Andrew Gerrand [Mon, 25 Oct 2010 03:37:30 +0000 (14:37 +1100)]
container/list: fix Remove bug and use pointer to self as identifier
Remove wasn't nil'ing the *Element.id. This property was exploited
by MoveToFront and MoveToBack internally, so I renamed the existing
Remove to "remove", and created an exported wrapper "Remove" that does
the right thing for the user's sake.
Also, saved an allocation by using *List as the id rather than *byte.
Rob Pike [Fri, 22 Oct 2010 22:16:34 +0000 (15:16 -0700)]
gobs: error cleanup part 1.
Remove err from the encoderState and decoderState types, so we're
not always copying to and from various copies of the error, and then
use panic/recover to eliminate lots of error checking.
another pass might take a crack at the same thing for the compilation phase.
Rob Pike [Fri, 22 Oct 2010 18:17:40 +0000 (11:17 -0700)]
gob: allow exchange of interface values
The implemetation describes each value as a string identifying the
concrete type of the value, followed by the usual encoding of that
value. All types to be exchanged as contents of interface values
must be registered ahead of time with the new Register function.
Although this would not seem strictly necessary, the linker garbage
collects unused types so without some mechanism to guarantee
the type exists in the binary, there could be unpleasant surprises.
Moreover, the receiver needs a reflect.Type of the value to be
written in order to be able to save the data. A Register function
seems necessary.
The implementation may require defining types in the middle of
of sending a value. The old code never did this. Therefore there
has been some refactoring to make the encoder and decoder
work recursively.
This change changes the internal type IDs. Existing gob archives
will break with this change. Apologies for that. If this is a deal
breaker it should be possible to create a conversion tool.
Error handling is too complicated in this code. A subsequent
change should clean it up.
Robert Griesemer [Fri, 22 Oct 2010 17:03:14 +0000 (10:03 -0700)]
go ast/parser/printer: permit elision of composite literal types for composite literal elements
gofmt: added -s flag to simplify composite literal expressions through type elision where possible
Robert Griesemer [Fri, 22 Oct 2010 15:58:52 +0000 (08:58 -0700)]
go spec: relaxed syntax for array, slice, and map composite literals
For elements which are themselves composite literals, the type may
be omitted if it is identical to the element type of the containing
composite literal.
Russ Cox [Thu, 21 Oct 2010 15:39:47 +0000 (11:39 -0400)]
ld: abandon symbol-driven archive loading
Load the entire archive file instead.
Reduces I/O by avoiding additional passes
through libraries to resolve symbols.
Go packages always need all the files anyway
(most often, all 1 of them).
Russ Cox [Thu, 21 Oct 2010 15:25:14 +0000 (11:25 -0400)]
encoding/binary: give LittleEndian, BigEndian specific types
Giving them specific types has the benefit that
binary.BigEndian.Uint32(b) is now a direct call, not an
indirect via a mutable interface value, so it can potentially
be inlined.
Recent changes to the spec relaxed the rules for comparison,
so this code is still valid:
func isLittle(o binary.ByteOrder) { return o == binary.LittleEndian }
The change does break this potential idiom:
o := binary.BigEndian
if foo {
o = binary.LittleEndian
}
That must rewrite to give o an explicit binary.ByteOrder type.
On balance I think the benefit from the direct call and inlining
outweigh the cost of breaking that idiom.
Russ Cox [Thu, 21 Oct 2010 04:56:20 +0000 (06:56 +0200)]
arm: prop up software floating point
Just enough to make mov instructions work,
which in turn is enough to make strconv work
when it avoids any floating point calculations.
That makes a bunch of other packages pass
their tests.
Should suffice until hardware floating point
is available.
Enable package tests that now pass
(some due to earlier fixes).
Looks like there is a new integer math bug
exposed in the fmt and json tests.
Russ Cox [Wed, 20 Oct 2010 19:16:39 +0000 (12:16 -0700)]
6l: correct logic for morestack choice
The frame that gets allocated is for both
the args and the autos. If together they
exceed the default frame size, we need to
tell morestack about both so that it allocates
a large enough frame.
Sanity check stack pointer in morestack
to catch similar bugs.
Rob Pike [Wed, 20 Oct 2010 04:25:28 +0000 (21:25 -0700)]
reflect: add InterfaceValue.Get to enable setting of an interface
value (through unsafe means) without having a reflect.Type
of type *interface{} (pointer to interface). This is needed to make
gob able to handle interface values by a method analogous to
the way it handles maps.
Russ Cox [Tue, 19 Oct 2010 22:07:19 +0000 (18:07 -0400)]
5l, 6l, 8l: link pclntab and symtab as ordinary rodata symbols
That is, move the pc/ln table and the symbol table
into the read-only data segment. This eliminates
the need for a special load command to map the
symbol table into memory, which makes the
information available on systems that couldn't handle
the magic load to 0x99000000, like NaCl and ARM QEMU
and Linux without config_highmem=y. It also
eliminates an #ifdef and some clumsy code to
find the symbol table on Windows.
The bad news is that the binary appears to be bigger
than it used to be. This is not actually the case, though:
the same amount of data is being mapped into memory
as before, and the tables are still read-only, so they're
still shared across multiple instances of the binary as
they were before. The difference is just that the tables
aren't squirreled away in some section that "size" doesn't
know to look at.
This is a checkpoint.
It probably breaks Windows and breaks NaCl more
than it used to be broken, but those will be fixed.
The logic involving -s needs to be revisited too.
Yuval Pavel Zholkover [Mon, 18 Oct 2010 16:32:55 +0000 (12:32 -0400)]
8l, runtime: initial support for Plan 9
No multiple processes/locks, managed to compile
and run a hello.go (with print not fmt). Also test/sieve.go
seems to run until 439 and stops with a
'throw: all goroutines are asleep - deadlock!'
- just like runtime/tiny.
based on Russ's suggestions at:
http://groups.google.com/group/comp.os.plan9/browse_thread/thread/cfda8b82535d2d68/243777a597ec1612
Build instructions:
cd src/pkg/runtime
make clean && GOOS=plan9 make install
this will build and install the runtime.
When linking with 8l, you should pass -s to suppress symbol
generation in the a.out, otherwise the generated executable will not run.
This is runtime only, the porting of the toolchain has already
been done: http://code.google.com/p/go-plan9/source/browse
in the plan9-quanstro branch.
Russ Cox [Sun, 17 Oct 2010 15:41:23 +0000 (11:41 -0400)]
5l: stop using R12 as SB
Because the SB is only good for 8k and Go programs
tend to have much more data than that, SB doesn't
save very much. A fmt.Printf-based hello world program
has 360 kB text segment. Removing SB makes the text
500 bytes (0.14%) longer.
Russ Cox [Fri, 15 Oct 2010 19:18:47 +0000 (15:18 -0400)]
6l: function at a time code layout
Also change the span-dependent jump algorithm
to use fewer iterations:
* resolve forward jumps at their targets (comefrom list)
* mark jumps as small or big and only do small->big
* record whether a jump failed to be encodable
These changes mean that a function with only small
jumps can be laid out in a single iteration, and the
vast majority of functions take just two iterations.
I was seeing a maximum of 5 iterations before; the
max now is 3 and there are fewer that get even that far.
Russ Cox [Fri, 15 Oct 2010 03:48:40 +0000 (23:48 -0400)]
5l, 6l, 8l: accumulate data image during import
Using explicit relocations internally, we can
represent the data for a particular symbol as
an initialized block of memory instead of a
linked list of ADATA instructions. The real
goal here is to be able to hand off some of the
relocations to the dynamic linker when interacting
with system libraries, but a pleasant side effect is
that the memory image is much more compact
than the ADATA list, so the linkers use less memory.
Russ Cox [Wed, 13 Oct 2010 20:20:22 +0000 (16:20 -0400)]
various: avoid %ld etc
The Plan 9 tools assume that long is 32 bits.
We converted all instances of long to int32 when
importing the code but missed the print formats.
Because int32 is always int on the compilers we use,
it is never correct to use %lux, %ld, etc. Convert to %ux, %d, etc.
(It matters because on 64-bit gcc, long is 64 bits,
so we were printing 32-bit quantities with 64-bit formats.)