pointer.  An example that triggers the bad behavior on a 64bit
machine http://play.golang.org/p/GrNFakAYLN
        rv1 := reflect.ValueOf(complex128(0))
        rt := rv1.Type()
        rv2 := rv1.Convert(rt)
        rv3 := reflect.New(rt).Elem()
        rv3.Set(rv2)

Running the code fails with the following:
        panic: reflect: internal error: storeIword of 16-byte value

I've tested on a 64bit machine and verified this fixes the panic.  I
haven't tested on a 32bit machine so I haven't verified the other
cases, but they follow logically.

R=golang-dev, r, iant
CC=golang-dev
https://golang.org/cl/12805045

R=r, adg
CC=golang-dev, toddwang
https://golang.org/cl/13149043

Update #5000
Should reduce the flakiness a little. Malloc counting is important
to general testing but not to the build dashboard, which uses -short.

R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/12866047

R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/12801045

Add eq, lt, etc. to allow one to do simple comparisons.
It's basic types only (booleans, integers, unsigned integers,
floats, complex, string) because that's easy, easy to define,
and covers the great majority of useful cases, while leaving
open the possibility of a more sweeping definition later.

{{if eq .X .Y}}X and Y are equal{{else}}X and Y are unequal{{end}}

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/13091045

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12975044

R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/13140043

Not a user-visible API change.  Just a minor garbage
win in contrived cases.

Also, the package is io/ioutil.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12987044

This is potentially an API-breaking change, but it is an important bug fix.

The CL https://golang.org/cl/7305072/ added stuff to make
the tar file look more like a file system internally, including providing an
implementation of os.FileInfo for the file headers within the archive.
But the code is incorrect because FileInfo.Name is supposed to return
the base name only; this implementation returns the full path. A round
trip test added in the same shows this in action, as the slashes are
preserved as we create a header using the local implementation of
FileInfo.

The CL here changes the behavior of the tar (and zip) FileInfo to honor
the Go spec for that interface. It also clarifies that the FileInfoHeader
function, which takes a FileInfo as an argument, will therefore create
a header with only the base name of the file recorded, and that
subsequent adjustment may be necessary.

There may be code out there that depends on the broken behavior.
We can call out the risk in the release notes.

Fixes #6180.

R=golang-dev, dsymonds, adg, bradfitz
CC=golang-dev
https://golang.org/cl/13118043

Generated by addca.

R=gobot
CC=golang-dev
https://golang.org/cl/12831044

If xcode-select is available, use it to determine the path to the
DVTFoundation.xcplugindata file.

Fixes #5997.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12741047

Update #5199
Update #6146

R=golang-dev, dvyukov
CC=golang-dev
https://golang.org/cl/13112044

Fixes #6146.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12927048

Update #6146

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12927047

Fixes #5377

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12991046

The os windows source uses file as the receiver, not f.
TBR=golang-dev@googlegroups.com

R=adg
CC=golang-dev
https://golang.org/cl/12922044

Some crashed, some didn't. Make a nil receiver always
return ErrInvalid rather than crash.
Fixes #5824.
The program in the bug listing is silent now, at least on my Mac.

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/13108044

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/12926046

(For go.tools/ssa/interp/interp_test, which runs a subset of
these tests under this assumption.)

R=rsc, r
CC=golang-dev
https://golang.org/cl/13108043

This has broken the 32-bit builds.

««« original CL description
cmd/gc, runtime: use type information to scan interface values

R=golang-dev, rsc, dvyukov
CC=golang-dev
https://golang.org/cl/12785045
»»»

R=khr, golang-dev, khr
CC=golang-dev
https://golang.org/cl/13010045

It's a full answer to the first question; let's let readers know that.

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/12788048

When the coverage report file is older than the file we're
showing the coverage report for, then we show a simple message
to state this fact.

R=adonovan, dominik.honnef, bradfitz
CC=golang-dev
https://golang.org/cl/12919044

Generated by addca.

R=gobot
CC=golang-dev
https://golang.org/cl/13073044

R=golang-dev, rsc, dvyukov, khr
CC=golang-dev
https://golang.org/cl/11386044

Fixes #6167.

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/13052043

In the crash stack trace race cgocall() calls endcgo(),
this means that m->racecall is wrong.
Indeed this can happen is a goroutine is rescheduled to another M
during race call.
Disable preemption for race calls.
Fixes #6155.

R=golang-dev, rsc, cshapiro
CC=golang-dev
https://golang.org/cl/12866045

R=golang-dev, rsc, dvyukov
CC=golang-dev
https://golang.org/cl/12785045

Fixes #5665.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12807044

Before this CL, the import stack was a) not printed and b) overwritten later
in the build, destroying the information about the cycle. This CL fixes both.

I made time depend on os (os already depends on time) and with this CL the error is:

/Users/r/go/src/pkg/fmt/print.go:10:2: import cycle not allowed
package code.google.com/p/XXX/YYY:
        imports fmt
        imports os
        imports time
        imports os

Doesn't give line numbers for the actual imports, as requested in the bug, but
I don't believe that's important.

Fixes #4292.

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/13100043

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/12744048

syntax/*: update messages
sliceerr3.go: bizarre new error fixed by deleting a space.

I could have sworn I ran all.bash before submitting the CL that triggered these.

TBR=golang-dev@googlegroups.com

R=golang-dev
CC=golang-dev
https://golang.org/cl/12812044

Fixes #5691.

R=golang-dev, bradfitz, daniel.morsing, rsc
CC=golang-dev
https://golang.org/cl/10255047

Given
        if (i == 0)
                x++
The old message was
        x.go:6: syntax error: unexpected semicolon or newline before {
Now we see
        x.go:6: syntax error: missing { after if clause

Fixes #5687

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/12822045

Update #5305.
This handles the case where the nil pointers are inside a slice.
A top-level nil pointer is harder, maybe fundamentally broken by gob's model.
Thinking required.
However, a slice is the important case since people don't expect to be sending
top-level nils much, but they can arise easily in slices.

R=golang-dev, josharian, adg
CC=golang-dev
https://golang.org/cl/13042044

The shell's -e doesn't work across "eval"; need to error-check by hand.
The recent spate of Darwin build failures pointed out that if the first
run of cmd/dist fails, we keep going. We shouldn't.

R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/13098043

This might fix the mkdtemp problem on the darwin builders if they
have TMPDIR set to a path ending in a slash; at worse this will
result in cleaner path names.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/13097043

The tar/archive code from golang has a problem with linknames with length >
100. A pax header is added but the original header still written with a too
long field length.

As it is clear that pax support is incomplete I have added missing
implementation parts.

This commit contains code from the golang project in the folder tar/archiv.

The following pax header records are now automatically written:

- gname)
- linkpath
- path
- uname

The following fields can be written with PAX, but the default is to use the
star binary extension.

- gid  (value > 2097151)
- size (value > 8589934591)
- uid (value > 2097151)

The string fields are written when the value is longer as the field or if the
string contains a char that is not encodable as 7-bit ASCII value.

The change was tested against a current ubuntu-cloud image tarball comparing
the compressed result.

+ added some automated tests for the new functionality.

Fixes #6056.

R=dsymonds
CC=golang-dev
https://golang.org/cl/12561043

Generated by addca.

R=gobot
CC=golang-dev
https://golang.org/cl/13096043

Fixes #5285.

R=golang-dev, dvyukov
CC=golang-dev
https://golang.org/cl/13045044

R=rsc, bradfitz, adg
CC=golang-dev
https://golang.org/cl/13004046