The Plan 9 symbol table format defines big-endian symbol values
for portability, but we want to be able to generate an ELF object file
and let the host linker link it, as part of the solution to issue 4069.
The symbol table itself, since it is loaded into memory at run time,
must be filled in by the final host linker, using relocation directives
to set the symbol values. On a little-endian machine, the linker will
only fill in little-endian values during relocation, so we are forced
to use little-endian symbol values.
To preserve most of the original portability of the symbol table
format, we make the table itself say whether it uses big- or
little-endian values. If the table begins with the magic sequence
fe ff ff ff 00 00
then the actual table begins after those six bytes and contains
little-endian symbol values. Otherwise, the table is in the original
format and contains big-endian symbol values. The magic sequence
looks like an "end of table" entry (the fifth byte is zero), so legacy
readers will see a little-endian table as an empty table.
All the gc architectures are little-endian today, so the practical
effect of this CL is to make all the generated tables little-endian,
but if a big-endian system comes along, ld will not generate
the magic sequence, and the various readers will fall back to the
original big-endian interpretation.
R=ken2
CC=golang-dev
https://golang.org/cl/7066043
A composite literal may be parenthesized when
used as operand for the unary operator &.
R=rsc, iant, r, ken
CC=golang-dev
https://golang.org/cl/6996053
CL6449105 changed godoc id attributes to ensure uniqueness.
This CL updates links to godoc pages in documents that used
the old id attributes.
R=golang-dev, dsymonds
CC=golang-dev, speter.go1
https://golang.org/cl/7015051
An error during the compilation can be more precise
than an error at link time.
For 'func init', the error happens always: you can't forward
declare an init func because the name gets mangled.
For other funcs, the error happens only with the special
(and never used by hand) -= flag, which tells 6g the
package is pure go.
The go command now passes -= for pure Go packages.
Fixes#3705.
R=ken2
CC=golang-dev
https://golang.org/cl/6996054
Enable cgo on OpenBSD.
The OpenBSD ld.so(1) does not currently support PT_TLS sections. Work
around this by fixing up the TCB that has been provided by librthread
and reallocating a TCB with additional space for TLS. Also provide a
wrapper for pthread_create, allowing zeroed TLS to be allocated for
threads created externally to Go.
Joint work with Shenghou Ma (minux).
Requires change 6846064.
Fixes#3205.
R=golang-dev, minux.ma, iant, rsc, iant
CC=golang-dev
https://golang.org/cl/6853059
This CL breaks Go 1 API compatibility but it doesn't matter because
previous ListenUnixgram doesn't work in any use cases, oops.
The public API change is:
-pkg net, func ListenUnixgram(string, *UnixAddr) (*UDPConn, error)
+pkg net, func ListenUnixgram(string, *UnixAddr) (*UnixConn, error)
Fixes#3875.
R=rsc, golang-dev, dave
CC=golang-dev
https://golang.org/cl/6937059
This is language change. It is a backward-compatible
change but for code that relies on a run-time panic
when calling delete on a nil map (unlikely).
Fixes#4253.
R=rsc, r, iant, ken, bradfitz, rogpeppe
CC=golang-dev
https://golang.org/cl/6909060
We have been using all three terms "indices", "indexes",
and "index expressions" indiscriminatly for index values.
With this change, "index" refers to an index value,
"indices" is the plural of "index", and "index expression"
refers to an array, slice, or map indexed by an index: a[x].
R=r, rsc, iant, ken, mtj
CC=golang-dev
https://golang.org/cl/6912056
New in Go 1 will be nanosecond precision in the result of time.Now on Linux.
This will break code that stores time in external formats at microsecond
precision, reads it back, and expects to get exactly the same time.
Code like that can be fixed by using time.Now().Round(time.Microsecond)
instead of time.Now() in those contexts.
R=golang-dev, bradfitz, iant, remyoudompheng
CC=golang-dev
https://golang.org/cl/6903050
The spec didn't preclude invalid type assertions and
type switches, i.e., cases where a concrete type doesn't
implement the interface type in the assertion in the first
place. Both, the gc and gccgo compiler exclude these cases.
This is documenting the status quo.
Also:
- minor clean up of respective examples
- added sentence about default case in select statements
Fixes#4472.
R=rsc, iant, r, ken
CC=golang-dev
https://golang.org/cl/6869050
Also:
- 'for' statements with a range clause do not accept send-only
channels
- '_, _ = range ch' is not equivalent to "_ = range ch" if ch
is a channel (rewriting the latter to the former leads to
an invalid range clause).
These clarifications document the status quo.
R=rsc, r, iant, ken
CC=golang-dev
https://golang.org/cl/6874053
1) Built-ins are restricted like for expression statements.
This documents the status quo.
2) Calls cannot be parenthesized. The spec is not clear. gccgo
permits it already, gc doesn't. Be explicit in the spec.
Fixes#4462.
R=rsc, iant, r, ken, lvd
CC=golang-dev
https://golang.org/cl/6861043
We have the notion of a PackageName, not package identifier.
As is, it could construed that imports that rename a package
don't have an "imported package identifier" but a local one.
R=r, rsc, iant, ken, dsymonds
CC=golang-dev
https://golang.org/cl/6858049
Do not skip the first symbol in the symbol table. Any other indexes
into the symbol table (for example, indexes in relocation entries)
will now refer to the symbol following the one that was intended.
Add an object that contains debug relocations, which debug/dwarf
failed to decode correctly. Extend the relocation tests to cover
this case.
Note that the existing tests passed since the symbol following the
symbol that required relocation is also of type STT_SECTION.
Fixes#4107.
R=golang-dev, mikioh.mikioh, iant, iant
CC=golang-dev
https://golang.org/cl/6848044
This is an attempt at making the interaction between
these three constructs clearer. Specifically:
- return statements terminate a function, execute deferred
functions, return to the caller, and then execution
continues after the call
- panic calls terminate a function, execute deferred
functions, return to the caller, and then re-panic
- deferred functions are executed before a function _returns_
to its caller
The hope is that with this change it becomes clear when a
deferred function is executed (when a function returns),
and when it is not (when a program exits).
R=r, rsc, iant, ken, iant
CC=golang-dev
https://golang.org/cl/6736071
Both gc and gccgo always checked this for constant
expressions but the spec only mentions run-time
exceptions.
This CL also requires that constant divisors
must not be zero in non-constant integer expressions:
This is consistent with the spirit of the most
recent changes and it is consistent with constant
expressions. We don't want to specify the effect for
non-integer expressions (f/0.0 where f is a float or
complex number) because there the result f/g is not
further specified if a non-constant g is 0.
R=r, rsc, iant, ken, andybalholm, iant
CC=golang-dev
https://golang.org/cl/6710045
This is a language change: Until now, the spec required run-
time panics for some of these errors. Note that gc and gccgo
implemented this inconsistently, and that both compilers already
reported compile-time errors in some cases. This change makes
make() behave along the same vein as index expressions.
This addresses the spec aspect of issue 4085.
R=r, rsc, iant, ken
CC=golang-dev
https://golang.org/cl/6725053
