxLatencyWriter behavior.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6033043
Add a case to ignore doctype tokens.
Clean up the flow of control to more clearly match the spec.
Pass one more test.
R=nigeltao
CC=golang-dev
https://golang.org/cl/6062047
Add case for doctype tokens (which are ignored).
This CL does not change the status of any tests.
R=golang-dev, nigeltao
CC=golang-dev
https://golang.org/cl/6061047
Don't require the home directory to exist. Just check
that it returns something.
Fixes#3531
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/6056045
Currently, the html package only runs a limited subset of the tests
in the testdata directory. This tends to limit development of the
parser to fixing the bug that causes the first failing test.
This CL gives it the ability to run all the tests and produce a
log showing the status of each test. (It does it when tests are run with
'go test --update-logs') The status is listed as PASS, FAIL, or PARSE
(PARSE means that parsing produced the correct tree, but rendering and
re-parsing does not produce the same tree).
When 'go test' is run without --update-logs, it runs the tests marked
'PASS' in the logs (and the parsing portion of the tests marked 'PARSE').
Thus it will fail if there has been a regression since the last
time the logs were updated.
My goal for this CL is to allow develoment of the html package to
be less test-driven, while still having the advantages of regression
tests. In other words, one can work on any portion of the parser
and quickly see whether he is breaking things or improving them.
Current statistics of the tests:
$ grep ^PASS *.log|wc -l
1017
$ grep ^PARSE *.log|wc -l
46
$ grep ^FAIL *.log|wc -l
181
R=nigeltao
CC=golang-dev
https://golang.org/cl/6031049
This change comes from CL 5536043,
created by Andrey Mirtchovski. His
description follows:
"The plan9 exec child handler does not manage
dup-ed fds from the parent correctly: when a
dup-ed file descriptor appears in the child's fd
list it is closed when first encountered and then
subsequent attempt to dup it later in Pass 2 fails,
resulting in 'fork/exec: fd out of range or not
open'."
R=golang-dev, rminnich, ality
CC=golang-dev, mirtchovski, rsc
https://golang.org/cl/6009046
Update runtime defs for openbsd. Add struct __tfork, which will be
needed by an upcoming change.
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/6007050
Fixes#3495.
I adapted fmt.TestCountMallocs to fix the
existing tests. As the resulting tests did not
appear to belong to either itoa or ftoa I moved
them into their own file.
R=bradfitz, fullung
CC=golang-dev
https://golang.org/cl/5985072
When SNI based certificate selection is enabled, we previously used
the default private key even if we selected a non-default certificate.
Fixes#3367.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5987058
It's a common error to reference unexported field names in templates,
especially for newcomers. This catches the error at parse time rather than
execute time so the rare few who check errors will notice right away.
These were always an error, so the net behavior is unchanged.
Should break no existing code, just identify the error earlier.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/6009048
below do not support '.
This makes package html consistent with package text/template's
HTMLEscape function.
Fixes#3489.
R=rsc, mikesamuel, dsymonds
CC=golang-dev
https://golang.org/cl/5992071
crypto/tls is tested, in part, by replaying recorded TLS connections
and checking that the bytes sent by the Go code haven't changed.
Previously we used GnuTLS's debug output and extracted the bytes of
the TLS connection using a Python script. That wasn't great, and I
think GnuTLS removed that level of debugging in a more current
release.
This change records the connection with Go code and adds a test for
ECDHE-AES clients generating using this method.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5988048
For completeness, we also expose the Canonical Combining Class of a rune.
This does not increase the data size.
R=r
CC=golang-dev
https://golang.org/cl/5931043
Update the threxit and thrsleep syscalls to match the ABI of the
OpenBSD 5.1 kernel. These changes are backwards compatible with
older kernels.
Fixes#3311.
R=golang-dev, rsc, devon.odell
CC=golang-dev
https://golang.org/cl/5777079
This leads to ~30kB improvement on code size for ARM machines with VFP/NEON.
Example: go test -c math
GOARM=5 GOARM=6
Old: 1884200 1839144
New: 1884165 1805245
-: 35 33899
R=rsc, bradfitz, dave, kai.backman
CC=golang-dev
https://golang.org/cl/5975060
Change 5660047 moved an FLDCW instruction
that disables invalid operand traps into
runtime·asminit, which is called from
runtime·mstart. Thus, runtime·check is being
called prior to setting the appropriate control bits,
which on any QNaN comparison will cause Plan 9
to take an invalid operand trap. This change loads
the control bits (for Plan 9) prior to runtime·check.
Ideally, this should be done before the QNaN checks
on any system, but possibly other kernels simply
don't ever trap on invalid operands.
R=golang-dev, rminnich
CC=golang-dev, john, rsc
https://golang.org/cl/5939045
Without an explicit signal for a truncation, copy propagation
will sometimes propagate a 32-bit truncation and end up
overwriting uses of the original 64-bit value.
The case that arose in practice is in C but I believe
that the same could plausibly happen in Go.
The main reason we didn't run into the same in Go
is that I (perhaps incorrectly?) drop MOVL AX, AX
during gins, so the truncation was never generated, so
it didn't confuse the optimizer.
Fixes#1315.
Fixes#3488.
R=ken2
CC=golang-dev
https://golang.org/cl/6002043
Assignment of a computed uint64 value to an
address derived with a function call was executing
the call after computing the value, which trashed
the value (held in registers).
long long *f(void) { return 0; }
void g(int x, int y) {
*f() = (long long)x | (long long)y<<32;
}
Before:
(x.c:3) TEXT g+0(SB),(gok(71))
...
(x.c:4) ORL AX,DX
(x.c:4) ORL CX,BX
(x.c:4) CALL ,f+0(SB)
(x.c:4) MOVL DX,(AX)
(x.c:4) MOVL BX,4(AX)
After:
(x.c:3) TEXT g+0(SB),(gok(71))
(x.c:4) CALL ,f+0(SB)
...
(x.c:4) ORL CX,BX
(x.c:4) ORL DX,BP
(x.c:4) MOVL BX,(AX)
(x.c:4) MOVL BP,4(AX)
Fixes#3501.
R=ken2
CC=golang-dev
https://golang.org/cl/5998043
Block signals during thread creation, otherwise the new thread can
receive a signal prior to initialisation completing.
Fixes#3102.
R=golang-dev, rsc, devon.odell, minux.ma
CC=golang-dev
https://golang.org/cl/5757064
It is a bug that Caller and Callers disagree about the offset of the skip
parameter. Document the bug.
R=rsc, dsymonds, r, iant
CC=golang-dev
https://golang.org/cl/5976064
On windows Mercurial installed with easy_install typically creates
an hg.bat batch file in Python Scripts directory, which cannot be used
with CreateProcess unless full path is specified. Work around by
launching hg via cmd.exe /c.
Additionally, fix a rare FormatMessageW crash.
Fixes#3093.
R=golang-dev, rsc, alex.brainman, aram, jdpoirier, mattn.jp
CC=golang-dev
https://golang.org/cl/5937043
I have a small web server that simply sets several cookies
along with an expires header, and then returns. In the
cpuprofile for a 200k request benchmark, time.Time.Format()
was showing up as 8.3% of cpu usage. Giving the buffer an
inital capacity to avoid reallocs on append drops it down to
7.6%.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5992058
Building go currently fails on NetBSD, thus the regeneration
scripts do not run correctly. With these changes, at least
the go_bootstrap tool builds correctly.
R=golang-dev, r, bradfitz
CC=golang-dev
https://golang.org/cl/5695064
The shouldEscape function did not correctly escape the reserved characters listed in RFC 3986 §2.2, breaking some strict web servers.
Fixes#3433.
R=rsc
CC=golang-dev
https://golang.org/cl/5970050
Tested using 6g and gccgo on x86_64 GNU/Linux and using gccgo
on PowerPC GNU/Linux (which is big-endian).
R=golang-dev, bradfitz, mikioh.mikioh, iant
CC=golang-dev
https://golang.org/cl/5975073
Previously we checked that de ≡ 1 mod φ(n). Since φ(n) is a multiple
of |(ℤ/nℤ)*|, this encompassed the new check, but it was too strict as
keys generated by GnuTLS would be rejected when gcd(p-1,q-1)≠1.
(Also updated the error strings in crypto/rsa to contain the package name, which some were missing.)
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5867043
Before, "go get -v foo/bar" was assuming "foo" was a hostname
and trying to perform discovery on it. Now, require a dot in
the first path component (the hostname).
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5981057
Special case for encoding 4 zeros as 'z' didn't
update source slice, causing 'index out of bounds'
panic in destination slice.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5970078
For mysterious reasons, the existing string was just wrong: it was missing a colon.
There is no apparent reason for this discrepancy.
This should be safe to fix because existing uses would not be RFC822-compliant;
people cannot be depending on it to generate correct mail headers.
Fixes#3444.
R=golang-dev, dsymonds, iant, rsc
CC=golang-dev
https://golang.org/cl/5969072
Without this fix, an erroneous template causes a panic; should be caught safely.
The bug did not affect correct templates.
Fixes#3267.
R=golang-dev, dsymonds, rsc
CC=golang-dev
https://golang.org/cl/5900065
"go1" dominates. Delete the text about weekly and release.
We can revisit this once the situation changes.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5969043
Not a complete fix for issue 3342, but fixes the trivial case.
There may still be a race in the instants before and after
a scavenger-induced garbage collection.
Intended to be "obviously safe": a call to runtime·gosched
before main.main is no different than a call to runtime.Gosched
at the beginning of main.main, and it is (or had better be)
safe to call runtime.Gosched at any point during main.
Update #3342.
R=iant
CC=golang-dev
https://golang.org/cl/5919052
Another attempt at https://golang.org/cl/5754088.
Before, we only consulted $GOBIN for source code
found in $GOROOT, but that's confusing to explain
and less useful. The new behavior lets users set
GOBIN=$HOME/bin and have all go-compiled binaries
installed there.
Tested a few cases in test.bash.
Ran all.bash with and without $GOBIN and it works.
Even so, I expect it to break the builders,
like it did last time, we can debug from there.
Fixes#3269 (again).
Fixes#3396.
Fixes#3397.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5927051
When we find a package in DIR/src/foo, we only let it
be known as foo if there is no other foo in an earlier
GOPATH directory or the GOROOT directory.
The GOROOT check was looking in GOROOT/src/foo
instead of GOROOT/src/pkg/foo, which meant that
the import paths "lib9", "libbio", "libmach", and so
on were unavailable, and the import paths "math",
"errors", and so on were available. Correct this.
Fixes#3390.
R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/5927050
Issue 3207 was caused by setting GOPATH=GOROOT.
This is a common mistake, so diagnose it at command start
and also correct the bug that it caused in get (downloading
to GOROOT/src/foo instead of GOROOT/src/pkg/foo).
Issue 3268 was caused by recognizing 'packages' that
had installed binaries but no source. This behavior is not
documented and causes trouble, so remove it. We can
revisit the concept of binary-only packages after Go 1.
Fixes#3207.
Fixes#3268.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5930044
Breaks closure test when GOMAXPROCS=2 or more.
««« original CL description
runtime: restore deadlock detection in the simplest case.
Fixes#3342.
R=iant, r, dave, rsc
CC=golang-dev, remy
https://golang.org/cl/5844051
»»»
R=rsc
CC=golang-dev
https://golang.org/cl/5924045
- Renamed ExportData -> FindGcExportData
and base it on an a bufio.Reader rather
than a filename so it can be used in
environments where object files are
stored elsewhere.
- Factor former GcImporter into GcImportData
and GcImport. Implementations with different
storage locations for object files can build
a customized GcImport using GcImportData.
This is pkg/exp only - no impact on Go 1.
R=golang-dev, lvd, rsc
CC=golang-dev
https://golang.org/cl/5574069
Otherwise we won't fail if something goes wrong.
This shell programming stuff is tricky.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/5905062
They could be tested but that requires more than seems wise right now.
Update #2648.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/5908054
The existing code that tried to prevent ECC ciphersuites from being
selected when there were no mutual curves still left |suite| set.
This lead to a panic on a nil pointer when there were no acceptable
ciphersuites at all.
Thanks to George Kadianakis for pointing it out.
R=golang-dev, r, bradfitz
CC=golang-dev
https://golang.org/cl/5857043
Let's tell the world that Go's sort is O(n log n).
Surely this is a feature we intend to keep.
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/5867045
In a browser with many open tabs, the tab titles become short
and uninformative because they all start with the same prefix
("Package ", "Directory ", etc.).
Permit use of shorter tab titles that start with the relevant
information first.
Fixes#3365.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5865056
They looked out of place in godoc.
Includes documenting sort stability.
Fixes#3356
R=golang-dev, gri, trolleriprofessorn
CC=golang-dev
https://golang.org/cl/5855044
This is because we disallow local import for non-local packages, if
GOROOT happens to be under one of GOPATH, then some tests will fail
to build.
Fixes#3337.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5852043
- use GO_GCFLAGS and GO_LDFLAGS if supplied
- build misc\dashboard\builder and misc\goplay
- run tests in test\bench\go1
- check api compatibility
R=golang-dev, r, kardianos, bradfitz
CC=golang-dev
https://golang.org/cl/5847063
The two optimizations for small structs and arrays
were missing the implicit cast from ideal bool.
Fixes#3351.
R=rsc, lvd
CC=golang-dev
https://golang.org/cl/5848062
If a CertificateRequest is received we have to reply with a
Certificate message, even if we don't have a certificate to offer.
Fixes#3339.
R=golang-dev, r, ality
CC=golang-dev
https://golang.org/cl/5845067
As discussed on golang-dev, windows will use
"\" as path separator. No escaping allowed.
R=golang-dev, r, mattn.jp, rsc, rogpeppe, bsiegert, r
CC=golang-dev
https://golang.org/cl/5825044
It is unprecedented to add tests to package syscall, especially
system-specific ones. Not a policy worth changing right before Go 1
is cut.
The sole existing test, passfd_test.go, contains the line
// +build linux darwin probablyfreebsd probablyopenbsd
which argues that this is not a subject to be undertaking likely.
Note that passfd_test.go also went in just now. It's the only test
in syscall.
Deleting for now, will reconsider after Go 1.
R=golang-dev, bradfitz, r, dsymonds
CC=golang-dev
https://golang.org/cl/5846063
Adds new file api/go1.txt, locking down the current API.
Any changes to the API will need to update that file.
run.bash (but not make.bash, or Windows) will check for
accidental API changes.
R=golang-dev, dsymonds, rsc
CC=golang-dev
https://golang.org/cl/5820070
They have been deleted from package unsafe.
Also delete their appearance in exp/types.
Fixes#3338.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/5847056
This makes the last error-reporting CL a bit less
aggressive. errPrintedOutput is a sentinel value
that should not be wrapped.
R=gri
CC=golang-dev
https://golang.org/cl/5845052
Fixes part of issue 3253.
We still need to support scattered relocations though.
R=golang-dev, bsiegert, rsc, iant
CC=golang-dev
https://golang.org/cl/5822050
There was a small window during program initialization
where a signal could come in before the handling mechanisms
were set up to handle it. Delay the signal-handler installation
until we're ready for the signals.
Fixes#3314.
R=golang-dev, dsymonds, mikioh.mikioh
CC=golang-dev
https://golang.org/cl/5833049
windows cmd.exe can't show utf-8 correctly basically.
chcp 65001 may make it show, but most people don't have fonts which can
show it.
R=golang-dev, rsc, adg, gri, r
CC=golang-dev
https://golang.org/cl/5820060
A "gccgoprefix" flag is added and used by the go tool,
to mirror the -fgo-prefix flag for gccgo, whose value
is required to know how to access functions from C.
Trying to export Go methods or unexported Go functions
will not work.
Also fix go test on "main" packages.
Updates #2313.
Fixes#3262.
R=mpimenov, rsc, iant
CC=golang-dev
https://golang.org/cl/5797046
Fixes#3324.
Robert suggested not reporting errors until the end of the output.
which I'd also like to do, but errPrintedOutput makes that a bigger
change than I want to do before Go 1. This change should at least
remove the confusion we had.
# Building packages and commands for linux/amd64.
runtime
errors
sync/atomic
unicode
unicode/utf8
math
sync
unicode/utf16
crypto/subtle
io
syscall
hash
crypto
crypto/md5
hash/crc32
crypto/cipher
crypto/hmac
crypto/sha1
go install unicode: copying /tmp/go-build816525784/unicode.a to /home/rsc/g/go/pkg/linux_amd64/unicode.a: short write
hash/adler32
container/list
container/ring
...
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5837054
This function uses 48-byte of precious non-split stack for every callback
function, and without this CL, it can easily overflow the non-split stack.
I encountered this when trying to enable misc/cgo/test on windows/amd64.
R=rsc
CC=golang-dev
https://golang.org/cl/5784075
A node spans multiple lines if the line difference
between start and end point is > 0 (rather than > 1).
Fixes some odd cases introduced by CL 5706055;
pointed out by dsymonds.
Added corresponding test case. The other change
in the .golden file reverts to the status before
the CL mentioned above and is correct.
gofmt -w src misc changes godoc.go back to where
it was before the CL mentioned above.
Fixes#3304.
R=dsymonds, rsc
CC=golang-dev
https://golang.org/cl/5820044
1) Poor error checking in variable declarations admitted
$x=2 or even $x%2.
2) Need white space or suitable termination character
after identifiers, so $x+2 doesn't parse, in case we want it
to mean something one day.
Number 2 in particular prevents mistakes that we will have
to honor later and so is necessary for Go 1.
Fixes#3270.
Fixes#3271.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5795073
Windows has paths like C:/Users/ADMIN~1. Also, it so happens
that go/parser allows ~ in import paths. So does the spec.
Fixes the build too.
R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/5777073
For URLs ending with /, the handler did not work, trying to append
".xml" to the path.
For instance, the "Share Memory by Communicating" returned the
following error:
open /Users/francisco.souza/lib/go/doc/codewalk/sharemem/.xml: no such file or directory
R=adg, minux.ma
CC=golang-dev
https://golang.org/cl/5797065
broke builders
««« original CL description
cmd/go: respect $GOBIN always
Before, we only consulted $GOBIN for source code
found in $GOROOT, but that's confusing to explain
and less useful. The new behavior lets users set
GOBIN=$HOME/bin and have all go-compiled binaries
installed there.
Fixes#3269.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5754088
»»»
TBR=bradfitz
CC=golang-dev
https://golang.org/cl/5794065
Before, we only consulted $GOBIN for source code
found in $GOROOT, but that's confusing to explain
and less useful. The new behavior lets users set
GOBIN=$HOME/bin and have all go-compiled binaries
installed there.
Fixes#3269.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5754088
It's the best we can do before Go 1.
For issue 3250; not a fix but at least less mysterious.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5797068
Undo CL 5783076 and apply correct fix.
The /doc hack is wrong. The code to handle this case was
already there and just needs a simple fix:
// We didn't find any directories containing Go files.
// If some directory returned successfully, use that.
- if len(all) == 0 && first != nil {
+ if !haveGo {
for _, d := range first {
haveName[d.Name()] = true
all = append(all, d)
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5783079
They were necessary to produce a canonical script
when we checked in build scripts, but now they're just
getting in the way.
Fixes#3279.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5796068
The generated syscall files for Windows are still breaking "go tool api"
(unknown function []byte); I'll look at fixing that separately.
Fixes#3285.
R=bradfitz
CC=golang-dev
https://golang.org/cl/5777062
1) The -D argument should always be a pseudo-import path,
like _/Users/rsc/foo/bar, never a standard import path,
because we want local imports to always resolve to pseudo-paths.
2) Disallow local imports in non-local packages. Otherwise
everything works but you get two copies of a package
(the real one and the "local" one) in your binary.
R=golang-dev, bradfitz, yiyu.jgl
CC=golang-dev
https://golang.org/cl/5787055
To make sure that there is no resource leak,
I suggest to fix the 'fakedb' driver such as it fails when any
Stmt is not closed.
First, add a check in fakeConn.Close().
Then, fix all missing Stmt.Close()/Rows.Close().
I am not sure that the strategy choose in fakeConn.Prepare/prepare* is ok.
The weak point in this patch is the change in Tx.Query:
- Tests pass without this change,
- I found it by manually analyzing the code,
- I just try to make Tx.Query look like DB.Query.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5759050
In a transaction, on a Stmt.Query error, it was possible for a
connection to be added to a db's freelist twice. Should use
the local releaseConn function instead.
Thanks to Gwenael Treguier for the failing test.
Also in this CL: propagate driver errors through releaseConn
into *DB.putConn, which conditionally ignores the freelist
addition if the driver signaled ErrBadConn, introduced in a
previous CL.
R=golang-dev, gary.burd
CC=golang-dev
https://golang.org/cl/5798049
We should check the CRC32s of files on EOF, even if there's no
data descriptor (in streamed files), as long as there's a non-zero
CRC32 in the file header / TOC.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5794045
We now always write the "optional" streaming data descriptor
signature, which turns out to be required for OS X.
Also, handle reading the data descriptor with or without the
signature, per the spec's recommendation. Fix data descriptor
reading bugs found in the process.
Fixes#3252
R=golang-dev, alex.brainman, nigeltao, rsc
CC=golang-dev
https://golang.org/cl/5787062
So that we don't duplicate knowledge about which OS/ARCH combination
supports cgo.
Also updated src/run.bash and src/sudo.bash to use 'go env'.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5792055
It may have to switch stacks, since we are calling
a DLL instead of a system call.
badcallback says where it is, because it is being called
on a Windows stack already.
R=golang-dev, alex.brainman
CC=golang-dev
https://golang.org/cl/5782060
Implement runtime·write, like on the other systems,
and also runtime·badcallback, in assembly to reduce
stack footprint.
TBR=golang-dev
CC=golang-dev
https://golang.org/cl/5785055
When a very low-level system call that should never fail
does fail, we call notok, which crashes the program.
Often, we are then left with only the program counter as
information about the crash, and it is in notok.
Instead, inline calls to notok (it is just one instruction
on most systems) so that the program counter will
tell us which system call is unhappy.
R=golang-dev, gri, minux.ma, bradfitz
CC=golang-dev
https://golang.org/cl/5792048
1. consistent usage section (go tool xxx)
2. reformat cmd/ld document with minor correction
document which -H flags are valid on which ld
document -d flag can't be used on Windows.
document -Hwindowsgui
R=golang-dev, r, rsc
CC=golang-dev
https://golang.org/cl/5782043
The test.bash file generates .so file using gcc, builds the executable
using the go tool and then run it with the $LD_LIBRARY_PATH variable
pointing to the directory where the .so file lives.
Fixes#2982.
R=rsc, remyoudompheng
CC=golang-dev
https://golang.org/cl/5788043
Before:
$ go run x.go
signal 11 (core dumped)
$
After:
$ go run x.go
runtime: cgo callback on thread not created by Go.
signal 11 (core dumped)
$
For issue 3068.
Not a fix, but as much of a fix as we can do before Go 1.
R=golang-dev, rogpeppe, gri
CC=golang-dev
https://golang.org/cl/5781047
doc: convert to use godoc built-in templates
tmpltohtml is gone, to avoid having a second copy of the code.
Instead, godoc -url /doc/go1.html will print the actual HTML
served for that URL. "make" will generate files named go1.rawhtml
etc, which can be fed through tidy.
It can be hard to tell from the codereview diffs, but all the
tmpl files have been renamed to be html files and then
have "Template": true added.
R=golang-dev, adg, r, gri
CC=golang-dev
https://golang.org/cl/5782046
Was missing recompilation of packages imported only
by external test packages (package foo_test), primarily
because Root was not set, so those packages looked like
they were from a different Go tree, so they were not
recompiled if they already existed.
Also clean things up so that only one call to computeStale
is needed.
Fixes#3238.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5786048
Factored the error synchronization code into two functions
syncStmt and syncDecl. Because they may return w/o advancing
the scanner, there is potential for endless loops across
multiple parse functions; typically caused by an incorrect
token list in these functions (e.g., adding token.ELSE to
syncStmt will cause the parser to go into an endless loop
for test/syntax/semi7.go without this mechanism). This would
indicate a compiler bug, exposed only in an error situation
for very specific source files. Added a mechanism to force
scanner advance if an endless loop is detected. As a result,
error recovery will be less good in those cases, but the parser
reported a source error already and at least doesn't get stuck.
R=rsc, rsc
CC=golang-dev
https://golang.org/cl/5784046
The spec is looser than the current implementation.
The spec edit was made in CL 4444050 (May 2011)
but I never implemented it.
Fixes#3244.
R=ken2
CC=golang-dev
https://golang.org/cl/5785049
.syso files are system objects copied directly
into the package archive.
Fixes#1552.
R=alex.brainman, iant, r, minux.ma, remyoudompheng
CC=golang-dev
https://golang.org/cl/5778043
gofmt reports now a single, accurate error for
the test case of issue 3106.
Also: Added test harness for general error
checking and two test cases for now.
Fixes#3106.
R=rsc, bradfitz
CC=golang-dev
https://golang.org/cl/5755062
I am not sure why RunTests and RunExamples are
exported, but I assume that because they are we
should not change the signature, so I added an
unexported global shared by Main and RunTests.
Fixes#3237.
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/5779043
If it didn't reach the limit, we can try extending the arena
before resorting to random memory mappings and praying for the
kernel to be kind.
Fixes#3173.
R=rsc, rsc
CC=golang-dev
https://golang.org/cl/5725045
Allows the Windows part of CL 5700087 to land.
I had build tested CL 5753060 (which allows
crypto/x509 to use cgo and io/ioutil), and
didn't spot any errors on Windows.
Turns out I was wrong.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5753065
This moves the various CA root fetchers from crypto/tls into crypto/x509.
The move was brought about by issue 2997. Windows doesn't ship with all
its root certificates, but will instead download them as-needed when using
CryptoAPI for certificate verification.
This CL changes crypto/x509 to verify a certificate using the system root
CAs when VerifyOptions.RootCAs == nil. On Windows, this verification is
now implemented using Windows's CryptoAPI. All other root fetchers are
unchanged, and still use Go's own verification code.
The CL also fixes the hostname matching logic in crypto/tls/tls.go, in
order to be able to test whether hostname mismatches are honored by the
Windows verification code.
The move to crypto/x509 also allows other packages to use the OS-provided
root certificates, instead of hiding them inside the crypto/tls package.
Fixes#2997.
R=agl, golang-dev, alex.brainman, rsc, mikkel
CC=golang-dev
https://golang.org/cl/5700087
The compiler must be changed with the Set method
so that the buildToolchain gets updated too.
Fixes#3231.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5768044
The old way to find a port was to listen :0 and then
look at what port it picked, close the listener, and then
immediately try to listen on that port.
On some Windows 7 machines that sequence fails at
the second listen, because the first one is still lingering
in the TCP/IP stack somewhere. (Ironically, most of these
are used in tests of a "second listen", which in this case
ends up being the third listen.)
Instead of this race, just return the listener from the
function, replacing usableLocalPort+Listen with
usableListenPort.
Fixes#3219.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5769045
Does not actually test so files.
««« original CL description
misc/cgo: re-enable testso
Also enabled it for darwin.
R=rsc
CC=golang-dev
https://golang.org/cl/5754063
»»»
R=golang-dev, minux.ma, r, f
CC=golang-dev
https://golang.org/cl/5756075
I don't know what's out there, but something
is answering to 127.0.71.111:80 on our builder,
so use a different port.
Also insert a check that the dial fails, which
would have diagnosed this problem.
Fixes#3016.
R=golang-dev, mikioh.mikioh, r
CC=golang-dev
https://golang.org/cl/5754062
I don't know enough about multicast.
Should this be disabled on all systems, not just Windows?
R=golang-dev
CC=golang-dev
https://golang.org/cl/5754060
$ go run
go run: no go files listed
$ go run ../../pkg/math/bits.go
go run: cannot run non-main package
$
Fixes#3168.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5755064
By default the all.bash tests must not ever announce
on an external address. It's not just an OS X issue.
R=golang-dev, mikioh.mikioh
CC=golang-dev
https://golang.org/cl/5753067
We need a compact, reasonably efficient IsPrint. That adds about 2K of data,
plus a modest amount of code, but now strconv is a near-leaf package.
R=r, bradfitz, adg, rsc, minux.ma
CC=golang-dev
https://golang.org/cl/5756050
In order to land 5700087 (which moves the knowledge of how to get the
root certificates for the system from crypto/tls to crypto/x509), we
need to relax the restrictions on crypto/x509. Afterwards, we can
probably tighten them up in crypto/tls.
R=golang-dev, rsc, krautz
CC=golang-dev
https://golang.org/cl/5753060
Either documentation or implementation
of go run's flags is wrong currently.
This change assumes the documentation
to be right.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5752054
This CL permits using arbitrary, non-VCS-qualified URLs as
aliases for fully VCS-qualified and/or well-known code hosting
sites.
Example 1) A VCS-qualified URL can now be shorter.
Before:
$ go get camlistore.org/r/p/camlistore.git/pkg/blobref
After:
$ go get camlistore.org/pkg/blobref
Example 2) A custom domain can be used as the import,
referencing a well-known code hosting site.
Before:
$ go get github.com/bradfitz/sonden
After:
$ go get bradfitz.com/pkg/sonden
The mechanism used is a <meta> tag in the HTML document
retrieved from fetching:
https://<import>?go-get=1 (preferred)
http://<import>?go-get=1 (fallback)
The meta tag should look like:
<meta name="go-import" content="import-alias-prefix vcs full-repo-root">
The full-repo-root must be a full URL root to a repository containing
a scheme and *not* containing a ".vcs" qualifier.
The vcs is one of "git", "hg", "svn", etc.
The import-alias-prefix must be a prefix or exact match of the
package being fetched with "go get".
If there are multiple meta tags, only the one with a prefix
matching the import path is used. It is an error if multiple
go-import values match the import prefix.
If the import-alias-prefix is not an exact match for the import,
another HTTP fetch is performed, at the declared root (which does
*not* need to be the domain's root).
For example, assuming that "camlistore.org/pkg/blobref" declares
in its HTML head:
<meta name="go-import" content="camlistore.org git https://camlistore.org/r/p/camlistore" />
... then:
$ go get camlistore.org/pkg/blobref
... looks at the following URLs:
https://camlistore.org/pkg/blobref?go-get=1http://camlistore.org/pkg/blobref?go-get=1https://camlistore.org/?go-get=1http://camlistore.org/?go-get=1
Ultimately it finds, at the root (camlistore.org/), the same go-import:
<meta name="go-import" content="camlistore.org git https://camlistore.org/r/p/camlistore" />
... and proceeds to trust it, checking out git //camlistore.org/r/p/camlistore at
the import path of "camlistore.org" on disk.
Fixes#3099
R=r, rsc, gary.burd, eikeon, untheoretic, n13m3y3r, rsc
CC=golang-dev
https://golang.org/cl/5660051
This exercises the Import function but more importantly
gives us a place to write down the policy for dependencies
within the Go tree. It also forces us to look at the dependencies,
which may lead to adjustments.
Surprises:
- go/doc imports text/template, for HTMLEscape (could fix)
- it is impossible to use math/big without fmt (unfixable)
- it is impossible to use crypto/rand without math/big (unfixable)
R=golang-dev, bradfitz, gri, r
CC=golang-dev
https://golang.org/cl/5732062
In the test, verify the copied constants are correct.
Also put the test into package utf16 rather than utf16_test;
the old location was probably due creating the test from
utf8, but the separation is not needed here.
R=golang-dev, bradfitz, rsc, rsc, r
CC=golang-dev
https://golang.org/cl/5752047
Also, tweak run.go to use no more than 2x the
number of CPUs, and only one on ARM.
53.85u 13.33s 53.69r ./run
50.68u 12.13s 18.85r go run run.go
Fixes#2833.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5754047
* Splits into three server tests.
- TestStreamConnServer for tcp, tcp4, tcp6 and unix networks
- TestSeqpacketConnServer for unixpacket networks
- TestDatagramPacketConnServer for udp, udp4, udp6 and unixgram networks
* Adds both PacketConn and Conn test clients to datagram packet conn tests.
* Fixes wildcard listen test cases on dual IP stack platform.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5701066
The dependency was there only to pull in two constants.
Now we define them locally and verify equality in the test.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5754046
GOROOT_FINAL is a build parameter that means "eventually
the Go tree will be installed here". Make the file name information
match that eventual location.
Fixes#3180.
R=ken, ken
CC=golang-dev
https://golang.org/cl/5742043
CL 3075041 says ARM is not little-endian, but my test suggests otherwise.
My test program is:
package main
import ("fmt"; "syscall"; "os")
func main() {
err := syscall.Fallocate(1, 1/*FALLOC_FL_KEEP_SIZE*/, 0, int64(40960));
fmt.Fprintln(os.Stderr, err)
}
Without this CL, ./test > testfile will show: file too large; and strace shows:
fallocate(1, 01, 0, 175921860444160) = -1 EFBIG (File too large)
With this CL, ./test > testfile will show: <nil>; and strace shows:
fallocate(1, 01, 0, 40960) = 0
Quoting rsc:
"[It turns out that] ARM syscall ABI requires 64-bit arguments to use an
(even, odd) register pair, not an (odd, even) pair. Switching to "big-endian"
worked because it ended up using the high 32-bits (always zero in the tests
we had) as the padding word, because the 64-bit argument was the last one,
and because we fill in zeros for the rest of the system call arguments, up to
six. So it happened to work."
I updated mksyscall_linux.pl to accommodate the register pair ABI requirement,
and removed all hand-tweaked syscall routines in favor of the auto-generated
ones. These including: Ftruncate, Truncate, Pread and Pwrite.
Some recent Linux/ARM distributions do not bundle kernel asm headers,
so instead we always get latest asm/unistd.h from git.kernel.org (just like
what we do for FreeBSD).
R=ken, r, rsc, r, dave, iant
CC=golang-dev
https://golang.org/cl/5726051
An X.509 path length constrains the number of certificate that may
follow in the chain. This is a little simplistic for a first pass as it
doesn't check self-signed certificates (which don't count towards the
length), but it's conservatively simplistic.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5727057
asn1 didn't have an omitempty tag, so the list of additional primes in
an RSA private key was serialised as an empty SEQUENCE, even for
version 1 structures. This tripped up external code that didn't handle
v2.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5729062
UTCTime only has a two digit date field and year values from 50 should
be 1950, not 2050.
R=golang-dev, r, rsc
CC=golang-dev
https://golang.org/cl/5729063
This CL changes the behavior of Dial and Listen API family.
Previous Dial and Listen allow a combo of "tcp6" and IPv4 or IPv6
IPv4-mapped address as its argument, but it also makes slightly
different behaviors between Linux and other platforms. This CL fixes
such differences across over platforms by tweaking IP-level socket
option IPV6_V6ONLY. Consequently new Dial and Listen API family will
reject arguments consists of "tcp6" and IPv4 or IPv6 IPv4-mapped
address.
This CL also adds a bit clarified unicast listener tests.
Fixes#2581.
R=rsc, minux.ma
CC=golang-dev
https://golang.org/cl/5677086
The motivation for this CL is to support $GOPATH well.
Since we already have a FileSystem interface, implement a
Plan 9-style name space. Bind each of the $GOPATH src
directories onto the $GOROOT src/pkg directory: now
everything is laid out exactly like a normal $GOROOT and
needs very little special case code.
The filter files are no longer used (by us), so I think they
can just be deleted. Similarly, the Mapping code and the
FileSystem interface were two different ways to accomplish
the same end, so delete the Mapping code.
Within the implementation, since FileSystem is defined to be
slash-separated, use package path consistently, leaving
path/filepath only for manipulating operating system paths.
I kept the -path flag, but I think it can be deleted too.
Fixes#2234.
Fixes#3046.
R=gri, r, r, rsc
CC=golang-dev
https://golang.org/cl/5711058
The current package comment doesn't mention varints and
protocol buffers. Also, the first sentence is incomprehensible
without further context as "fixed-size values" is undefined.
R=rsc
CC=golang-dev
https://golang.org/cl/5715048
Missed a case for variadic functions with too few arguments.
The code passes, and with the right error, but might as well record the test case.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5732050
The recent addition of automatic function invocation generated
some troublesome ambiguities. Restore the previous behavior
and compensate by providing a "call" builtin to make it easy to
do what the automatic invocation did, but in a clear and explicit
manner.
Fixes#3140.
At least for now.
R=golang-dev, dsymonds, r
CC=golang-dev
https://golang.org/cl/5720065
I tried before to make relative imports work by simply
invoking the compiler in the right directory, so that
an import of ./foo could be resolved by ./foo.a.
This required creating a separate tree of package binaries
that included the full path to the source directory, so that
/home/gopher/bar.go would be compiled in
tmpdir/work/local/home/gopher and perhaps find
a ./foo.a in that directory.
This model breaks on Windows because : appears in path
names but cannot be used in subdirectory names, and I
missed one or two places where it needed to be removed.
The model breaks more fundamentally when compiling
a test of a package that lives outside the Go path, because
we effectively use a ./ import in the generated testmain,
but there we want to be able to resolve the ./ import
of the test package to one directory and all the other ./
imports to a different directory. Piggybacking on the compiler's
current working directory is then no longer possible.
Instead, introduce a new compiler option -D prefix that
makes the compiler turn a ./ import into prefix+that,
so that import "./foo" with -D a/b/c turns into import
"a/b/c/foo". Then we can invent a package hierarchy
"_/" with subdirectories named for file system paths:
import "./foo" in the directory /home/gopher becomes
import "_/home/gopher/foo", and since that final path
is just an ordinary import now, all the ordinary processing
works, without special cases.
We will have to change the name of the hierarchy if we
ever decide to introduce a standard package with import
path "_", but that seems unlikely, and the detail is known
only in temporary packages that get thrown away at the
end of a build.
Fixes#3169.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5732045
Except for the tests, this is mostly deleting code:
- removed several exprListModes:
blankStart: easily done explicitly, and trailing blanks
are cleaned up by the trimmer post-pass
blankEnd: never used
commaSep: all exprLists calls had this set
- added test cases for multi-line returns
(for a later fix of issue 1207)
- no formatting changes
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5672062
Fixes a problem Rob is having with goprotobuf.
Cannot add a test because the same case is more broken
when using ./ imports. That still needs to be fixed,
and is one aspect of issue 3169.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5725043
Also remove useless "install" argument to pkgpath now that go/build
defines package install locations.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5714059
Build environments will often define stock LDFLAGS
that are not compatible with the gc ld, causing
non-obvious failures midway through the build.
R=golang-dev, rsc, minux.ma, rsc
CC=golang-dev
https://golang.org/cl/5724044
Also change Wait to say "exit" instead of "exit or stop".
I notice that Pid is not implemented on all systems.
Should we fix that?
Fixes#3138.
R=golang-dev, alex.brainman, r
CC=golang-dev
https://golang.org/cl/5710056
Presumably something about the very large go/build
doc comment breaks the build constraint parser in
cmd/dist. I don't feel like debugging C code right now,
so move it into its own file. If cmd/dist decides doc.go
is not part of the package, it will still build correctly.
R=golang-dev
TBR=golang-dev
CC=golang-dev
https://golang.org/cl/5722043
The docs look awkward - there is a paragraph
"For example:" with a plain text example,
and straight below it a real Example.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5717048
* Install tools into tool dir always
(Fixes issue 3049. Fixes issue 2868. Fixes issue 2925.)
* Make packages depend on compiler, linker (Fixes issue 3036.)
* Do not recompile packages across roots (Fixes issue 3149.)
* Allow use of binary-only packages (Fixes issue 2775.)
* Avoid duplicate cgo dependencies (Fixes issue 3001.)
* Show less in go get -x. (Fixes issue 2926.)
* Do not force repo root for existing checkout (Fixes issue 2969.)
* Show full syntax error list always (Fixes issue 2811.)
* Clean arguments before processing (Fixes issue 3034.)
* Add flags for compiler, linker arguments (Fixes issue 2996.)
* Pass flags in make.bash (Fixes issue 3091.)
* Unify build flags, defined in one place.
* Clean up error messages (Fixes issue 3075. Fixes issue 2923.)
* Support local import paths (Fixes issue 3118.)
* Allow top-level package outside $GOPATH (Fixes issue 3009.)
In addition to these fixes, all commands now take a list of
go files as a way to specify a single package, just as go build and
go run always have. This means you can:
go list -json x.go
go fix x.go
go vet x.go
go test x_test.go
Preliminary tests in test.bash.
Mainly testing things that the ordinary build does not.
I don't mind if the script doesn't run on Windows.
I expect that gccgo support is now broken, and I hope that
people will feel free to file issues and/or send CLs to fix it. :-)
R=golang-dev, dsymonds, r, rogpeppe
CC=golang-dev
https://golang.org/cl/5708054
This is an API change, but one I have been promising would
happen when it was clear what the go command needed.
This is basically a complete replacement of what used to be here.
build.Tree is gone.
build.DirInfo is expanded and now called build.Package.
build.FindTree is now build.Import(package, srcDir, build.FindOnly).
The returned *Package contains information that FindTree returned,
but applicable only to a single package.
build.ScanDir is now build.ImportDir.
build.FindTree+build.ScanDir is now build.Import.
The new Import API allows specifying the source directory,
in order to resolve local imports (import "./foo") and also allows
scanning of packages outside of $GOPATH. They will come back
with less information in the Package, but they will still work.
The old go/build API exposed both too much and too little.
This API is much closer to what the go command needs,
and it works well enough in the other places where it is
used. Path is gone, so it can no longer be misused. (Fixes issue 2749.)
This CL updates clients of go/build other than the go command.
The go command changes are in a separate CL, to be submitted
at the same time.
R=golang-dev, r, alex.brainman, adg
CC=golang-dev
https://golang.org/cl/5713043
go test -short # like in the build; no external stuff
go test # long tests, + external
go test -external=false # long tests, no external
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5696079
The two string comparison optimizations were
missing the implicit cast from ideal bool.
Fixes#3119.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5696071
The strikes against it are:
1. It does not take path boundaries into account.
2. It assumes that Windows==case-insensitive file system
and non-Windows==case-sensitive file system, neither of
which is always true.
3. Comparing ToLower against ToLower is not a correct
implementation of a case-insensitive string comparison.
4. If it returns true on Windows you still don't know how long
the matching prefix is in bytes, so you can't compute what
the suffix is.
R=golang-dev, r, dsymonds, r
CC=golang-dev
https://golang.org/cl/5712045
In -test.short mode, skip measuring the upper bound of time
sleeps. The API only guarantees minimum bounds on sleeps,
anyway, so this isn't a bug we're ignoring as much as it is
simply observing bad builder virtualization and/or loaded
machines.
We keep the test in full mode where developers will
presumably be running on a lightly-loaded, native, fast
machine.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5713044
They are broken and hard to make work.
They have never worked: if you import "/tmp/x"
from "/home/rsc/p.c" then the compiler rewrites
this into import "/home/rsc/tmp/x", which is
clearly wrong.
Also we just disallowed the : character in import
paths, so import "c:/foo" is already not allowed.
Finally, in order to support absolute paths well in
a build tool we'd have to provide a mechanism to
instruct the compiler to resolve absolute imports
by looking in some other tree (where the binaries live)
and provide a mapping from absolute path to location
in that tree. This CL avoids adding that complexity.
This is not part of the language spec (and should not be),
so no spec change is needed.
If we need to make them work later, we can.
R=ken2
CC=golang-dev
https://golang.org/cl/5712043
Fixes#2919 I believe. (gets as far as sending a CONNECT
request to my little dummy logging proxy that doesn't actually
support CONNECT now.) Untested with a real CONNECT-supporting
proxy, though.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5708055
This CL mostly deletes code.
Using existing position information is
just as good to determine if a new section
is needed; no need to track exact multi-
line information. Eliminates the need to
carry around a multiLine parameter with
practically every function.
Applied gofmt -w src misc resulting in only
a minor change to godoc.go. In return, a couple
of test cases are now formatted better.
Not Go1-required, but nice-to-have as it will
simplify fixes going forward.
R=rsc
CC=golang-dev
https://golang.org/cl/5706055
Don't try to print obviously corrupt slices or interfaces.
Doesn't actually solve 3047 or 2818, but seems a good idea anyway.
R=rsc, bsiegert
CC=golang-dev
https://golang.org/cl/5708061
We should use DialUnix or ListenPacket for unixgram networks
because Dial doesn't take a local UnixAddr.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5706043
Work around profiling kernel bug with signal masks.
Still broken on 64-bit Snow Leopard kernel,
but I think we can ignore that one and let people
upgrade to Lion.
Add new trivial tools addr2line and objdump to take
the place of the GNU tools of the same name, since
those are not installed on OS X.
Adapt pprof to invoke 'go tool addr2line' and
'go tool objdump' if the system tools do not exist.
Clean up disassembly of base register on amd64.
Fixes#2008.
R=golang-dev, bradfitz, mikioh.mikioh, r, iant
CC=golang-dev
https://golang.org/cl/5697066
If stdout and stderr are indeed the same file (not a tty), which is
often the case, fully-buffered stdout will make it harder to see
progresses, for example, ./make.bash 2>&1 | tee log
R=r, rsc
CC=golang-dev
https://golang.org/cl/5700070
When debugging ASTs, it's useful to also
see the comments on occasion. Usage:
gotype -ast -comments file.go
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5703043
I've elected to omit escaping the output of Marshalers for now.
I haven't thought through the implications of that;
I suspect that double escaping might be the undoing of that idea.
Fixes#3127.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5694098
This fixes the build of package net for GOOS=NetBSD.
Of course, a real implementation would be even better.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5693065
Thanks to dr.volker.dobler for tracking this down.
Filed a long-term issue (3142) which may eventually
resolve this problem w/o the need for a manual fix.
R=iant
CC=golang-dev
https://golang.org/cl/5698078
The main change is simple: Both the Directory and DirEntry
struct have an extra field 'HasPkg' indicating whether the
directory contains any package files. The remaining changes
are more comments and adjustments to the template files.
Fixes#3121.
R=golang-dev, bradfitz, sameer
CC=golang-dev
https://golang.org/cl/5699072
Import paths with spaces are now invalid.
The builders would've caught this if they were running
the long tests. I've removed the check for short tests
in this package since the current tests are fast enough
already.
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/5694082
Allows one to disable everything but the example being debugged.
This time for sure.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5700079
For grouped type declarations, go/doc introduces
fake individual declarations. Don't use the original
location of the "type" keyword because it will lead
to an overly large source code range for that fake
declaration, and thus an overly large selection shown
via godoc (e.g.: click on the AssignStmt link for:
http://golang.org/pkg/go/ast/#AssignStmt ).
Also: Don't create a fake declaration if not needed.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5694061
Otherwise
go list -f "{{if .Stale}}{{.ImportPath}}{{end}}" all
and similar commands can print pages of empty lines.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5696058
Also allow multiple invalid import statements in a
single file.
Fixes#3021. The changes to go/parser and the
language specifcation have already been committed.
R=rsc, gri
CC=golang-dev
https://golang.org/cl/5672084
We were not aligning the code size,
so read-only data, which follows in the same
segment, could be arbitrarily misaligned.
Fixes#2506.
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/5693055
Avoids global array buffer overflows if they are
indexed using some of the values between NTYPE
and NALLTYPE. It is entirely likely that not all of these
are necessary, but this is the C compiler and not worth
worrying much about. This change takes up only a
few more bytes of memory and makes the behavior
deterministic.
Fixes#3078.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5693052
Alternatively, we could expand the ewidth array
in [568]c/txt.c to have NALLTYPES elements and
give all types above NTYPE a width of -1.
I don't think it's worth it since TDOT and TOLD
are the only two type values above NTYPE that
are passed to typ:
$ /tmp/cctypes
cc/dcl.c:683: t->down = typ(TOLD, T);
cc/dcl.c:919: return typ(TDOT, T);
$
Fixes#3063.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5694047
Replaced /*-style comments with line comments;
there are two many *'s already in those lines.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5699051
Although Intel considers the three-argument form of IMUL to be a
variant of IMUL, I couldn't make 6l able to differentiate it without
huge changes, so I called it IMUL3.
R=rsc
CC=golang-dev
https://golang.org/cl/5686055
Makes it possible for client code to maintain its own profiles,
and also reduces the API surface by giving us a type that
models built-in profiles.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5684056
This also fixes MarshalIndent's example after the
recent formatting convention changes.
Fixes#2831.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5671062
Using reg as the flag word was unfortunate, since the
default value is not 0 but NREG (==16), which happens
to be the bit NOPTR now. Clear it.
If I say this will fix the build, it won't.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5690072
Not a Go 1 issue, but appeared to be fairly easy to fix.
- Note that a few existing test cases look slightly worse but
those cases were not representative for real code. All real
code looks better now.
- Manual move of the comment in go/scanner/example_test.go
before applying gofmt.
- gofmt -w $GOROOT/src $GOROOT/misc
Fixes#3062.
R=rsc
CC=golang-dev
https://golang.org/cl/5674093
Ignore signals while we are spawning a new thread. Previously, a
signal arriving just before runtime.minit setting up the signal
handler triggers a "double fault" in signal trampolining.
Fixes#3017.
R=rsc, mikioh.mikioh, minux.ma, adg
CC=golang-dev
https://golang.org/cl/5684060
dodata will convert to SNOPTRDATA if appropriate.
Should fix arm build (hope springs eternal).
TBR=golang-dev
CC=golang-dev
https://golang.org/cl/5687074
This is a manual undo of CL 5674098.
It does not implement the even less strict spec
that we just agreed on, but it gets us back where
we were at the last weekly.
R=ken2
CC=golang-dev
https://golang.org/cl/5683069
cc: add #pragma textflag to set it
runtime: mark mheap to go into noptr-bss.
remove special case in garbage collector
Remove the ARM from.flag field created by CL 5687044.
The DUPOK flag was already in p->reg, so keep using that.
Otherwise test/nilptr.go creates a very large binary.
Should fix the arm build.
Diagnosed by minux.ma; replacement for CL 5690044.
R=golang-dev, minux.ma, r
CC=golang-dev
https://golang.org/cl/5686060
Accept certain non-compliant response headers
(in particular, when spaces preceed the colon).
All major browser and curl seem to support this,
and at least one webserver seems to send these.
*shrug*
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/5690059
When installing pprof into the tools directory, it needs to
have execute permissions on unix-like systems.
Fixes issues 3077.
R=golang-dev, rsc, minux.ma
CC=golang-dev
https://golang.org/cl/5675095
Before we were using "ESMTP" in the banner as a clue,
but that is not required by the RFC and breaks mailing
to smtp.yandex.ru.
Fixes#3045.
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5687066
A fault during malloc might lead to the program's
first call to findfunc, which would in turn call malloc.
Don't do that.
Fixes#1777.
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/5689047