Implement the foster-parenting algorithm for content that is inside a table
but not in a cell.

Also fix a bug in reconstructing the active formatting elements.

Pass test 30 in tests1.dat:
<a><table><td><a><table></table><a></tr><a></table><b>X</b>C<a>Y

R=nigeltao
CC=golang-dev
https://golang.org/cl/5309052

The additional test case in parse_test.go is:
<select><b><option><select><option></b></select>X

R=andybalholm
CC=golang-dev
https://golang.org/cl/5293051

- no explicit API change, but new(big.Rat) now creates a big.Rat value
  of 0 that is immediately usable, in sync. w/ the conventions elsewhere
- various cleanups along the way

R=r
CC=golang-dev
https://golang.org/cl/5301056

R=rsc
CC=golang-dev
https://golang.org/cl/5293058

Also:
- changed semantics of return values for [Int|Rat].SetString
  if an error occured (returned value is nil); will expose
  hidden errors where return values are not checked
- added more tests
- various cleanups throughout

Fixes #2384.

R=r
CC=golang-dev
https://golang.org/cl/5312044

Because gotest's args are mingled with the tests, it's
hard to get the usage message to print. This CL adds
explicit support for -help, spelled several different ways.
Gotest has special flags like -file that are somewhat
hidden otherwise.

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5298052

Fixes #2386

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5312045

server.go/channel.go:
* rename Server to ServerConfig to match Client.
* rename ServerConnection to ServeConn to match Client.
* add Listen/Listener.
* ServerConn.Handshake(), general cleanups.

client.go:
* fix bug where fmt.Error was not assigned to err

R=rsc, agl
CC=golang-dev
https://golang.org/cl/5265049

Requires CL 5285044

client.go:
* add Dial, ClientConn, ClientChan, ClientConfig and Cmd.

doc.go:
* add Client documentation.

server.go:
* adjust for readVersion change.

transport.go:
* return an os.Error not a bool from readVersion.

R=rsc, agl, n13m3y3r
CC=golang-dev
https://golang.org/cl/5162047

Avoids test failure due to undefined map iteration order.

R=rsc, iant, iant
CC=golang-dev
https://golang.org/cl/5297048

R=gri
CC=golang-dev
https://golang.org/cl/5292048

R=alex.brainman, mattn.jp
CC=golang-dev
https://golang.org/cl/5311047

The first additional test case in parse_test.go is:
<!--><div>--<!-->

The second one is unrelated to the comment change, but also passes:
<p><hr></p>

R=andybalholm
CC=golang-dev
https://golang.org/cl/5299047

R=golang-dev, n13m3y3r
CC=golang-dev
https://golang.org/cl/5303050

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5296044

Adds check that, for example, a Scan method taking
a first argument of type fmt.ScanState has the correct
signature to satisfy fmt.Scanner.

Similarly, a ReadByte should return byte, os.Error.

These are important to check because various pieces
of code (fmt, gob, json, flate) do dynamic checks
for these methods, so code with incorrect signatures
would not be flagged at compile time.

These become even more important to check when
rune is introduced.

R=r
CC=golang-dev
https://golang.org/cl/5305044

Makes tables.go output consistent across maketable runs.
(It was already inconsistent across architectures; the new
map iteration order just make it inconsistent across runs.)

R=r
CC=golang-dev
https://golang.org/cl/5303046

Shows first line of any response that the codereview server
has identified as an LGTM.  Example output below.

5305046:
        big: update for fmt interface changes

        Nothing terribly interesting here.

        Reviewer: gri@golang.org
                gri: LGTM
        CC: golang-dev@googlegroups.com
        Files:
                src/pkg/big/int.go
                src/pkg/big/nat.go
                src/pkg/big/nat_test.go
                src/pkg/big/rat.go

5307044:
        exp/template/html: use rune

        Nothing terribly interesting here.

        Reviewer: mikesamuel@gmail.com, nigeltao@golang.org
                mikesamuel: I don't see a type def for rune.  Assuming that's a new intrinsic, LGTM.
        CC: golang-dev@googlegroups.com
        Files:
                src/pkg/exp/template/html/css.go
                src/pkg/exp/template/html/css_test.go
                src/pkg/exp/template/html/html.go
                src/pkg/exp/template/html/js.go

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/5297045

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5297044

R=gri
CC=golang-dev
https://golang.org/cl/5302044

I had a brain fart in af84b15fbae2 and messed up the names in
generated certificates.

R=rsc, bradfitz
CC=golang-dev
https://golang.org/cl/5315046

It hasn't been primitive in a while.

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/5287041

R=golang-dev, dave, iant
CC=golang-dev
https://golang.org/cl/5299045

regenerated on a debian sid arm5 host.

R=bradfitz, iant
CC=golang-dev
https://golang.org/cl/5298046

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/5295044

R=golang-dev
CC=golang-dev
https://golang.org/cl/5292045

R=rsc
CC=golang-dev
https://golang.org/cl/5302041

R=mikesamuel
CC=golang-dev
https://golang.org/cl/5313044

Store the reflect.Value in the internal print state. Code is simpler, cleaner,
and a little faster - back to what it was before the change.

R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/5299046

The mechanism to record the error in the call is already in place.
Fixes #2382.

R=golang-dev, dsymonds, bradfitz, r
CC=golang-dev
https://golang.org/cl/5307043

cssEscaper escapes using the CSS convention: `\` + hex + optional-space

It outputs the space when the escape could be followed by
a hex digit to distinguish a "\na" from "\u00aa".

It did not output a space when the escape is followed by a space
character so did not distinguish "\n " from "\n".

Currently when doing lookahead, it does not distinguish spaces that
will be escaped later by the same function from ones that will not.
This is correct but suboptimal.

R=nigeltao
CC=golang-dev
https://golang.org/cl/5306042

R=golang-dev, rsc, rsc
CC=golang-dev
https://golang.org/cl/5283052

Pass tests1.dat, test 26:
#data
<script><div></script></div><title><p></title><p><p>
#document
| <html>
|   <head>
|     <script>
|       "<div>"
|     <title>
|       "<p>"
|   <body>
|     <p>
|     <p>

Thanks to Andy Balholm for driving this change.

R=andybalholm
CC=golang-dev
https://golang.org/cl/5301042

Move string table to the end of the __LINKEDIT segment.

This change allows Apple's codesign(1) utility to successfully sign
Go binaries, as long as they don't contain DWARF data (-w flag to
8l/6l).  This is because codesign(1) expects the string table to be
the last part of the file.

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5271050

1. Generate TOC for package pages using template,
instead of using JavaScript magic.  This makes the
pages generated by godoc -html easier to export
to other systems.

2. Make TOC one column.  It's hard to do two columns
portably without invoking JavaScript.

3. Since the TOC is only one column, show full type
signatures for functions and methods.  Many times
that's all you need to see anyway.

4. Name the section after the TOC "Overview".
Naming it something is important, to set it off
from the TOC and so that there's a quick link to
it in the TOC.

For now, some illustrative examples:

http://swtch.com:6060/pkg/io/
http://swtch.com:6060/pkg/strings/
http://swtch.com:6060/pkg/tabwriter/
http://swtch.com:6060/pkg/unicode/

Fixes #1982.

R=gri, bradfitz, r
CC=golang-dev
https://golang.org/cl/5303044

R=rsc, alex.brainman, bradfitz
CC=bsiegert, go.peter.90, golang-dev
https://golang.org/cl/4188047

If the length of the interpreter string
pushes us over the ELFRESERVE limit, the
resulting error message will be comical.

I was doing some ELF tinkering with a
modified version of 8l when I hit this.

To be clear, the stock linkers wouldn't
hit this without adding about forty more
section headers.  We're safe for now. ;)

Also, remove a redundant call to cflush.

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5268044

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/5272050

Fixes #2379.

R=r, rsc
CC=golang-dev
https://golang.org/cl/5305043

There is no semantic change here, just better errors.
If a function says it takes a byte, and you pass it an int,
the compiler error now says that you need a byte, not
that you need a uint8.

Groundwork for rune.

R=ken2
CC=golang-dev
https://golang.org/cl/5300042