Rob Pike [Fri, 7 Sep 2012 17:28:24 +0000 (10:28 -0700)]
spec: an initial BOM can be ignored
After further deliberation, let's back down to the Unicode proposal.
Ignoring aBOMinations anywhere means that things like
grep unsafe *.go
might fail because there's a BOM in the middle: unBOMsafe.
Joel Sing [Fri, 7 Sep 2012 03:32:40 +0000 (13:32 +1000)]
cgo: use debug data section for ELF
When generating enums use the debug data section instead of the
DWARF debug info, if it is available in the ELF file. This allows
mkerrors.sh to work correctly on OpenBSD/386 and NetBSD/386.
Dave Cheney [Thu, 6 Sep 2012 07:50:26 +0000 (17:50 +1000)]
crypto/tls: fix data race on conn.err
Fixes #3862.
There were many areas where conn.err was being accessed
outside the mutex. This proposal moves the err value to
an embedded struct to make it more obvious when the error
value is being accessed.
As there are no Benchmark tests in this package I cannot
feel confident of the impact of this additional locking,
although most will be uncontended.
Nigel Tao [Thu, 6 Sep 2012 01:10:47 +0000 (11:10 +1000)]
image/jpeg: fix quantization tables to be in zig-zag order, not natural
order.
JPEG images generated prior to this CL are still valid JPEGs, as the
quantization tables used are encoded in the wire format. Such JPEGs just
don't use the recommended quantization tables.
The old code was a depth first graph traversal that could, under the
right conditions, end up re-exploring the same subgraphs multiple
times, once for each way to arrive at that subgraph at a given depth.
The new code uses a breadth first search to make sure that it only
visits each reachable embedded struct once.
Also add fast path for the trivial case.
benchmark old ns/op new ns/op delta
BenchmarkFieldByName1 1321 187 -85.84%
BenchmarkFieldByName2 6118 5186 -15.23%
BenchmarkFieldByName3 8218553 42112 -99.49%
R=gri, r
CC=golang-dev
https://golang.org/cl/6458090
Alan Donovan [Tue, 4 Sep 2012 18:40:49 +0000 (14:40 -0400)]
runtime: discard SIGPROF delivered to non-Go threads.
Signal handlers are global resources but many language
environments (Go, C++ at Google, etc) assume they have sole
ownership of a particular handler. Signal handlers in
mixed-language applications must therefore be robust against
unexpected delivery of certain signals, such as SIGPROF.
The default Go signal handler runtime·sigtramp assumes that it
will never be called on a non-Go thread, but this assumption
is violated by when linking in C++ code that spawns threads.
Specifically, the handler asserts the thread has an associated
"m" (Go scheduler).
This CL is a very simple workaround: discard SIGPROF delivered to non-Go threads. runtime.badsignal(int32) now receives the signal number; if it returns without panicking (e.g. sig==SIGPROF) the signal is discarded.
I don't think there is any really satisfactory solution to the
problem of signal-based profiling in a mixed-language
application. It's not only the issue of handler clobbering,
but also that a C++ SIGPROF handler called in a Go thread
can't unwind the Go stack (and vice versa). The best we can
hope for is not crashing.
Note:
- I've ported this to all POSIX platforms, except ARM-linux which already ignores unexpected signals on m-less threads.
- I've avoided tail-calling runtime.badsignal because AFAICT the 6a/6l don't support it.
- I've avoided hoisting 'push sig' (common to both function calls) because it makes the code harder to read.
- Fixed an (apparently incorrect?) docstring.
Patrick Crosby [Tue, 4 Sep 2012 01:27:20 +0000 (11:27 +1000)]
net/http/pprof: updated documentation (run an http server)
Added instructions for starting an http server
to the godoc header for this package. With the old
instructions, the example "go tool pprof..." commands
wouldn't work unless you happen to be running an http
server on port 6060 in your application.
Shenghou Ma [Sun, 2 Sep 2012 19:49:03 +0000 (03:49 +0800)]
doc/progs: use test/run.go for testing on Windows
cgo[1-4].go, go1.go couldn't be tested now
(cgo[1-4].go can only be tested when cgo is enabled, go1.go
contain a list of filenames in the current directory)
codereview.py: correct error handling without decorator
The decorator hides the number of function arguments from Mercurial,
so Mercurial cannot give proper error messages about commands
invoked with the wrong number of arguments.
Left a 'dummy' hgcommand decorator in place as a way to document
what functions are hg commands, and just in case we need some other
kind of hack in the future.
There was mail on golang-nuts a few weeks ago
from someone who understood the message perfectly
and knew he had a cyclic dependency but assumed
that Go, like Python or Java, was supposed to handle it.
R=golang-dev, bradfitz, dave
CC=golang-dev
https://golang.org/cl/6488069
exp/locale/collate: Added functionality to parse and process LDML files
for both locale-specific exemplar characters and tailorings to
the collation table.
Some specifices:
- Moved stringSet to the beginning of the file and added some functionality
to parse command line files.
- openReader now modifies the input URL for localFiles to guarantee that
any http source listed in the generated file is indeed this source.
- Note that the implementation of the Tailoring API used by maketables.go
is not yet checked in. So for now adding tailorings are simply no-ops.
- The generated file of exemplar characters will be used somewhere else.
Here is a snippet of how the body of the generated file looks like:
type exemplarType int
const (
exCharacters exemplarType = iota
exContractions
exPunctuation
exAuxiliary
exCurrency
exIndex
exN
)
var exemplarCharacters = map[string][exN]string{
"af": [exN]string{
0: "a á â b c d e é è ê ë f g h i î ï j k l m n o ô ö p q r s t u û v w x y z",
3: "á à â ä ã æ ç é è ê ë í ì î ï ó ò ô ö ú ù û ü ý",
4: "a b c d e f g h i j k l m n o p q r s t u v w x y z",
},
...
}
exp/locale/collate: first changes that introduce implementation of tailorings:
- Elements in the array are now sorted as a linked list. This makes it easier to
apply tailorings.
- Added code to sort entries by collation elements.
- Added logical reset points. This is used for tailoring relative to certain
properties, rather than characters.
NOTE: all code for type entry should now be in order.go. To keep the diffs for
this CL reasonable, though, the existing code is left in builder.go. I'll move
this in a separate CL.
Dave Cheney [Fri, 31 Aug 2012 23:17:14 +0000 (09:17 +1000)]
syscall: add PtraceSyscall(pid int, signal int)
Fixes #3525.
PTRACE_SYSCALL behaves like PTRACE_CONT and can deliver
a signal to the process. Ideally PtraceSingleStep should
support the signal argument, but its interface is frozen
by Go1.
Dave Cheney [Fri, 31 Aug 2012 10:17:59 +0000 (20:17 +1000)]
cgo/misc/test: burn CPU to improve sleep accuracy
Fixes #4008.
Run a background goroutine that wastes CPU to trick the
power management into raising the CPU frequency which,
by side effect, makes sleep more accurate on arm.
=== RUN TestParallelSleep
--- PASS: TestParallelSleep (1.30 seconds)
_cgo_gotypes.go:772: sleep(1) slept for 1.000458s
R=minux.ma, r
CC=golang-dev
https://golang.org/cl/6498060
Rob Pike [Thu, 30 Aug 2012 04:42:53 +0000 (21:42 -0700)]
text/template: make spaces significant
Other than catching an error case that was missed before, this
CL introduces no changes to the template language or API.
For simplicity, templates use spaces as argument separators.
This means that spaces are significant: .x .y is not the same as .x.y.
In the existing code, these cases are discriminated by the lexer,
but that means for instance that (a b).x cannot be distinguished
from (a b) .x, which is lousy. Although that syntax is not
supported yet, we want to support it and this CL is a necessary
step.
This CL emits a "space" token (actually a run of spaces) from
the lexer so the parser can discriminate these cases. It therefore
fixes a couple of undisclosed bugs ("hi".x is now an error) but
doesn't otherwise change the language. Later CLs will amend
the grammar to make .X a proper operator.
There is one unpleasantness: With space a token, three-token
lookahead is now required when parsing variable declarations
to discriminate them from plain variable references. Otherwise
the change isn't bad.
The CL also moves the debugging print code out of the lexer
into the test, which is the only place it's needed or useful.
Step towards resolving issue 3999.
It still remains to move field chaining out of the lexer
and into the parser and make field access an operator.
Rob Pike [Wed, 29 Aug 2012 21:46:57 +0000 (14:46 -0700)]
go_spec.html: clarify rune and string literals
No changes to the meaning, just clearer language and more
examples, including illegal rune and string literals.
In particular, "character literal" and "character constant"
are now called "rune literal" and "rune constant" and the
word "character" always refers to the source text, not
program values.
Andrew Balholm [Wed, 29 Aug 2012 17:06:37 +0000 (10:06 -0700)]
cmd/yacc/units.txt: fix exchange rates
In the example "units" program for goyacc, the exchange rates were
reciprocals of the correct amounts. Turn them right-side-up
and update them to current figures.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6495053
Dave Cheney [Tue, 28 Aug 2012 23:05:30 +0000 (09:05 +1000)]
net/http/httputil: fix race in DumpRequestOut
Fixes #3892.
Swapping the order of the writers inside the MultiWriter ensures
the request will be written to buf before http.ReadRequest completes.
The fencedBuffer is not required to make the test pass on
any machine that I have access too, but as the buf is shared
across goroutines, I think it is necessary for correctness.
Tobias Columbus [Mon, 27 Aug 2012 19:59:16 +0000 (03:59 +0800)]
misc/vim: fix for autocompletion
Vim autocompletion respects the $GOPATH variable and does not
ignore dashes ('-'), dots ('.') and underscores ('_') like found
in many remote packages.
Environment variable $GOROOT is determined by calling
'go env GOROOT' instead of relying on the user's environment
variables.
Joel Sing [Sun, 26 Aug 2012 10:57:47 +0000 (20:57 +1000)]
runtime: use netbsd signal ABI v2
Use version 2 of the NetBSD signal ABI - both version 2 and version 3
are supported by the kernel, with near identical behaviour. However,
the netbsd32 compat code does not allow version 3 to be used, which
prevents Go netbsd/386 binaries from running in compat mode on a
NetBSD amd64 kernel. Switch to version 2 of the ABI, which is the
same version currently used by NetBSD's libc.
Rob Pike [Fri, 24 Aug 2012 20:00:24 +0000 (13:00 -0700)]
text/template: catch (A).X as a parse error
This shouldn't be an error (see issue 3999), but until it's handled
correctly, treat it as one to avoid confusion. Without this CL,
(A).X parses as two arguments.
Rob Pike [Fri, 24 Aug 2012 19:37:23 +0000 (12:37 -0700)]
text/template: allow grouping of pipelines using parentheses
Based on work by Russ Cox. From his CL:
This is generally useful but especially helpful when trying
to use the built-in boolean operators. It lets you write:
{{if not (f 1)}} foo {{end}}
{{if and (f 1) (g 2)}} bar {{end}}
{{if or (f 1) (g 2)}} quux {{end}}
instead of
{{if f 1 | not}} foo {{end}}
{{if f 1}}{{if g 2}} bar {{end}}{{end}}
{{$do := 0}}{{if f 1}}{{$do := 1}}{{else if g 2}}{{$do := 1}}{{end}}{{if $do}} quux {{end}}
The result can be a bit LISPy but the benefit in expressiveness and readability
for such a small change justifies it.
I believe no changes are required to html/template.
Marcel van Lohuizen [Fri, 24 Aug 2012 13:56:07 +0000 (15:56 +0200)]
exp/locale/collate: add code to ignore tests with (unpaired) surrogates.
In the regtest data, surrogates are assigned primary weights based on
the surrogate code point value. Go now converts surrogates to FFFD, however,
meaning that the primary weight is based on this code point instead.
This change drops tests with surrogates and lets the tests pass.
Julien Schmidt [Fri, 24 Aug 2012 02:29:47 +0000 (19:29 -0700)]
database/sql: stop reuse of bad connections
The second parameter for sql.putConn() (err) is always nil. As a result bad
connections are reused, even if the driver returns an driver.ErrBadConn.
Unsing a pointer to err instead achievs the desired behavior.
See http://code.google.com/p/go/issues/detail?id=3777 for more details.
Fixes #3777.
Shenghou Ma [Thu, 23 Aug 2012 06:35:26 +0000 (14:35 +0800)]
cmd/6g: fix float32/64->uint64 conversion
CVTSS2SQ's rounding mode is controlled by the RC field of MXCSR;
as we specifically need truncate semantic, we should use CVTTSS2SQ.
Fixes #3804.
R=rsc, r
CC=golang-dev
https://golang.org/cl/6352079
Nigel Tao [Thu, 23 Aug 2012 06:17:22 +0000 (16:17 +1000)]
cmd/8g: roll back the small integer constant optimizations introduced
in 13416:67c0b8c8fb29 "faster code, mainly for rotate" [1]. The codegen
can run out of registers if there are too many small-int arithmetic ops.
An alternative approach is to copy 6g's sbop/abop codegen to 8g, but
this change is less risky.