sigaction, sigprocmask, sigaltstack, and raiseproc.
Fix bug in mstart_stub where we weren't saving callee-saved registers,
so if an m finished the pthread library calling mstart_stub would
sometimes fail.
Update #17490
Update #22805
Change-Id: Ie297ede0997910aa956834e49e85711b90cdfaa7
Reviewed-on: https://go-review.googlesource.com/116875
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Previously, go/doc would only consider functions and slices that
return types of T or any number of pointers to T: *T, **T, etc. This
change expands the definition of a constructor to include functions
that return arrays of a type (or pointer to that type) in its first
return.
With this change, the following return types also classify a function
as a constructor of type T:
[1]T
[1]*T
[1]**T
(and so on)
Fixes#22856.
Change-Id: I37957c5f2d6a7b2ceeb3fbaef359057f2039393d
Reviewed-on: https://go-review.googlesource.com/85355
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Previously the Transport had good support for 100 Continue responses,
but other 1xx informational responses were returned as-is.
But per https://tools.ietf.org/html/rfc7231#section-6.2:
> A client MUST be able to parse one or more 1xx responses received
> prior to a final response, even if the client does not expect one. A
> user agent MAY ignore unexpected 1xx responses.
We weren't doing that. Instead, we were returning any 1xx that wasn't
100 as the final result.
With this change we instead loop over up to 5 (arbitrary) 1xx
responses until we find the final one, returning an error if there's
more than 5. The limit is just there to guard against malicious
servers and to have _some_ limit.
By default we ignore the 1xx responses, unless the user defines the
new httptrace.ClientTrace.Got1xxResponse hook, which is an expanded
version of the previous ClientTrace.Got100Continue.
Still remaining:
* httputil.ReverseProxy work. (From rfc7231#section-6.2: "A proxy MUST
forward 1xx responses unless the proxy itself requested the
generation of the 1xx response."). Which would require:
* Support for an http.Handler to generate 1xx informational responses.
Those can happen later. Fixing the Transport to be resilient to others
using 1xx in the future without negotiation (as is being discussed
with HTTP status 103) is most important for now.
Updates #17739
Change-Id: I55aae8cd978164643fccb9862cd60a230e430486
Reviewed-on: https://go-review.googlesource.com/116855
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
On dragonfly, freebsd and solaris the sendfile syscall does not update
the read position of the source fd. Update it after sendfile so
successive calls start at the correct position.
Fixes#25809
Change-Id: Iaac79f89704b75b8038d4bb60eaf793a262cdd8f
Reviewed-on: https://go-review.googlesource.com/117895
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This CL takes advantage of the ability to record vet-specific export data,
added in CL 108558, to save information about observed printf wrappers.
Then calls to those wrappers from other packages can be format-checked.
This found a few real mistakes using previously-unrecognized printf
wrappers in cmd/compile. It will no doubt find real mistakes in external code.
Change-Id: I9c29c92d89bbdc984571a174a96e6054585e9cd4
Reviewed-on: https://go-review.googlesource.com/108559
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
This CL makes it possible for vet to write down notes about one package
and then access those notes later, when analyzing other code importing
that package. This is much like what the compiler does with its own export
data for type-checking, so we call it "vet-export" data or vetx data.
The next CL in the stack makes vet actually use this functionality.
Change-Id: Ic70043ab407dfbfdb3f30eaea7c0e3c8197009cf
Reviewed-on: https://go-review.googlesource.com/108558
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Keep searching for a package that is both findable and importable. The
current code would always guarantee that a package was findable but
exited if it was not importable.
Fixes#25478
Change-Id: I237b7dfafb930cae02538c4a2e4d5ce0c1058478
Reviewed-on: https://go-review.googlesource.com/114295
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Rob Pike <r@golang.org>
Run-TryBot: Bryan C. Mills <bcmills@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
For memmove/memclr using jump tables only reduces overall
function performance for both amd64 and 386.
Benchmarks for 32-bit memclr:
name old time/op new time/op delta
Memclr/5-8 8.01ns ± 0% 8.94ns ± 2% +11.59% (p=0.000 n=9+9)
Memclr/16-8 9.05ns ± 0% 9.49ns ± 0% +4.81% (p=0.000 n=8+8)
Memclr/64-8 9.15ns ± 0% 9.49ns ± 0% +3.76% (p=0.000 n=9+10)
Memclr/256-8 16.6ns ± 0% 16.6ns ± 0% ~ (p=1.140 n=10+9)
Memclr/4096-8 179ns ± 0% 166ns ± 0% -7.26% (p=0.000 n=9+8)
Memclr/65536-8 3.36µs ± 1% 3.31µs ± 1% -1.48% (p=0.000 n=10+9)
Memclr/1M-8 59.5µs ± 3% 60.5µs ± 2% +1.67% (p=0.009 n=10+10)
Memclr/4M-8 239µs ± 3% 245µs ± 0% +2.49% (p=0.004 n=10+8)
Memclr/8M-8 618µs ± 2% 614µs ± 1% ~ (p=0.315 n=10+8)
Memclr/16M-8 1.49ms ± 2% 1.47ms ± 1% -1.11% (p=0.029 n=10+10)
Memclr/64M-8 7.06ms ± 1% 7.05ms ± 0% ~ (p=0.573 n=10+8)
[Geo mean] 3.36µs 3.39µs +1.14%
For less predictable data, like loop iteration dependant sizes,
branch table still shows 2-5% worse results.
It also makes code slightly more complicated.
This CL removes TODO note that directly suggest trying this
optimization out. That encourages people to spend their time
in a quite hopeless endeavour.
The code used to implement branch table used a 32/64-entry table
with pointers to TEXT blocks that implemented every associated
label work. Most last entries point to "loop" code that is
a fallthrough for all other sizes that do not map into specialized
routines. The only inefficiency is extra MOVL/MOVQ required
to fetch table pointer itself as MOVL $sym<>(SB)(AX*4) is not valid
in Go asm (it works in other assemblers):
TEXT ·memclrNew(SB), NOSPLIT, $0-8
MOVL ptr+0(FP), DI
MOVL n+4(FP), BX
// Handle 0 separately.
TESTL BX, BX
JEQ _0
LEAL -1(BX), CX // n-1
BSRL CX, CX
// AX or X0 zeroed inside every text block.
MOVL $memclrTable<>(SB), AX
JMP (AX)(CX*4)
_0:
RET
Change-Id: I4f706931b8127f85a8439b95834d5c2485a5d1bf
Reviewed-on: https://go-review.googlesource.com/115678
Run-TryBot: Iskander Sharipov <iskander.sharipov@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
CL 115975 changed TestGoExec to check symbol types.
However, this test is failing on Plan 9, because
there is no read-only data segment symbol on Plan 9.
This change fixes TestGoExec to replace the check
of read-only data segment symbol (R) by data segment
symbol (D) on Plan 9.
Fixes#25820.
Change-Id: I7164cd9056fa1dfcd1dc1b0f87653290c14c85fa
Reviewed-on: https://go-review.googlesource.com/118035
Run-TryBot: David du Colombier <0intro@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Currently .DepOnly is set when go list -test is invoked to help
distinguish those packages that matched the command line spec from those
which are dependencies (of test packages). This is also useful when
calling go list -deps for the same reason.
Change-Id: Ifc0e68dad0fd01355928793ef803691dee5f4f29
Reviewed-on: https://go-review.googlesource.com/112755
Run-TryBot: Paul Jolly <paul@myitcv.org.uk>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Russ Cox <rsc@golang.org>
An upcoming change to cmd/go will enable this functionality, which
allows vet to write down information about one package for use by
later invocation of vet that analyze code importing that package.
We've intended to do this for a long time, but the build caching was
necessary to have a decent way to manage the vet-specific export data.
This is also an experiment in building scalable whole-program analyses.
In the long term we'd like to allow other analyses to be invoked this way.
Change-Id: I34e4b70445786b2e8707ff6a0c00947bf1491511
Reviewed-on: https://go-review.googlesource.com/117099
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This adds the support to enable the race detector for ppc64le.
Added runtime/race_ppc64le.s to manage the calls from Go to the
LLVM tsan functions, mostly converting from the Go ABI to the
PPC64 ABI expected by Clang generated code.
Changed racewalk.go to call racefuncenterfp instead of racefuncenter
on ppc64le to allow the caller pc to be obtained in the asm code
before calling the tsan version.
Changed the set up code for racecallbackthunk so it doesn't use
the autogenerated save and restore of the link register since that
sequence uses registers inconsistent with the normal ppc64 ABI.
Made various changes to recognize that race is supported for
ppc64le.
Ensured that tls_g is updated and accessible from race_linux_ppc64le.s
so that the race ctx can be obtained and passed to tsan functions.
This enables the race tests for ppc64le in cmd/dist/test.go and
increases the timeout when running the benchmarks with the -race
option to avoid timing out.
Updates #24354, #23731
Change-Id: Ib97dc7ac313e6313c836dc7d2fb698f9d8fba3ef
Reviewed-on: https://go-review.googlesource.com/107935
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
"Syntax analysis" sounds more familiar and fits the
item before, which says "lexical analysis".
If there was specific intention to the original wording,
I, as a reader, would like to see it instead of this
confusing wording.
Change-Id: Id32dbf75300a86b21cb9f35e54526184fe5df6cb
Reviewed-on: https://go-review.googlesource.com/117696
Reviewed-by: Robert Griesemer <gri@golang.org>
It has been a long time since the last time the vendored zoneinfo in
lib/time was updated, and we're well into the freeze. Update it to the
lastest release from IANA.
Updates #22487
Change-Id: Ib9a8eb409554848285fc88363dbb04ed9d6d9eb0
Reviewed-on: https://go-review.googlesource.com/117855
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
When using plugins with goroutines calling cgo, we hit a case where
an intermittent SIGSEGV occurs when referencing an address that is based
on r2 (TOC address). When the failure can be generated in gdb, the
contents of r2 is wrong even though the value in the current stack's
slot for r2 is correct. So that means it somehow switched to start
running the code in this function without passing through the beginning
of the function which had the correct value of r2 and stored it there.
It was noted that in runtime.gogo when the state is restored from
gobuf, r2 is not restored from its slot on the stack. Adding the
instruction to restore r2 prevents the SIGSEGV.
This adds a testcase under testplugin which reproduces the problem if
the program is run multiple times. The team who reported this problem
has verified it fixes the issue on their larger, more complex
application.
Fixes#25756
Change-Id: I6028b6f1f8775d5c23f4ebb57ae273330a28eb8f
Reviewed-on: https://go-review.googlesource.com/117515
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Hardware AES support in Go on s390x currently requires ECB, CBC
and CTR modes be available. It also requires that either the
GHASH or GCM facilities are available. The existing checks missed
some of these constraints.
While we're here simplify the cpu package on s390x, moving masking
code out of assembly and into Go code. Also, update SHA-{1,256,512}
implementations to use the cpu package since that is now trivial.
Finally I also added a test for internal/cpu on s390x which loads
/proc/cpuinfo and checks it against the flags set by internal/cpu.
Updates #25822 for changes to vet whitelist.
Change-Id: Iac4183f571643209e027f730989c60a811c928eb
Reviewed-on: https://go-review.googlesource.com/114397
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Skip it like on freebsd until there is proper a fix for #25809
Updates #25809
Change-Id: Id53c433aee75f2a992ab6a8d58d98fd1f8a6c1c6
Reviewed-on: https://go-review.googlesource.com/117698
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Add test for freebsd issue #25809.
This test also fails on my Windows 10 Version 1803.
My hope is that adding new test will break one of our builders.
Updates #25722
Updates #25809
Change-Id: Ia103bc708b8fa3b9af57613acc44893f90b3fa18
Reviewed-on: https://go-review.googlesource.com/117775
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TOKEN_ALL_ACCESS was changed at some stage by Microsoft.
Updates #25775
Change-Id: I3e18914207a0020b2ebfb99f4e57aa55f9de813b
Reviewed-on: https://go-review.googlesource.com/117635
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Although not part of http://mimesniff.spec.whatwg.org,
for WASM streaming compilation to happen, the response
needs to have the application/wasm MIME type
as mentioned here:
https://webassembly.github.io/spec/web-api/index.html#streaming-modules.
And all current browsers prevent streaming compilation
from happening if this MIME type is not present in the response.
The magic number is mentioned here:
https://webassembly.org/docs/binary-encoding
Since we are already adding WASM support, it makes sense
to support this MIME type.
Change-Id: I8dd7b413a8c438a5c23c29d843b42f6da2a20ba4
Reviewed-on: https://go-review.googlesource.com/113396
Reviewed-by: Richard Musiol <neelance@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Users of this field are better off using Result instead.
Fixes#25763.
Change-Id: I4391afa6ed3873107628630adc1d409d77fb3f20
Reviewed-on: https://go-review.googlesource.com/117675
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The Solaris assembler uses a different syntax for section directives.
Fixes https://gcc.gnu.org/PR85429.
Change-Id: I1e54dffee3290046dbb68ba4e90ab795c6b72571
Reviewed-on: https://go-review.googlesource.com/109140
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This is extension to https://golang.org/cl/113955 that handled
duplicated "unresolved relocation" errors.
For platforms with trampoline support, additional error is generated
per each undefined symbol. This breaks TestUndefinedRelocErrors test
on these platforms.
Proposed fix:
1. Changes error text to be identical to normal undefined reloc.
If relocation is undefined, jump to it will be unresolved
as well.
2. Introduces a map that can be used by all sites that
handle this kind of errors which makes it easier
to report such errors exactly once.
Errors on ppc64 before this change (note first 4 lines):
main.defined1: unresolved inter-package jump to main.undefined()
main.defined1: unresolved inter-package jump to main.undefined()
main.defined2: unresolved inter-package jump to main.undefined()
main.defined2: unresolved inter-package jump to main.undefined()
main.defined1: relocation target main.undefined not defined
main.defined2: relocation target main.undefined not defined
runtime.main_main·f: function main is undeclared in the main package
After this change:
main.defined1: relocation target main.undefined not defined
main.defined2: relocation target main.undefined not defined
runtime.main_main·f: function main is undeclared in the main package
Because of (1), errors output is the same on all platforms now.
Fixes#25753
Change-Id: Ic3084202a6fc5d4a6d2d0a93344f012b37fe58ed
Reviewed-on: https://go-review.googlesource.com/116676
Run-TryBot: Iskander Sharipov <iskander.sharipov@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
os.Args is usually never empty and the flag package panics if it is.
This commit makes os.Args default to ["js"] for js/wasm.
Change-Id: Iba527145686487b052da438fca40159e57e61a81
Reviewed-on: https://go-review.googlesource.com/117475
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
On 32-bit architectures without native 64-bit atomic instructions,
64-bit atomics are emulated using spinlocks. However,
the sigprof handling code expects to be able to perform
64-bit atomic operations in signal handlers. Spinning on an
acquired spinlock in a signal handler leads to a livelock.
This is issue #20146.
The original fix for #20146 did not include arm in
the list of architectures that need to work around portability
issues in the sigprof handler code. The unit test designed to
catch this issue does not fail on arm builds because arm uses
striped spinlocks, and thus the livelock takes many minutes
to reproduce. This is issue #24260. (This patch doesn't completely
fix#24260 on go1.10.2 due to issue #25785, which is probably
related to the arm cas kernel helpers. Those have been removed
at tip.)
With this patch applied, I was able to run the reproducer for
issue #24260 for more than 90 minutes without reproducing the
livelock. Without this patch, the livelock took as little as
8 minutes to reproduce.
Fixes#20146
Updates #24260
Change-Id: I64bf53a14d53c4932367d919ac55e17c99d87484
Reviewed-on: https://go-review.googlesource.com/117057
Run-TryBot: Philip Hofer <phofer@umich.edu>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
CL 116975 added TestCoverHTML. However, this test is failing
on Plan 9, because the GNU diff tool is called "ape/diff"
instead of "diff" on Plan 9.
This change replaces the "diff" command by the "ape/diff"
command on Plan 9.
Fixes#25795.
Change-Id: I15b49868cd09f3f977aa13fffdfc430c882bf757
Reviewed-on: https://go-review.googlesource.com/117415
Run-TryBot: David du Colombier <0intro@gmail.com>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Turns out it doesn't currently matter, as these ops are always issued
together with a BTSL which does clobber flags. So I can't write a test
that currently fails. But better to be future-proof.
BS{F,R}Q generates flags, so it doesn't need to be marked as clobbering.
Change-Id: I70daea154023fd435fac696bf3a384803c647cd3
Reviewed-on: https://go-review.googlesource.com/117375
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This is a refinement of CL 114855, which fixed the empty clause case,
but broke some other cases where segment boundaries can coincide
for other reasons.
Fixes#25767.
Change-Id: I2a387c83f9d651c8358f3e11b03f6167af0eb8bf
Reviewed-on: https://go-review.googlesource.com/116976
Run-TryBot: David Symonds <dsymonds@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
This change improves upon cycle detection by taking into account
cycles involving constants, variables, _and_ types. All new code
(except for the additional tests) is guarded by the useCycleMarking
(internal) flag and thus can be disabled on short notice if it
introduced new problems. (The intent is to remove this flag shortly
after 1.11 is released.)
The test suite has been extended with various additional (and mostly
esoteric) test cases which now correctly report cycles. A handful of
existing test cases now report additional errors, but those are mostly
esoteric cases. Overall, this is an improvement over the status quo.
Fixes#8699.
For #20770.
Change-Id: I6086719acd0d5200edca4a3dbe703d053496af31
Reviewed-on: https://go-review.googlesource.com/116815
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
This adds a case for what was fixed in 4fe688c to prevent regression;
a follow-on change will address #25767.
Change-Id: Iced8cc10e2993ef7caf7e9c59ffbc7147d78ddd7
Reviewed-on: https://go-review.googlesource.com/116975
Run-TryBot: David Symonds <dsymonds@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
If a package is checked out in the right place, but uses, for instance,
an unusual Git remote configuration, don't refuse to update just because
cmd/go can't parse it. In most cases, `git pull -ff-only` (or the
equivalent in other VCSes) works just fine.
Updates #25432.
Change-Id: I1952a0e6e03f185f34029b186e1756c9549689e1
Reviewed-on: https://go-review.googlesource.com/113456
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Andrew Gerrand <adg@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
Users are sometimes confused why session tickets are not enabled even if
SessionTicketsDisabled is false.
Change-Id: I3b783d2cf3eed693a3ad6acb40a8003db7e0b648
Reviewed-on: https://go-review.googlesource.com/117255
Reviewed-by: Adam Langley <agl@golang.org>
Code has ended up depending on things like RSA's key generation being
deterministic given a fixed random Reader. This was never guaranteed and
would prevent us from ever changing anything about it.
This change makes certain calls randomly (based on the internal
fastrand) read an extra byte from the random Reader. This helps to
ensure that code does not depend on internal details.
I've not added this call in the key generation of ECDSA and DSA because,
in those cases, key generation is so obvious that it probably is
acceptable to do the obvious thing and not worry about code that depends
on that.
This does not affect tests that use a Reader of constant bytes (e.g. a
zeroReader) because shifting such a stream is a no-op. The stdlib uses
this internally (which is fine because it can be atomically updated if
the crypto libraries change).
It is possible that external tests could be doing the same and would
thus break if we ever, say, tweaked the way RSA key generation worked.
I feel that addressing that would be more effort than it's worth.
Fixes#21915
Change-Id: I84cff2e249acc921ad6eb5527171e02e6d39c530
Reviewed-on: https://go-review.googlesource.com/64451
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
