The pointer will eventually let us find *T given T.
This CL just makes room for it, always storing a zero.
R=r, r2
CC=golang-dev
https://golang.org/cl/4221046
In CL 4188061 I changed malg to allocate the requested
number of bytes n, not n+StackGuard, so that the
allocations would use rounder numbers.
The allocation of the signal stack asks for 32k and
then used g->stackguard as the base, but g->stackguard
is StackGuard bytes above the base. Previously, asking
for 32k meant getting 32k+StackGuard bytes, so using
g->stackguard as the base was safe. Now, the actual base
must be computed, so that the signal handler does not
run StackGuard bytes past the top of the stack.
Was causing flakiness mainly in programs that use the
network, because they sometimes write to closed network
connections, causing SIGPIPEs. Was also causing problems
in the doc/progs test.
Also fix Makefile so that changes to stack.h trigger rebuild.
R=bradfitzgo, r, r2
CC=golang-dev
https://golang.org/cl/4230044
Avoids deadlocks like the one below, in which a stack split happened
in order to call lock(&stacks), but then the stack unsplit cannot run
because stacks is now locked.
The only code calling stackalloc that wasn't on a scheduler
stack already was malg, which creates a new goroutine.
runtime.futex+0x23 /home/rsc/g/go/src/pkg/runtime/linux/amd64/sys.s:139
runtime.futex()
futexsleep+0x50 /home/rsc/g/go/src/pkg/runtime/linux/thread.c:51
futexsleep(0x5b0188, 0x300000003, 0x100020000, 0x4159e2)
futexlock+0x85 /home/rsc/g/go/src/pkg/runtime/linux/thread.c:119
futexlock(0x5b0188, 0x5b0188)
runtime.lock+0x56 /home/rsc/g/go/src/pkg/runtime/linux/thread.c:158
runtime.lock(0x5b0188, 0x7f0d27b4a000)
runtime.stackfree+0x4d /home/rsc/g/go/src/pkg/runtime/malloc.goc:336
runtime.stackfree(0x7f0d27b4a000, 0x1000, 0x8, 0x7fff37e1e218)
runtime.oldstack+0xa6 /home/rsc/g/go/src/pkg/runtime/proc.c:705
runtime.oldstack()
runtime.lessstack+0x22 /home/rsc/g/go/src/pkg/runtime/amd64/asm.s:224
runtime.lessstack()
----- lessstack called from goroutine 2 -----
runtime.lock+0x56 /home/rsc/g/go/src/pkg/runtime/linux/thread.c:158
runtime.lock(0x5b0188, 0x40a5e2)
runtime.stackalloc+0x55 /home/rsc/g/go/src/pkg/runtime/malloc.c:316
runtime.stackalloc(0x1000, 0x4055b0)
runtime.malg+0x3d /home/rsc/g/go/src/pkg/runtime/proc.c:803
runtime.malg(0x1000, 0x40add9)
runtime.newproc1+0x12b /home/rsc/g/go/src/pkg/runtime/proc.c:854
runtime.newproc1(0xf840027440, 0x7f0d27b49230, 0x0, 0x49f238, 0x40, ...)
runtime.newproc+0x2f /home/rsc/g/go/src/pkg/runtime/proc.c:831
runtime.newproc(0x0, 0xf840027440, 0xf800000010, 0x44b059)
...
R=r, r2
CC=golang-dev
https://golang.org/cl/4216045
A terminal panic (one that prints a stack trace and exits)
has been calling runtime.breakpoint before calling exit,
so that if running under a debugger, the debugger can
take control. When not running under a debugger, though,
this causes an additional SIGTRAP on Unix and pop-up
dialogs on Windows.
Support for debugging Go programs has gotten good
enough that we can rely on the debugger to set its own
breakpoint on runtime.exit if it wants to look around.
R=r, r2
CC=golang-dev
https://golang.org/cl/4222043
Much yet to come, but this is a safe first step, introducing
an in-the-future configurable Client object (where policy for
cookies, auth, redirects will live) as well as introducing a
ClientTransport interface for sending requests.
The CL intentionally ignores everything around the creation
and configuration of Clients and merely ports/wraps the old
interfaces to/around Client/ClientTransport.
R=rsc, dsymonds, nigeltao, bradfitzwork
CC=golang-dev
https://golang.org/cl/4182086
The existing code assumed that signals only arrived
while executing on the goroutine stack (g == m->curg),
not while executing on the scheduler stack (g == m->g0).
Most of the signal handling trampolines correctly saved
and restored g already, but the sighandler C code did not
have access to it.
Some rewriting of assembly to make the various
implementations as similar as possible.
Will need to change Windows too but I don't
understand how sigtramp gets called there.
R=r
CC=golang-dev
https://golang.org/cl/4203042
With this change, a panic trace due to a signal arriving while
running on the scheduler stack during a lessstack
(a stack unsplit) will trace through the lessstack to show
the state of the goroutine that was unsplitting its stack.
R=r
CC=golang-dev
https://golang.org/cl/4206042
There are further changes required for things like
recursive map types. Recursive struct types work
but the mechanism needs generalization. The
case handled in this CL is pathological since it
cannot be represented at all by gob, so it should
be handled separately. (Prior to this CL, encode
would recur forever.)
R=rsc
CC=golang-dev
https://golang.org/cl/4206041
There are some minor irregularities in the printer
output (some paren's are present that should be
removed), but these are unrelated issues.
Will review in a 2nd step.
R=rsc
CC=golang-dev
https://golang.org/cl/4188068
As a result, parsing a "control clause" is now sufficiently
different for if, switch, and for statements that the code
is not factored out anymore. The code is a bit longer but
clearer in each individual case.
Reflect the changes in AST.
R=r, r2
CC=golang-dev
https://golang.org/cl/4173075
Fix problems found.
On amd64, various library routines had bigger
stack frames than expected, because large function
calls had been added.
runtime.assertI2T: nosplit stack overflow
120 assumed on entry to runtime.assertI2T
8 after runtime.assertI2T uses 112
0 on entry to runtime.newTypeAssertionError
-8 on entry to runtime.morestack01
runtime.assertE2E: nosplit stack overflow
120 assumed on entry to runtime.assertE2E
16 after runtime.assertE2E uses 104
8 on entry to runtime.panic
0 on entry to runtime.morestack16
-8 after runtime.morestack16 uses 8
runtime.assertE2T: nosplit stack overflow
120 assumed on entry to runtime.assertE2T
16 after runtime.assertE2T uses 104
8 on entry to runtime.panic
0 on entry to runtime.morestack16
-8 after runtime.morestack16 uses 8
runtime.newselect: nosplit stack overflow
120 assumed on entry to runtime.newselect
56 after runtime.newselect uses 64
48 on entry to runtime.printf
8 after runtime.printf uses 40
0 on entry to vprintf
-8 on entry to runtime.morestack16
runtime.selectdefault: nosplit stack overflow
120 assumed on entry to runtime.selectdefault
56 after runtime.selectdefault uses 64
48 on entry to runtime.printf
8 after runtime.printf uses 40
0 on entry to vprintf
-8 on entry to runtime.morestack16
runtime.selectgo: nosplit stack overflow
120 assumed on entry to runtime.selectgo
0 after runtime.selectgo uses 120
-8 on entry to runtime.gosched
On arm, 5c was tagging functions NOSPLIT that should
not have been, like the recursive function printpanics:
printpanics: nosplit stack overflow
124 assumed on entry to printpanics
112 after printpanics uses 12
108 on entry to printpanics
96 after printpanics uses 12
92 on entry to printpanics
80 after printpanics uses 12
76 on entry to printpanics
64 after printpanics uses 12
60 on entry to printpanics
48 after printpanics uses 12
44 on entry to printpanics
32 after printpanics uses 12
28 on entry to printpanics
16 after printpanics uses 12
12 on entry to printpanics
0 after printpanics uses 12
-4 on entry to printpanics
R=r, r2
CC=golang-dev
https://golang.org/cl/4188061
Other than maybe cleaning the code up a bit, this has
little practical effect for now, but lays the foundation
for remembering the method set of a type, which can
be expensive.
R=rsc
CC=golang-dev
https://golang.org/cl/4193041
It was observed that the interface was generic enough
that several other types implemented it too.
Fixes#1530.
R=rsc
CC=golang-dev
https://golang.org/cl/4169063
Removed a redefinition of the request URL which is never used and
redundant checking of the return value from send().
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/4184061
Using make -C $* rather than (cd $* ; make) results in a small,
but measurable improvement in build times where compilation is
not the major component. eg.
before - ~/go/src/pkg$ time make
real 0m1.176s
user 0m0.639s
sys 0m0.399s
after - ~/go/src/pkg$ time make
real 0m0.916s
user 0m0.571s
sys 0m0.243s
There are other places in the distribution src/make.common for example
that could also benefit from this change.
R=adg
CC=golang-dev, rsc
https://golang.org/cl/4174055
PKCS#1 v2.1 section 7.1.1 says that the result of an OAEP encryption
is "an octet string of length $k$". Since we didn't left-pad the
result it was previously possible for the result to be smaller when
the most-significant byte was zero.
Fixes#1519.
R=rsc
CC=golang-dev
https://golang.org/cl/4175059
Flags defined in the testing package may conflict
with real flags defined in the main package, or in
any other imported package.
This change makes them less likely to be used for
other purposes.
R=r, rsc, r2
CC=golang-dev
https://golang.org/cl/4167055
I'm not sure if it's 100% correct wrt the HTML5 specification,
but the test suite has plenty of HTML comment test cases, and
we'll shake out any tokenization bugs as the parser improves its
coverage.
R=gri
CC=golang-dev
https://golang.org/cl/4186055
notes:
Darwin is very particular about joining a multicast group if the
listneing socket is not created in "udp4" mode, the other supported
OS's are more flexible.
A simple example sets up a socket to listen on the mdns/bonjour
group 224.0.0.251:5353
// ensure the sock is udp4, and the IP is a 4 byte IPv4
socket, err := net.ListenUDP("udp4", &net.UDPAddr {
IP: net.IPv4zero,
// currently darwin will not allow you to bind to
// a port if it is already bound to another process
Port: 5353,
})
if err != nil {
log.Exitf("listen %s", err)
}
defer socket.Close()
err = socket.JoinGroup(net.IPv4(224, 0, 0, 251))
if err != nil {
log.Exitf("join group %s", err)
}
R=adg, rsc
CC=golang-dev
https://golang.org/cl/4066044
Yesterday's change was too simple-minded and failed if an
interface value was being discarded. We need to parse the
data stream and remember any type information that arrives.
Also fix a minor bug when ignoring an interface: toss only what
we know about, not everything.
R=rsc
CC=golang-dev
https://golang.org/cl/4179045
Now that packet/ is checked in, we can add its Makefile. Also, a couple
of updates to error/ and s2k/ for bugfixes and to use the new crypto
package.
R=bradfitzgo
CC=golang-dev
https://golang.org/cl/4179043
(The unittest for Signature may seem a little small, but it's tested by
the higher level code.)
R=bradfitzgo
CC=golang-dev
https://golang.org/cl/4173043
In line with other functions such as Fprintf, put the
thing to be written first.
Apologies for the breakages this is sure to cause.
R=rsc, gri, adg, eds, r2, aam
CC=golang-dev
https://golang.org/cl/4169042
BSD and Darwin require an extra page between
end and the first mapping, and Windows has various
memory in the way too.
Fixes#1464.
R=r, r2
CC=golang-dev
https://golang.org/cl/4167041
Add Error type to enable clients to distinguish
between local and remote errors.
Also return "connection shut down error" after
the first error return rather than returning the
same error each time.
R=r
CC=golang-dev
https://golang.org/cl/4080058
Neither gofmt nor godoc are making use of a Styler (for
token-specific formatting) anymore. Stylers interacted in complicated
ways with HTML-escaping which was why the printer needed an HTML mode
in the first place.
godoc now uses a more powerful and general text formatting
function that does HTML escaping, text selection, and can
handle token-specific formatting if so desired (currently
used only for comments).
As a consequence, cleaned up uses of go/printer in godoc;
simplified the various write utility functions, and also
removed the need for the "html" template format (in favor of
html-esc which now does the same and is used more pervasively).
Applied gofmt -w src misc to verify no changes occured,
and tested godoc manually.
There should be no visible changes except that (type) code
snippets presented for godoc package documentation now
uses the same formatting as for general source code and
thus comments get the comment-specific color here as well
(not the case at the moment).
(TODO: godoc needs a good automatic test suite).
R=rsc
CC=golang-dev
https://golang.org/cl/4152042
Even if local, it requires communication with a daemon
which may not be available. This is creating problems
for getting an Ubuntu package going in Launchpad's PPA.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/3989062
Faster in most cases, and not prone to memory leaks. Named "Do" to match with similarly named method on Vector.
R=gri
CC=golang-dev
https://golang.org/cl/4134046
Currently, when an importer closes the connection, the exporter gives an
error message 'netchan export: error decoding client header:EOF'. This
change causes the exporter to look for an EOF during the parse of the
header, and silences the log message in that case.
R=r
CC=golang-dev, rog
https://golang.org/cl/4132044
Previously, the outer loop would continue until we selected the
client's least preferable ciphersuite.
R=golang-dev, r2
CC=golang-dev
https://golang.org/cl/4029056
OS X, at least, appears to test |byte == 255|, not |byte != 0| to
establish if a bool is true or false.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/4128064
Many recently issued certificates are chained: there's one or more
intermediate certificates between the host certificate and the root CA
certificate. This change causes the code to load any number of
certificates from the certificate file. This matches the behaviour of
common webservers, and the output of OpenSSL's command line tools.
R=golang-dev, r2
CC=golang-dev
https://golang.org/cl/4119057
If a multi-line raw string is the first token on a line, it
should not be indented because the following lines (belonging
to the raw string) are not indented either.
Adjusted src of ebnf/ebnf_test.go manually as it now is formatted
as expected.
gofmt -w src misc
Fixes#1072.
R=r
CC=golang-dev
https://golang.org/cl/4119056
signed or unsigned integers using %v or the formatless scanner.
That is, Sscan("0x11", &i) or Sscanf("0x11", "%v", &i) will now
set i to 17. If a format other than %v is presented, the behavior
is as before.
Fixes#1469.
R=rsc
CC=golang-dev
https://golang.org/cl/4131042
Must be invoked as ./clean.bash --gomake make
(or --gomake gmake, depending on the name of
GNU make).
R=niemeyer
CC=golang-dev
https://golang.org/cl/4023065
Note that DSA public key support is nascent and the verification
functions clearly don't support it yet. I'm intending to get RSA keys
working first.
R=bradfitzgo
CC=golang-dev
https://golang.org/cl/3973054
ReadRune. (If you have a Reader but not a RuneReader, use bufio.)
The matching code is a few percent slower but significantly cleaner.
R=rsc
CC=golang-dev
https://golang.org/cl/4125046
As discussed in the mailing list, this adds a simple barrier
implementation to the sync package which enables one or more
goroutines to wait for a counter to go down to zero.
R=rsc, rog, r
CC=golang-dev
https://golang.org/cl/3770045
notes:
* due to Issue 1466 the Msghdr struct for
src/pkg/syscall/ztypes_darwin_386.go
src/pkg/syscall/ztypes_darwin_amd64.go
had to be edited after the godefs generation.
* ztypes_*.go files for linux, freebsd and darwin
have been prepared on the correct host OS and ARCH.
While the total increase is a dozen lines per file
the diff is larger due to a change to godefs,
http://code.google.com/p/go/source/detail?r=c79e30afe9c8
while has altered the names of Pad members which
causes gofmt to realign the affected structs
R=rsc, mikioh
CC=golang-dev
https://golang.org/cl/4119053
Since nobody suggested major changes to the higher level API, I'm
splitting up the lower level code for review. This is the first of the
changes for the packet reading/writing code.
It deliberately doesn't include a Makefile because the package is
incomplete.
R=bradfitzgo
CC=golang-dev
https://golang.org/cl/4080051
* Don't require lines to be full.
* Don't forget to flush the line buffer.
* Update the test so that it doesn't happen to include only full lines
in order to test the above.
* Always write the line after the header as GNUPG expects it.
R=bradfitzgo
CC=golang-dev
https://golang.org/cl/4124043
OpenPGP changed its OCFB mode for more modern packets (for example, the
MDC symmetrically encrypted packet). This change adds a bool to
determine which mode is used.
R=bradfitzgo, r, rsc
CC=golang-dev
https://golang.org/cl/4126041
GC is still single-threaded.
Multiple threads will happen in another CL.
Garbage collection pauses are typically
about half as long as they were before this CL.
R=brainman, iant, r
CC=golang-dev
https://golang.org/cl/3975046
so that spawned processes avoid inheriting pipes.
Implement CloseOnExec.
Make file and pipe handles inheritable.
R=rsc, brainman, vincent.vanackere
CC=golang-dev
https://golang.org/cl/4126047
Follow morestack, so that crashes during a stack split
give complete traces. Also mark stack segment boundaries
as an aid to debugging.
Correct various line number bugs with yet another attempt
at interpreting the pc/ln table. This one has a chance at
being correct, because I based it on reading src/cmd/ld/lib.c
instead of on reading the documentation.
Fixes#1138.
Fixes#1430.
Fixes#1461.
throw: runtime: split stack overflow
runtime.throw+0x3e /home/rsc/g/go2/src/pkg/runtime/runtime.c:78
runtime.throw(0x81880af, 0xf75c8b18)
runtime.newstack+0xad /home/rsc/g/go2/src/pkg/runtime/proc.c:728
runtime.newstack()
runtime.morestack+0x4f /home/rsc/g/go2/src/pkg/runtime/386/asm.s:184
runtime.morestack()
----- morestack called from stack: -----
runtime.new+0x1a /home/rsc/g/go2/src/pkg/runtime/malloc.c:288
runtime.new(0x1, 0x0, 0x0)
gongo.makeBoard+0x33 /tmp/Gongo/gongo_robot_test.go:344
gongo.makeBoard(0x809d238, 0x1, 0xf76092c8, 0x1)
----- stack segment boundary -----
gongo.checkEasyScore+0xcc /tmp/Gongo/gongo_robot_test.go:287
gongo.checkEasyScore(0xf764b710, 0x0, 0x809d238, 0x1)
gongo.TestEasyScore+0x8c /tmp/Gongo/gongo_robot_test.go:255
gongo.TestEasyScore(0xf764b710, 0x818a990)
testing.tRunner+0x2f /home/rsc/g/go2/src/pkg/testing/testing.go:132
testing.tRunner(0xf764b710, 0xf763b5dc, 0x0)
runtime.goexit /home/rsc/g/go2/src/pkg/runtime/proc.c:149
runtime.goexit()
R=ken2, r
CC=golang-dev
https://golang.org/cl/4000053
If the same directory was used for multiple builds,
it was possible for a stale version.go to contain the
wrong definitions for $GOOS and $GOARCH, because
they can change even if the hg version does not.
Split into multiple files to fix.
R=r, r2
CC=golang-dev
https://golang.org/cl/4124050
Shame on me: I fixed the same bug in 6l in 8691fcc6a66e
(https://golang.org/cl/2609041) and neglected
to look at 5l and 8l to see if they were affected.
On the positive side, the check I added in that CL is the
one that detected this bug.
Fixes#1457.
R=ken2
CC=golang-dev
https://golang.org/cl/3981052
This aligns the naming scheme with the testing package and
also lets govet work on more logging calls.
R=rsc
CC=golang-dev
https://golang.org/cl/4001048
Make the error message and the needed action more obvious
when a command isn't found to obtain the source code
of a project. Users seem to strugle with the existing
wording in practice.
R=rsc
CC=golang-dev
https://golang.org/cl/4058047
The callback mechanism has been made more flexible.
Eliminated one round of argument copying in Syscall.
Faster Get/SetLastError implemented.
Added gettime for gc perf profiling.
R=rsc, brainman, mattn, rog
CC=golang-dev
https://golang.org/cl/4058046
The crypto package is added as a common place to store identifiers for
hash functions. At the moment, the rsa package has an enumeration of
hash functions and knowledge of their digest lengths. This is an
unfortunate coupling and other high level crypto packages tend to need
to duplicate this enumeration and knowledge (i.e. openpgp).
crypto pulls this code out into a common location.
It would also make sense to add similar support for ciphers to crypto,
but the problem there isn't as acute that isn't done in this change.
R=bradfitzgo, r, rsc
CC=golang-dev
https://golang.org/cl/4080046
The "all:" target is the default for running gomake
by hand, but it is not used during the build.
The build runs make install and make test.
Save the build of maketables for the test phase
so that the packages it needs will have been
installed already.
R=r, r2
CC=golang-dev
https://golang.org/cl/4121043
The sanity checking in pass 2 is wrong
when a select is offering to communicate in
either direction on a channel and neither case
is immediately ready.
R=ken2
CC=golang-dev
https://golang.org/cl/3991047
The old heap maps used a multilevel table, but that
was overkill: there are only 1M entries on a 32-bit
machine and we can arrange to use a dense address
range on a 64-bit machine.
The heap map is in bss. The assumption is that if
we don't touch the pages they won't be mapped in.
Also moved some duplicated memory allocation
code out of the OS-specific files.
R=r
CC=golang-dev
https://golang.org/cl/4118042
- complex numbers now supported.
- entirely independent of standard decode code.
- parser has no read-ahead; that is, the scanning works
simply by reading the values as they arrive, not by trying
to count bytes for message boundaries, a proof of concept
for the pending rewrite of the regular decoder.
R=rsc, r2
CC=golang-dev
https://golang.org/cl/4084044
It is unmaintained and untested, and I think it's broken too.
It was a toy to show that Go can run on real hardware,
and it served its purpose.
The source code will of course remain in the repository
history, so it could be brought back if needed later.
R=r, r2, uriel
CC=golang-dev
https://golang.org/cl/3996047
The implementation of the position computation
was surprisingly broken. Implemented fixes and
added extra test cases.
There is a slight interface change: Calling
Pos() returns the current position; but if
called before Scan() that position may not
be the position of the next token returned
by Scan() (depending on the scan settings
and the source text) - this in contrast to
the original comment.
However, after calling Scan(), the Scanner's
Position field reports the position of the
scanned token, as before.
Fixes#1327.
R=rsc
CC=golang-dev
https://golang.org/cl/3972047
Also simplify sleeper algorithm and poll
occasionally so redundant sleeper goroutines
will quit sooner.
R=r, niemeyer, r2
CC=golang-dev
https://golang.org/cl/4063043
The docstring claims the function uses PSS message encoding,
when the function actually implements PKCS1-v1_5 encoding.
R=agl1, rsc
CC=danderson, golang-dev
https://golang.org/cl/4097042
Init may report an error on the first character and
thus one needs an ability to set the error handler
for Init. Was a design bug.
Added corresponding test cases and better documentation.
Also: Fixed a subtle infinite loop exposed by one of the
new test cases.
Fixes#1380.
R=rsc, gri
CC=golang-dev
https://golang.org/cl/4094041
When parsing numbers with an exponent (like "12e-1"), the JSON scanner
would only allow a lowercase 'e', while the RFC also allows the
uppercase 'E'.
R=adg
CC=golang-dev, rsc
https://golang.org/cl/3986042
(or at least a correct encoder, still to come).
Change the debug structure slightly to better represent
the grammar.
Minor tweaks for consistency in type.go.
R=rsc
CC=golang-dev
https://golang.org/cl/4007044
The test code used to do this:
for _, tc := range tests {
ch <- &tc
}
Note that &tc is always the same value here. As the value is
received from the channel, the sender can loop around and
change the contents of tc. This means that the receiver's
value is unstable and can change while it is in use.
R=adg, r2, rsc
CC=chris, golang-dev
https://golang.org/cl/3978043
Close of closed channel panics.
Receive from closed channel never panics,
even if done repeatedly.
Fixes#1349.
Fixes#1419.
R=gri, iant, ken2, r, gri1, r2, iant2, rog, albert.strasheim, niemeyer, ejsherry
CC=golang-dev
https://golang.org/cl/3989042
Re-implement the debugging helper to be independent of the existing
implementation. This is preparatory to a rewrite to clean up issue 1416.
Include a definition of the grammar of the data stream.
R=rsc
CC=golang-dev
https://golang.org/cl/3970045
This was broken after the last update (2011-01-20).
However, I'm not sure if the changed example is a
sensible use of import(), so I'd appreciate comments.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/4067043
The functionality we want (shared ppid) is implied
by CLONE_THREAD already, and CLONE_PARENT
causes problems if the Go program is pid 1 (init).
See issue 1406 for more details.
Fixes#1406.
R=adg, iant
CC=golang-dev
https://golang.org/cl/3971044
The o+i*p approach to visiting select cases in random
order stops being fair when there is some case that
is never ready. If that happens, then the case that follows
it in the order gets more chances than the others.
In general the only way to ensure fairness is to make
all permutations equally likely. I've done that by computing
one explicitly.
Makes the permutations correct for n >= 4 where
previously they were broken. For n > 12, there's not
enough randomness to do a perfect job but this should
still be much better than before.
Fixes#1425.
R=r, ken2, ejsherry
CC=golang-dev
https://golang.org/cl/4037043
With the current implementation, xml unmarshalling
will silently fail to unmarshal any paths passing
through the same element, such as:
type T struct {
A string "dummy>a"
B string "dummy>b"
}
This change tweaks the algorithm so that this works
correctly.
Also, using paths that would cause the same element to
unmarshal twice will error out ahead of time explaining
the problem, rather than silently misbehaving.
R=rsc
CC=golang-dev
https://golang.org/cl/4082041
When data is received for a channel, but that channel
is not ready to receive it, the central run() loop
is currently blocked, but this can lead to deadlock
and interference between independent channels.
This CL adds an explicit buffer size to netchan
channels (an API change) - the sender will not
send values until the buffer is non empty.
The protocol changes to send ids rather than channel names
because acks can still be sent after a channel is hung up,
we we need an identifier that can be ignored.
R=r, rsc
CC=golang-dev
https://golang.org/cl/2447042
Also:
* document special cases for Frexp and Ldexp
* handle ±Inf in Ldexp
* correctly return -0 on underflow in Ldexp
* test special cases for Ldexp
* test boundary cases for Frexp, Ilogb, Ldexp, and Logb
R=rsc
CC=golang-dev
https://golang.org/cl/3676041
Minor cleanup:
- removed a duplicate test case
- added a function to remove repeated code
- for consistency, replaced "return nil" with a panic at an
unreachable point
Fixes#1428.
R=golang-dev, r, rsc
CC=golang-dev
https://golang.org/cl/4057042
RFC2616 sections 4.4 and 14.16:
* Cannot use Content-Length with non-identity Transfer-Encoding
* Content-Range response is "bytes x-y/z" not "x-y/z"
R=rsc
CC=golang-dev
https://golang.org/cl/4018041
For Windows, the options for syscall.Wait4() aren't used.
Then this will be dummy value like WNOHANG, WSTOPPED.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/4075041
This introduces support for selecting which subelement
to unmarshal into a given struct field by providing a
nesting path separated by the > character.
R=rsc
CC=golang-dev
https://golang.org/cl/4066041
Have to set #defines correctly to get correct value.
Otherwise get a mask for use in implementing WIFSTOPPED(status).
Changed WSTOPPED definition in os because on
OS X WSTOPPED and WUNTRACED have different values
even though they seem to mean the same thing.
Fixes#1374.
R=r, r2
CC=golang-dev
https://golang.org/cl/4063042
The fault was lucky: when it wasn't faulting it was silently
copying a word from some other block and later putting
that same word back. If some other goroutine had changed
that word of memory in the interim, too bad.
The ARM code was inconsistent about whether the
"argument frame" included the saved LR. Including it made
some things more regular but mostly just caused confusion
in the places where the regularity broke. Now the rule
reflects reality: argp is always a pointer to arguments,
never a saved link register.
Renamed struct fields to make meaning clearer.
Running ARM in QEMU, package time's gotest:
* before: 27/58 failed
* after: 0/50
R=r, r2
CC=golang-dev
https://golang.org/cl/3993041
In this specific package crosscall2 is already defined in a .S
file anyhow. This avoids a warning about mismatched
alignment.
R=rsc
CC=golang-dev
https://golang.org/cl/4000043
Introduced a printer mode (pmode) type and corresponding
pmode values which permit easy toggling of the current
printer mode for fine-tuning of layout.
Use the printer mode to disable potential introduction of
line breaks before a closing '}' in composite literals.
Added extra test case. Applied gofmt to src and misc.
Fixes#1365.
R=r2
CC=golang-dev
https://golang.org/cl/4008041
Until now, each scan of a file added a new file to the file set.
With this change, a file can be re-scanned using the same *token.File
w/o changing the file set. Eventually this will enable the re-use of
cached source code in godoc (for the fulltext index). At the moment,
source files are read over and over again from disk.
This is the first step in that direction.
R=r, rsc, r2
CC=golang-dev
https://golang.org/cl/4001041
Also adds some tests for Issue 900 which was the reason
the current URL parsing is broken. (the previous fix
was wrong)
R=rsc, adg, dangabrad, bradfitzwork
CC=golang-dev
https://golang.org/cl/3910042
This implements the algorithm qsufsort using the sort package
as a sorting primitive. Its worst-case performance is O(N*log(N)), and it
uses only an additional slice of N ints of memory during creation.
Benchmarks (seconds):
old new
10k nulls 149 0.044
1M English corpus 32.0 3.6
R=gri, gri1
CC=golang-dev
https://golang.org/cl/3752044
error is needed by all the OpenPGP packages as they return a shared
family of error types.
armor implements OpenPGP armoring. It's very like PEM except:
a) it includes a CRC24 checksum
b) PEM values are small (a few KB) and so encoding/pem assumes that
they fit in memory. Armored data can be very large and so this
package presents a streaming interface.
R=r, nsz, rsc
CC=golang-dev
https://golang.org/cl/3786043
I missed that environment is used during runtime setup,
well before go init() functions run. Implemented os-dependent
runtime.goenvs functions to allow for different unix, plan9 and
windows versions of environment discovery.
R=rsc, paulzhol
CC=golang-dev
https://golang.org/cl/3787046
In the case where r.Method == "POST", was
calling Printf with an argument but "" format string,
causing a spurious %!EXTRA(...) message.
Also escape string properly in HTML generation.
R=r
CC=golang-dev
https://golang.org/cl/3923043
A few system call numbers on x86 Linux are
defined in terms of a previous definition,
e.g.,
#define __NR_timer_create 259
#define __NR_timer_settime (__NR_timer_create+1)
...
#define __NR_mq_open 277
#define __NR_mq_unlink (__NR_mq_open+1)
This change assumes the numbers are sorted
sequentially in the input file.
R=rsc, bradfitzwork
CC=golang-dev
https://golang.org/cl/3946041
The After code is trivially generalisable to provide support
for this, and it is possible to use AfterFunc to do
things that After cannot, such as waiting
for many events at varied times without an overhead
of one goroutine per event.
R=rsc, r
CC=golang-dev
https://golang.org/cl/3905041
- added position information for [ and ] brackets of Index and Slice expression nodes
- removed a TODO in go/printer
R=r, rsc
CC=golang-dev
https://golang.org/cl/3867045
If we don't do this, then when C code calls back to Go code
which panics, we lose space on the scheduler stack. If that
happens a lot, eventually there is no space left on the
scheduler stack.
R=rsc
CC=golang-dev
https://golang.org/cl/3898042
Not absolutely certain it's right to do this, but since there's
no error value coming back, it seems reasonable.
Fixes#1392.
R=rsc, adg
CC=golang-dev
https://golang.org/cl/3896042
A cursory reading of the cgo code suggests this
should be necessary, though I don't have access
to a FreeBSD machine for testing.
R=rsc, adg
CC=golang-dev
https://golang.org/cl/3746047
The old loop was a bit odd; change it to be more regular.
This also enables a diagnostic for Printf("%", 3): %!(NOVERB)
R=rsc, Kyle C
CC=golang-dev
https://golang.org/cl/3749044
- add End() method to all nodes; the text range of a node n is [n.Pos(), n.End())
- various small bug fixes in the process
- fixed several comments
R=r, rsc
CC=golang-dev
https://golang.org/cl/3769042
After a prefix match, the old code advanced the length of the
prefix. This is incorrect since the full match might begin
in the middle of the prefix. (Consider "aaaab+" matching
"aaaaaab").
Fixes#1373
R=rsc
CC=golang-dev
https://golang.org/cl/3795044
The bug was that for an anchored pattern such as ^x, the prefix
scan ignored the anchor, and could scan the whole file if there was
no x present. The fix is to do prefix matching after the anchor;
the cost miniscule; the speedups huge.
R=rsc, gri
CC=golang-dev
https://golang.org/cl/3837042
Implementation uses fast suffixarray lookup to find
initial matches if the regular expression starts with
a suitable prefix without meta characters.
R=r, rsc
CC=golang-dev
https://golang.org/cl/3720042
#pragma dynexport is no longer needed for
this use of cgo, since the gcc and gc code are
now linked together into the same binary.
It may still be necessary later.
On the Mac, you cannot use the GOT to resolve
symbols that exist in the current binary, so 6l and 8l
translate the GOT-loading mov instructions into lea
instructions.
On ELF systems, we could use the GOT for those
symbols, but for consistency 6l and 8l apply the
same translation.
The translation is sketchy in the extreme
(depending on the relocation being in a mov
instruction) but it verifies that the instruction
is a mov before rewriting it to lea.
Also makes typedefs global across files.
Fixes#1335.
Fixes#1345.
R=iant, r
CC=golang-dev
https://golang.org/cl/3650042
The former is a boolean function to test whether a string
contains a regular expression metacharacter; the second
returns the string used to compile the regexp.
R=gri, rsc
CC=golang-dev
https://golang.org/cl/3728041
It was only used by exp/4s, and even if it is general purpose, I think
it belongs in a graphics library atop exp/draw, not in exp/draw itself.
R=rsc
CC=golang-dev
https://golang.org/cl/3705041
This is largely based on ality's CL 2747042.
crypto/rc4: API break in order to conform to crypto/cipher's
Stream interface
cipher/cipher: promote to the default build
Since CBC differs between TLS 1.0 and 1.1, we downgrade and
support only 1.0 at the current time. 1.0 is what most of the
world uses.
Given this CL, it would be trival to add support for AES 256,
SHA 256 etc, but I haven't in order to keep the change smaller.
R=rsc
CC=ality, golang-dev
https://golang.org/cl/3659041
The recent linker changes broke NaCl support
a month ago, and there are no known users of it.
The NaCl code can always be recovered from the
repository history.
R=adg, r
CC=golang-dev
https://golang.org/cl/3671042
This Flush is equivalent to zlib's Z_SYNC_FLUSH.
The addition of the explicit Writer type opens the
door to adding a PartialFlush if needed for SSH
and maybe even FullFlush. It also opens the door
for a SetDictionary method to be added.
http://www.bolet.org/~pornin/deflate-flush.html
documents the various intricacies of flushing a
DEFLATE stream.
R=agl1, r
CC=golang-dev
https://golang.org/cl/3637041
The code used interfaces in a pretty, pedagogical way but not efficiently.
Remove unnecessary interface code for significant speedups.
Before:
regexp.BenchmarkLiteral 1000000 2629 ns/op
regexp.BenchmarkNotLiteral 100000 18131 ns/op
regexp.BenchmarkMatchClass 100000 26647 ns/op
regexp.BenchmarkMatchClass_InRange 100000 27092 ns/op
regexp.BenchmarkReplaceAll 100000 27014 ns/op
After:
regexp.BenchmarkLiteral 1000000 2077 ns/op
regexp.BenchmarkNotLiteral 100000 13738 ns/op
regexp.BenchmarkMatchClass 100000 20418 ns/op
regexp.BenchmarkMatchClass_InRange 100000 20999 ns/op
regexp.BenchmarkReplaceAll 100000 21825 ns/op
There's likely more to do without major surgery, but this is a simple, significant step.
R=rsc
CC=golang-dev
https://golang.org/cl/3572042
Too many programs complain that we even try.
This was a bit of security paranoia and not worth
the bother.
Fixes#1340.
R=r, r2
CC=golang-dev
https://golang.org/cl/3579042
cc: same
runtime: test cc alignment (required moving #define of offsetof to runtime.h)
fix bug260
Fixes#482.
Fixes#609.
R=ken2, r
CC=golang-dev
https://golang.org/cl/3563042
- change Walk signature to use an ast.Node instead of interface{}
- add Pos functions to a couple of ast types to make them proper nodes
- explicit nil checks where a node can be nil; incorrect ASTs cause Walk to crash
For now ast.Walk is exercised extensively as part of godoc's indexer;
so we have some confidence in its correctness. But this needs a test,
eventually.
Fixes#1326.
R=rsc, r
CC=golang-dev
https://golang.org/cl/3481043
This patch adds a new package: os/inotify, which
provides a Go wrapper to the Linux inotify system.
R=rsc, albert.strasheim, rog, jacek.masiulaniec
CC=golang-dev
https://golang.org/cl/2049043
At the moment, and for the forseeable future, it only checks arguments to print calls.
R=rsc, gri, niemeyer, iant2, rog, lstoakes, jacek.masiulaniec, cw
CC=golang-dev
https://golang.org/cl/3522041
One of my own experiments ended up getting mistakenly commited when
switching to Jacobian transformations.
R=rsc
CC=golang-dev
https://golang.org/cl/3473044
Formerly known as libcgo.
Almost no code here is changing; the diffs
are shown relative to the originals in libcgo.
R=r
CC=golang-dev
https://golang.org/cl/3420043
The name ParseFile was kept as it indicates that
this function parses a Go 'SourceFile' per the
Go spec. Similarly, the other functions (ParseExpr,
ParseStmtList, ParseDeclList) parse the corresponding
constructs as defined in the Go spec.
Fixes#1311.
R=r, rsc
CC=golang-dev
https://golang.org/cl/3453042
I have written a tool to verify Printf calls, and although it's not
ready to be reviewed yet it's already uncovered a spate of problems
in the repository. I'm sending this CL to break the changes into
pieces; as the tool improves it will find more, I'm sure.
R=rsc
CC=golang-dev
https://golang.org/cl/3427043
Specifically:
* lib/godoc:
- provide file set (FSet) argument to formatters where needed
* src/cmd:
- cgo, ebnflint, godoc, gofmt, goinstall: provide file set (fset) where needed
- godoc: remove local binary search with sort.Search (change by rsc),
extract file set for formatters
* src/pkg:
- exp/eval: remove embedded token.Position fields from nodes and replace
with named token.Pos fields; add corresponding Pos() accessor methods
- go/token: added file.Line(), changed signature of File.Position()
* test/fixedbugs/:
- bug206.go: change test to not rely on token.Pos details
* added various extra comments
* Runs all.bash
* gofmt formats all of src, misc w/o changes
* godoc runs
* performance:
- The new version of godoc consumes about the same space after indexing
has completed, but indexing is half the speed. Significant space savings
are expected from smaller ASTs, but since they are thrown away after a
file has been indexed, this is not visible anymore. The slower indexing
time is due to the much more expensive computation of line information.
However, with the new compressed position information, indexing can be
rewritten and simplified. Furthermore, computing the line info can be
done more efficiently.
New godoc, immediately after indexing completed (best of three runs):
PID COMMAND %CPU TIME #TH #PRTS #MREGS RPRVT RSHRD RSIZE VSIZE
44381 godoc 0.0% 0:38.00 4 19 149 145M 184K 148M 176M
2010/12/03 17:58:35 index updated (39.231s, 18505 unique words, 386387 spots)
2010/12/03 17:58:35 bytes=90858456 footprint=199182584
2010/12/03 17:58:36 bytes=47858568 footprint=167295224
Old godoc, immediately after indexing completed (best of three runs):
PID COMMAND %CPU TIME #TH #PRTS #MREGS RPRVT RSHRD RSIZE VSIZE
23167 godoc 0.0% 0:22.02 4 17 132 129M 184K 132M 173M
2010/12/03 14:51:32 index updated (24.892s, 18765 unique words, 393830 spots)
2010/12/03 14:51:32 bytes=66404528 footprint=163907832
2010/12/03 14:51:32 bytes=46282224 footprint=163907832
The different numbers for unique words/spots stem from the fact the the
two workspaces are not exactly identical. The new godoc maintains a large
file set data structure during indexing which (probably) is the reason
for the larger heap (90858456 vs 66404528) before garbage collection.
R=rsc, r
CC=golang-dev
https://golang.org/cl/3050041
8l was broken by commit 7ac0d2eed9, it caused .data to be page aligned in the file - which is not how Plan 9 expects things to be.
Also .rodata was layed out in a similar fashion.
Not sure when signame was introduced, but added a stub.
Removed the symo assignment in asm.c as it is not currently used.
Fix runtime breakage after commit 629c065d36 which prefixes all external symbols with runtime·.
R=rsc
CC=golang-dev
https://golang.org/cl/2674041
Note:
* Exp2 doesn't have a special case for very small arguments
* Exp2 hasn't been subject to a proper error analysis
Also:
* add tests for Exp2 with integer argument
* always test Go versions of Exp and Exp2
R=rsc
CC=Charlie Dorian, PeterGo, golang-dev
https://golang.org/cl/3481041
fmt.Printf("%U", 1) yields "U+0001"
It's essentially "U+%.4x" but lets you override the precision works in scan, too.
R=rsc
CC=golang-dev
https://golang.org/cl/3423043
Doing the tail recursion elimination explicitly
seems safer than leaving it to the compiler;
the code is still clean and easy to understand.
R=r, r2, gri
CC=golang-dev
https://golang.org/cl/3373041
Before one could say
{field}
or
{field|formatter}
Now one can also say
{field1 field2 field3}
or
{field1 field2 field3|formatter}
and the fields are passed as successive arguments to the formatter,
analogous to fmt.Print.
R=rsc, gri
CC=golang-dev
https://golang.org/cl/3385041
Make them more like Printf, with a ... final argument. This breaks
code with existing formatters but not the templates that use them.
R=rsc, gri
CC=golang-dev
https://golang.org/cl/3378041
Changed all uses of bytes.Add (aside from those testing bytes.Add) to append(a, b...).
Also ran "gofmt -s" and made use of copy([]byte, string) in the fasta benchmark.
R=golang-dev, r, r2
CC=golang-dev
https://golang.org/cl/3302042
This means that any Writer can be used safely
even when Output is called concurrently.
Fixes#1302.
R=r, nigeltao_gnome
CC=golang-dev
https://golang.org/cl/3300042
Make Split work on backslashes as well as on slashes under Windows
and support the "C:filename" special case. Also add corresponding
tests.
R=r, rsc, PeterGo, r2, brainman
CC=golang-dev
https://golang.org/cl/3008041
This will make it easier to use Pos values
together with suffix arrays by slightly de-
coupling the mapping of Pos values to global
offsets.
R=rsc
CC=golang-dev
https://golang.org/cl/3231041
cipher is intended to replace crypto/block over time. This
change only adds basic parts: CBC and CTR mode and doesn't add
the package to the top-level Makefile.
R=r, rsc
CC=golang-dev
https://golang.org/cl/3069041
Backwards incompatible change, but makes
it easier to reason about non-idiomatic searches:
now f specifies what is sought.
R=gri
CC=golang-dev
https://golang.org/cl/3195042
Change comment to be more generic,
with indexed data structure search as
one common use case.
Fix typo []data.
R=gri, rog
CC=golang-dev
https://golang.org/cl/3159041
A pos value represents a file-set specific, accurate
source position value. It is 8x smaller in size than
the corresponding Position value (4 bytes vs 32 bytes).
Using Pos values instead of Position values in AST
saves approx. 25MBytes of memory when running godoc
on the current repository.
This CL introduces the Pos, File, and FileSet data
types; it does not affect existing code. Another
(pending CL) will make the change to all dependent
source files.
Missing: tests
R=r
CC=golang-dev, rsc
https://golang.org/cl/2936041
Cleaner, but also results in a 25%+ performance improvement for Get()/SetValue() on my machine.
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/3072041
The need for a LastIndexAny function has come up in the discussion
for https://golang.org/cl/3008041/. This function is
implemented analogously to lastIndexFunc, using functions from
the utf8 package.
R=r, rsc, PeterGo
CC=golang-dev
https://golang.org/cl/3057041
* Add support for certificate policy identifiers
* Fix the version number of generated certificates
* Fix the parsing of version numbers
* Fix the case of multiple name entries (it should have been a list of
tagged values, not a tagged list of values).
R=r
CC=golang-dev
https://golang.org/cl/3044041
When searching a list of directories, the files which match
the pattern are accumulated in a slice. If the glob has a
wildcard for the directory, and the wildcard matches a file
rather than a directory, then the files found so far are
discarded. E.g., path.Glob("*/x") in a directory which
contains both files and subdirectories. This patch avoids
discarding matches found so far when a file is found.
R=r
CC=bsiegert, golang-dev
https://golang.org/cl/3015042
This is in preparation for a different position representation.
It also resolves situations where a node would be printed as
it's node position simply because the embedded token.Position
has a String method.
R=r
CC=golang-dev
https://golang.org/cl/2991041
When it is known that there is already at least one element in the
list, it is awkwardly verbose to use three lines and an extra
variable declaration to remove the first or last item (a common
case), rather than use a simple expression.
a stack:
stk.PushFront(x)
x = stk.Front().Remove().(T)
vs.
stk.PushFront(x)
e := stk.Front()
e.Remove()
x = e.Value.(T)
[An alternative CL might be to add PopFront and PopBack methods].
R=gri
CC=golang-dev
https://golang.org/cl/3000041
As discussed in http://groups.google.com/group/golang-dev/browse_thread/thread/926b7d550d98ec9e,
add a simple "path expander" function, which returns all the
files matching the given pattern. This function is called Glob
after glob(3) in libc.
Also add a convenience function, hasMeta, that checks whether
a string contains one of the characters which are specially handled
by Match.
R=rsc, r, r2
CC=golang-dev
https://golang.org/cl/2476041
Previously we checked the certificate chain from the leaf
upwards and expected to jump from the last cert in the chain to
a root certificate.
Although technically correct, there are a number of sites with
problems including out-of-order certs, superfluous certs and
missing certs.
The last of these requires AIA chasing, which is a lot of
complexity. However, we can address the more common cases by
using a pool building algorithm, as browsers do.
We build a pool of root certificates and a pool from the
server's chain. We then try to build a path to a root
certificate, using either of these pools.
This differs from the behaviour of, say, Firefox in that Firefox
will accumulate intermedite certificate in a persistent pool in
the hope that it can use them to fill in gaps in future chains.
We don't do that because it leads to confusing errors which only
occur based on the order to sites visited.
This change also enabled SNI for tls.Dial so that sites will return
the correct certificate chain.
R=rsc
CC=golang-dev
https://golang.org/cl/2916041
Prefix all external symbols in runtime by runtime·,
to avoid conflicts with possible symbols of the same
name in linked-in C libraries. The obvious conflicts
are printf, malloc, and free, but hide everything to
avoid future pain.
The symbols left alone are:
** known to cgo **
_cgo_free
_cgo_malloc
libcgo_thread_start
initcgo
ncgocall
** known to linker **
_rt0_$GOARCH
_rt0_$GOARCH_$GOOS
text
etext
data
end
pclntab
epclntab
symtab
esymtab
** known to C compiler **
_divv
_modv
_div64by32
etc (arch specific)
Tested on darwin/386, darwin/amd64, linux/386, linux/amd64.
Built (but not tested) for freebsd/386, freebsd/amd64, linux/arm, windows/386.
R=r, PeterGo
CC=golang-dev
https://golang.org/cl/2899041
Previously all the functions took two arguments: src, dst. This is the
reverse of the usual Go style and worth changing sooner rather than
later.
Unfortunately, this is a change that the type system doesn't help
with. However, it's not a subtle change: any unittest worth the name
should catch this.
R=rsc, r
CC=golang-dev
https://golang.org/cl/2751042
CAST5 is the default OpenPGP cipher.
(This won't make Rob any happier about the size of crypto/, of course.)
It already has dst, src in that order but it doesn't have any users yet so I figure it's better than changing it later.
R=rsc, gri, r
CC=golang-dev
https://golang.org/cl/2762042
There's no need to hold the client mutex when calling encode, since encode itself
uses a mutex to make the writes atomic. However, we need to keep
the messages ordered, so add a mutex for that purpose alone.
Fixes#1244.
R=rsc
CC=golang-dev
https://golang.org/cl/2833041
First step towards a more light-weight implementation of token.Position:
- only use token.Position for reporting token and error position
- use offsets only for scanner control
- no interface changes yet
R=rsc
CC=golang-dev
https://golang.org/cl/2825041
1) Be sure to use the eval-time encoder/decoder rather than
the compile-time decoder. In a few cases the receiver for
the compiling encoder was being pickled incorrectly into a
closure.
(This is the fix for issue 1238).
2) Get the innermost name right when given a pointer to an
unnamed type.
3) Use a count to delineate interface values, making it
possible to ignore values without having a concrete type
to encode into. This is a protocol change but only for the
new feature, so it shouldn't affect anyone. The old test
worked because, amazingly, it depended on bug #1.
Fixes#1238.
R=rsc, albert.strasheim
CC=golang-dev
https://golang.org/cl/2806041
Rather than updating the stripped-down regexp implementation embedded
in testing, delete it by passing the one function we need from the package
main file created by gotest.
R=rsc
CC=golang-dev
https://golang.org/cl/2761043
Adds softfloat64 to generic runtime
(will be discarded by linker when unused)
and adds test for it. I used the test to check
the software code against amd64 hardware
and then check the software code against
the arm and its simulation of hardware.
The latter should have been a no-op (testing
against itself) but turned up a bug in 5c causing
the vlrt.c routines to miscompile.
These changes make the cmath, math,
and strconv tests pass without any special
accommodations for arm.
R=ken2
CC=golang-dev
https://golang.org/cl/2713042
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.
Fixes#1224.
R=rsc, dsymonds
CC=golang-dev
https://golang.org/cl/2685042
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.
R=rsc
CC=golang-dev
https://golang.org/cl/2660042
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.
R=rsc
CC=golang-dev
https://golang.org/cl/2618042
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.
R=r, r2
CC=golang-dev
https://golang.org/cl/2427042
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.
R=ken2
CC=golang-dev
https://golang.org/cl/2638041
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.
R=ken2
CC=golang-dev
https://golang.org/cl/2609041
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.
R=rsc
CC=golang-dev
https://golang.org/cl/2597041
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.
Fixes#871.
R=ken2
CC=golang-dev
https://golang.org/cl/2587041
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.
R=rsc
CC=golang-dev
https://golang.org/cl/2273041
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.
R=ken2, r2, ken3
CC=golang-dev
https://golang.org/cl/2487042
After a fill(), there is nothing to back up. Make sure UnreadRune
recognizes the situation.
Fixes#1137.
(Stops the crash, but doesn't make UnreadRune usable after a Peek()).
R=rsc
CC=golang-dev
https://golang.org/cl/2498041
* Maintain Sym* list for text with individual
prog lists instead of using one huge list and
overloading p->pcond.
* Comment what each file is for.
* Move some output code from span.c to asm.c.
* Move profiling into prof.c, symbol table into symtab.c.
* Move mkfwd to ld/lib.c.
* Throw away dhog dynamic loading code.
* Throw away Alef become.
* Fix printing of WORD instructions in 5l -a.
Goal here is to be able to handle each piece of text or data
as a separate piece, both to make it easier to load the
occasional .o file and also to make it possible to split the
work across multiple threads.
R=ken2, r, ken3
CC=golang-dev
https://golang.org/cl/2335043
Use a bytes.Buffer in log writing instead of string concatenation.
Should reduce the number of allocations significantly.
R=rsc, r2
CC=golang-dev
https://golang.org/cl/2417042
New logging interface simplifies and generalizes.
1) Loggers now have only one output.
2) log.Stdout, Stderr, Crash and friends are gone.
Logging is now always to standard error by default.
3) log.Panic* replaces log.Crash*.
4) Exiting and panicking are not part of the logger's state; instead
the functions Exit* and Panic* simply call Exit or panic after
printing.
5) There is now one 'standard logger'. Instead of calling Stderr,
use Print etc. There are now triples, by analogy with fmt:
Print, Println, Printf
What was log.Stderr is now best represented by log.Println,
since there are now separate Print and Println functions
(and methods).
6) New functions SetOutput, SetFlags, and SetPrefix allow global
editing of the standard logger's properties. This is new
functionality. For instance, one can call
log.SetFlags(log.Lshortfile|log.Ltime|log.Lmicroseconds)
to get all logging output to show file name, line number, and
time stamp.
In short, for most purposes
log.Stderr -> log.Println or log.Print
log.Stderrf -> log.Printf
log.Crash -> log.Panicln or log.Panic
log.Crashf -> log.Panicf
log.Exit -> log.Exitln or log.Exit
log.Exitf -> log.Exitf (no change)
This has a slight breakage: since loggers now write only to one
output, existing calls to log.New() need to delete the second argument.
Also, custom loggers with exit or panic properties will need to be
reworked.
All package code updated to new interface.
The test has been reworked somewhat.
The old interface will be removed after the new release.
For now, its elements are marked 'deprecated' in their comments.
Fixes#1184.
R=rsc
CC=golang-dev
https://golang.org/cl/2419042
Package iterable has outlived its utility.
It is an interesting demonstration, but it encourages
people to use iteration over channels where simple
iteration over array indices or a linked list would be
cheaper, simpler, and have fewer races.
R=dsymonds, r
CC=golang-dev
https://golang.org/cl/2436041
An exercise in reflection and an unusual tool.
From the usage message:
usage: gotry [packagedirectory] expression ...
Given one expression, gotry attempts to evaluate that expression.
Given multiple expressions, gotry treats them as a list of arguments
and result values and attempts to find a function in the package
that, given the first few expressions as arguments, evaluates to
the remaining expressions as results. If the first expression has
methods, it will also search for applicable methods.
If there are multiple expressions, a package directory must be
specified. If there is a package argument, the expressions are
evaluated in an environment that includes
import . "packagedirectory"
Examples:
gotry 3+4
# evaluates to 7
gotry strings '"abc"' '"c"' 7-5
# finds strings.Index etc.
gotry regexp 'MustCompile("^[0-9]+")' '"12345"' true
# finds Regexp.MatchString
R=rsc, PeterGo, r2
CC=golang-dev
https://golang.org/cl/2352043
Permits one to easily put a timeout in a select:
select {
case <-ch:
// foo
case <-time.After(1e6):
// bar
}
R=r, rog, rsc, sameer1, PeterGo, iant, nigeltao_gnome
CC=golang-dev
https://golang.org/cl/2321043
This crops up in a lot of places.
It's just a one-liner, but doesn't add any dependancies.
Seems worth it.
R=r, r2
CC=golang-dev
https://golang.org/cl/2344041
On systems where the mmap succeeds
(e.g., sysctl -w vm.mmap_min_addr=0)
it changes the signal code delivered for a
nil fault from ``page not mapped'' to
``invalid permissions for page.''
TBR=r
CC=golang-dev
https://golang.org/cl/2294041
There are a variety of token pairs that if printed
without separating blank may combine into a different
token sequence. Most of these (except for INT + .)
don't happen at the moment due to the spacing
introduced between expression operands. However, this
will prevent errors should the expression spacing
change.
R=rsc
CC=golang-dev
https://golang.org/cl/2207044
Not all OS make that the default.
Can finally do this now that the syscall package
has the right definitions.
Fixes#679.
R=r
CC=golang-dev
https://golang.org/cl/2204048
THIS WILL BREAK THE BUILD.
The z files have socketpair code in them that was
written by hand; breaking the build with this is the first
step in getting rid of that hand-written code.
R=adg
TBR=adg
CC=golang-dev
https://golang.org/cl/2197050
work on FreeBSD even without /usr/src/sys.
work on systems where gcc -static is broken.
TBR so I can test my semi-automated z builder.
TBR=adg
CC=golang-dev
https://golang.org/cl/2215046
Replace Marshal with MarshalToMemory
(no one was using old Marshal anyway).
Swap arguments to Unmarshal.
Fixes#1133.
R=agl1
CC=golang-dev
https://golang.org/cl/2249045
This is a replacement for pending CL 2219042. It only contains
the raw suffixarray functionality with two methods:
- New create a new index from some data
- Lookup lookup occurences of a bytes slice in the data
Any other functionality (dealing with multiple data sets and
the corresponding position lists) is generic and doesn't have
to be part of this package.
Known performance bug: This implementation works fine for data sets
up to several megabytes as long as it doesn't contain very long
contiguous sequences of equal bytes. For instance, index creation for
all .go files under GOROOT (250KLOCs, approx. 9MB) takes ~50s on
2.66 GHz Intel Xeon as long as test/fixedbugs/257.go is excluded.
With that file, index creation times takes several days. 257.go contains
a string of 1M smiley faces.
There are more sophisticated suffixarray creation algorithms which
can handle very long common prefixes. The implementation can be
updated w/o the need to change the interface.
R=rsc, r, PeterGo
CC=golang-dev
https://golang.org/cl/2265041
Fixes#1124.
Implementation is suggested by Skip.
Test is suggested by PeterGo.
R=r, PeterGo, rsc
CC=golang-dev, skip.tavakkolian
https://golang.org/cl/2256041
* Add documentation about array arguments. Fixes issue 1125.
* Do not interpret x, y := z, w as special errno form. Fixes issue 952.
* Fix nested Go calls (brainman). Fixes issue 907.
R=r
CC=golang-dev
https://golang.org/cl/2214044
Regenerate zsyscall_linux_*.go files with recent changes to
mksyscall.sh.
Add socketpair to syscall_linux_amd64.go; for some reason it
was in the generated file but not in the source file.
R=rsc, r2
CC=golang-dev
https://golang.org/cl/2190044
The key/value format of X.500 names means that it's possible to encode
a name with multiple values for, say, organisation. RFC5280
doesn't seem to consider this, but there are Verisign root
certificates which do this and, in order to find the correct
root certificate in some cases, we need to handle it.
Also, CA certificates should set the CA flag and we now check
this. After looking at the other X.509 extensions it appears
that they are universally ignored/bit rotted away so we ignore
them.
R=rsc
CC=golang-dev
https://golang.org/cl/2249042
asn1: add support for T61String because this is the string type which
several www.google.com certificates are now using for fields like
CommonName
tls: force a handshake in Dial so that certificates are ready
afterwards.
Fixes#1114.
R=rsc
CC=golang-dev
https://golang.org/cl/2216043
The decoder was crashing when handling an rpc that expected
a struct but was delivered something else. This diagnoses the
problem. The other direction (expecting non-struct but getting
one) was already handled.
R=rsc
CC=golang-dev
https://golang.org/cl/2246041
* SHOW DB responds with a
"110 n databases present - text follows" -- parse it.
* Doing a Define() on a non-existing word gives error
"invalid definition count: no", when we really
want "552 no match".
R=rsc
CC=golang-dev
https://golang.org/cl/2211041
The panic NaN was a translation error.
The earliest version said panic "return sys.NaN()",
and when sys.NaN came along, it changed
to "panic sys.NaN()" instead of "return sys.NaN()".
R=r
CC=golang-dev
https://golang.org/cl/2106049
g is not in r15 anymore.
now it's in a per-thread memory segment,
which is valid even inside a signal handler,
so we can just refer to g directly.
Fixes#1082.
R=r
CC=golang-dev
https://golang.org/cl/2149045
1) hack regalloc to leave R9 (m) and R10 (g) alone.
the real fix is tricker, but this gets us running
2) fix up the few places in the package sources that
the shortage of registers affects, by simplifying
some expressions.
all of this should be reverted when the right fix is in.
Fixes#1084.
R=rsc
CC=golang-dev
https://golang.org/cl/2132046
Add windows NOTEST list to pkg make file.
5a make file. the change removes a space
character that was included when appending
an extension to TARG.
R=brainman, rsc
CC=golang-dev
https://golang.org/cl/2140046
Need to correct for deferproc's extra 2 words of stack or in some
cases (such as memory profiling) traceback can cause a crash.
Also bulletproof the closure test.
R=rsc
CC=golang-dev
https://golang.org/cl/2138047
Old code was using recursion to traverse object graph.
New code uses an explicit stack, cutting the per-pointer
footprint to two words during the recursion and avoiding
the standard allocator and stack splitting code.
in test/garbage:
Reduces parser runtime by 2-3%
Reduces Peano runtime by 40%
Increases tree runtime by 4-5%
R=r
CC=golang-dev
https://golang.org/cl/2150042
Also add exporter.Drain() to wait for completion.
This makes it possible for an Exporter to fire off a message
and wait (by calling Drain) for the message to be received,
even if a client has yet to call to retrieve it.
Once this design is settled, I'll do the same for import send.
Testing strategies welcome. I have some working stand-alone
tests.
R=rsc
CC=golang-dev
https://golang.org/cl/2137041
This is a common task, so it makes sense to have a helper to do it.
(App Engine doesn't like "Transfer-Encoding: chunked" for POST
bodies, which is the default for regular Post.)
R=rsc
CC=golang-dev
https://golang.org/cl/2113041
It's unlikely to affect any extant code, but I wanted to make this API change
before digging in for a rewrite.
R=rsc
CC=golang-dev
https://golang.org/cl/2112041
Two tests start a goroutine which runs exportSend, and then
the tests run importReceive. exportSend creates an export
channel. importReceive asks to receive values on that
channel. Because exportSend runs in a separate goroutine,
it's possible for the export client to receive the request for
values on the channel, from importReceive, before the
goroutine actually creates the export channel. That causes an
error: "export: no such channel: exportedSend". This patch
avoids the race by creating the export channel before starting
the goroutine.
There does not seem to be a similar race condition in the
tests which send data in the other direction.
R=r
CC=golang-dev
https://golang.org/cl/2026045
- refine/define Scope, Object, and Type structures
(note: scope.go has the addition of types, the rest is only re-organized
for better readability)
- implemented top-level of type checker:
resolve global type declarations (deal with double decls, cycles, etc.)
- temporary hooks for checking of const/var declarations, function/method bodies
- test harness for fine-grained testing (exact error locations)
with initial set of tests
This is a subset of the code for easier review.
R=rsc
CC=golang-dev
https://golang.org/cl/1967049