(no progress in presence of syntax errors)
- end parsing early if source doesn't start
proper package clause
R=iant
DELTA=18 (7 added, 6 deleted, 5 changed)
OCL=27840
CL=27842
- separate parsing from execution
- rearrange code for organizational clarity
- provide execution errors and parse-time errors
- implement .or for repeated
TBR=rsc
OCL=27650
CL=27650
Imagine your var has a value of zero. If you have a goroutine calling Set(5),
and another calling Increment(+1), then you only want one of these outcomes:
- Set completes first, and then Increment occurs => 6
- Increment completes first, and then Set occurs => 5
However, you could get a sequence:
- read (for Increment) 0
- set (for Set) 5
- write (for Increment) 1
This results in a value of 1, which is undesirable.
Kudos to dnadasi for catching this.
R=r
APPROVED=r
DELTA=3 (3 added, 0 deleted, 0 changed)
OCL=27625
CL=27625
This handles integer-valued vars in a singleton struct, and exports functions
for incrementing, setting and getting those vars, as well as rendering all the
vars in a standard format.
Demonstrate the use of the exvar package in the http/triv server.
R=dcross,r
APPROVED=r
DELTA=122 (122 added, 0 deleted, 0 changed)
OCL=27617
CL=27622
exact spec:
a) must be a multi-assignment w :=
b) a proper subset of the lhs
can be declared in same block
with the same type with no
"redeclaration" error
R=r
OCL=27610
CL=27610
this allows gotest to find the locally built package when doing
make
gotest
without this option, one would have to say
make install
gotest
which kinda defeats the purpose
based on discussions with rsc.
R=ken,rsc
DELTA=12 (10 added, 1 deleted, 1 changed)
OCL=27606
CL=27606
Change the representation of errors in "os" to be cleaner.
(But they are not really representative of the power of the new scheme.)
Step 3 will be to remove all references to os.NewError.
Step 4 will be to delete the second half of lib/os/error.go.
R=rsc
OCL=27587
CL=27587
import (
"vector" -> "container/vector"
"ast" -> "go/ast"
"sha1" -> "hash/sha1"
etc.
)
and update Makefiles. Because I did the conversion
semi-automatically, I sorted all the import blocks
as a post-processing. Some files have therefore
changed that didn't strictly need to.
Rename local packages to lower case.
The upper/lower distinction doesn't work on OS X
and complicates the "single-package directories
with the same package name as directory name"
heuristic used by gobuild and godoc to create
the correlation between source and binary locations.
Now that we have a plan to avoid globally unique
names, the upper/lower is unnecessary.
The renamings will cause trouble for a few users,
but so will the change in import paths.
This way, the two maintenance fixes are rolled into
one inconvenience.
R=r
OCL=27573
CL=27575
for example, if building in src/lib/container,
objects go in obj/container/, so that 6g -Iobj
will find "container/vector".
install packages in hierarchy in $GOROOT.
this change only updates gobuild.
another change will have to update all
the sources to refer to "container/vector" etc
and regenerate all the Makefiles.
there are some pretty lame functions here
(e.g., Mkdir, Remove, the Getenv("PWD"))
but i will implement better ones in another CL.
R=r
DELTA=117 (99 added, 2 deleted, 16 changed)
OCL=27550
CL=27574
as a reminder, the old conversion
was that you could write
var arr [10]byte;
var slice []byte;
slice = arr;
but now you have to write
slice = &arr;
the change eliminates an implicit &, so that
the only implicit &s left are in the . operator
and in string(arr).
also, removed utf8.EncodeRuneToString
in favor of string(rune).
R=r
DELTA=83 (1 added, 23 deleted, 59 changed)
OCL=27531
CL=27534
6g generates for functions: expect the keyword "func". The
older reflection syntax, without the "func", is still
recognized for simplicity in parsing interface reflection
strings.
R=r,rsc
DELTA=66 (31 added, 8 deleted, 27 changed)
OCL=27396
CL=27422
add test for string range.
test has minor failure: after loop the index == len(s); should be len(s)-1
in this case. according to spec, vars are left at position at last
iteration.
R=ken,rsc
DELTA=259 (161 added, 96 deleted, 2 changed)
OCL=27343
CL=27343
The HTML one here is just a stub - should use an HTML library to do the right thing.
R=rsc
DELTA=54 (47 added, 2 deleted, 5 changed)
OCL=27250
CL=27250
store only the original import path string (+ .a)
if 6g resolves it to an archive file.
let 6l re-resolve the .a at link time.
this lets libraries build against an archive
in the current directory but get used
against an installed archive.
R=r
OCL=27244
CL=27244
use path.Clean in web server to sanitize URLs.
http://triv/go/../../../etc/passwd
no longer serves the password file.
it redirects to
http://triv/etc/passwd
which then gets a 404.
R=r
DELTA=288 (286 added, 0 deleted, 2 changed)
OCL=27142
CL=27152
string(b)[0:n]
instead of the more direct string(b[0:n]).
convert to the more direct form.
R=r
DELTA=5 (0 added, 0 deleted, 5 changed)
OCL=27082
CL=27140
* make([100]int) was being compiled to
make([]int), kind of.
* []this = [100]that was working for any this, that.
turned up a typo in pipe_test.go
R=ken
OCL=27081
CL=27081
was copying a bit too much about t into n,
like whether the signature was queued to be printed.
(bug reported by anton)
was also editing t, meaning you could do
type T int;
func (p int) Meth() { }
both fixed.
R=ken
OCL=27052
CL=27052
this avoids problems people have run into with
multiple closures in the same package.
when preparing filename, only cut off .go, not .anything.
this fixes a bug tgs ran into with foo.pb.go and foo.go
in the same package.
also turn bad identifier chars from filename into
underscores: a-b.pb.go => a_b_pb
R=ken
OCL=27050
CL=27050
(this removes a lot of string() conversions down the road)
- a few minor adjustments
R=rsc
DELTA=11 (0 added, 0 deleted, 11 changed)
OCL=27029
CL=27038
- have explicit XSpec nodes for declarations
- have a general GenDecl node instead of DeclList
R=rsc
DELTA=164 (52 added, 52 deleted, 60 changed)
OCL=27005
CL=27027
160 - 75 was just barely not enough for deferproc + morestack.
added enum names and bumped to 256 - 128.
added explanation.
changed a few mal() (garbage-collected) to
malloc()/free() (manually collected).
R=ken
OCL=26981
CL=26981
- minor adjustments as suggested by rsc
2) Added parser_test fragment
3) Renamed some types in AST.go per rsc request
R=rsc
DELTA=2053 (2027 added, 0 deleted, 26 changed)
OCL=26963
CL=26971
remove use of _subv in vlrt.c
darwin/386/signal.c
darwin/386/*
linux/386/* (forgotten before)
can run empty program on darwin/386 now.
R=r
DELTA=1140 (1021 added, 114 deleted, 5 changed)
OCL=26942
CL=26968
into gc directory, where it gets included as ../gc/ldbody
this is similar to the assemblers including ../cc/lexbody
and ../cc/macbody.
* hook go-specific loader code into 8l.
* make current 8.out.h and 6.out.h backward compatible
with plan 9's versions. i had added some constants in
the middle of enums and have now moved them to the end.
this keeps us from invalidating old .8 and .6 files.
not sure how much it really matters, but easy to do.
R=r
DELTA=1314 (667 added, 623 deleted, 24 changed)
OCL=26938
CL=26941
type of the field. Use the field alignment to compute the
size of a structure.
This may help 8g but is mainly for gccgo. gccgo maintains the
standard C/C++ ABI for structure field alignment. For the
i386, this requires that a float64 field in a struct be
aligned on a 32-bit boundary, although for efficiency a
variable of type float64 or []float64 should be aligned on a
64-bit boundary.
I also removed the unused size field from structField.
R=r
DELTA=117 (75 added, 2 deleted, 40 changed)
OCL=26842
CL=26936
uses gcc to determine system header layouts and
emits simple C or Go. see comment in main.c.
R=r
DELTA=1069 (1067 added, 0 deleted, 2 changed)
OCL=26682
CL=26880