R=r
DELTA=13  (5 added, 6 deleted, 2 changed)
OCL=34337
CL=34343

R=ken
OCL=34339
CL=34341

R=r
DELTA=42  (21 added, 20 deleted, 1 changed)
OCL=34328
CL=34333

R=rsc
DELTA=20  (20 added, 0 deleted, 0 changed)
OCL=34330
CL=34332

main changes:
	gccgo is much slower on  binary-tree and fannkuch
	6g seems to struggle with alignment of loops in nbody (dignosis uncertain)
	6g is a little faster at regexps and concurrency

R=rsc
DELTA=93  (93 added, 0 deleted, 0 changed)
OCL=34320
CL=34322

R=austin
DELTA=26  (0 added, 26 deleted, 0 changed)
OCL=34294
CL=34317

exit is still handled poorly.

R=rsc
APPROVED=rsc
DELTA=6  (6 added, 0 deleted, 0 changed)
OCL=34288
CL=34315

R=rsc
DELTA=15  (3 added, 2 deleted, 10 changed)
OCL=34307
CL=34314

R=r
DELTA=7  (0 added, 4 deleted, 3 changed)
OCL=34310
CL=34312

R=ken
DELTA=9  (0 added, 0 deleted, 9 changed)
OCL=34309
CL=34311

R=rsc
DELTA=217  (216 added, 0 deleted, 1 changed)
OCL=34279
CL=34306

R=gri
DELTA=3  (0 added, 3 deleted, 0 changed)
OCL=34305
CL=34305

added iant's bug202 (in main code)
and ken's bug203 (in init function).
bug187 remains at large.

R=ken
OCL=34293
CL=34293

APPROVED=rsc
DELTA=1  (0 added, 0 deleted, 1 changed)
OCL=34289
CL=34289

R=austin
DELTA=142  (128 added, 4 deleted, 10 changed)
OCL=34275
CL=34284

R=ken
DELTA=3  (0 added, 0 deleted, 3 changed)
OCL=34283
CL=34283

comment out incomplete stuff

R=rsc
DELTA=61  (58 added, 0 deleted, 3 changed)
OCL=34278
CL=34281

reuse Thread in function calls.

R=austin
DELTA=59  (8 added, 7 deleted, 44 changed)
OCL=34266
CL=34266

TBR=gri
OCL=34264
CL=34264

R=rsc
DELTA=196  (194 added, 0 deleted, 2 changed)
OCL=34234
CL=34263

(was waiting until after release)

R=austin
DELTA=5  (5 added, 0 deleted, 0 changed)
OCL=34222
CL=34261

change eval functions from taking *Frame to *Thread

R=austin
DELTA=500  (7 added, 4 deleted, 489 changed)
OCL=34256
CL=34260

R=rsc
APPROVED=rsc
DELTA=31  (0 added, 4 deleted, 27 changed)
OCL=34251
CL=34259

R=r
OCL=34253
CL=34253

R=ken
OCL=34244
CL=34249

(6g had a few pretty major bugs
where it was more lax than it should
have been when checking when things
satisfied interfaces; i fixed them and
this turned up.)

R=r
DELTA=2  (0 added, 0 deleted, 2 changed)
OCL=34243
CL=34248

R=austin
DELTA=959  (404 added, 99 deleted, 456 changed)
OCL=34214
CL=34237

R=ken
OCL=34232
CL=34232

R=rsc
APPROVED=rsc
DELTA=1003  (1003 added, 0 deleted, 0 changed)
OCL=34223
CL=34227

R=r
APPROVED=r
DELTA=1  (0 added, 0 deleted, 1 changed)
OCL=34216
CL=34220

  * do full lookup for {.section},
  	so that it is always allowed to replace
	{Foo} with {.section Foo}{@}{.end}
  * treat False as empty so that .section can
  	be used to test bools

R=r
DELTA=29  (21 added, 0 deleted, 8 changed)
OCL=34215
CL=34219

R=austin
DELTA=17  (0 added, 16 deleted, 1 changed)
OCL=34213
CL=34218

interface, but implement it better this time.  Also, work
around a 6g bug in asInterface.

R=rsc
APPROVED=rsc
DELTA=39  (13 added, 1 deleted, 25 changed)
OCL=34211
CL=34217

R=austin
DELTA=958  (956 added, 0 deleted, 2 changed)
OCL=34180
CL=34212

only visible change is that evaluating an expression returns a
interface{} instead of a Value.

R=rsc
APPROVED=rsc
DELTA=60  (15 added, 26 deleted, 19 changed)
OCL=34206
CL=34208

R=ken
OCL=34205
CL=34207

R=rsc,r
DELTA=4  (3 added, 0 deleted, 1 changed)
OCL=34204
CL=34204

R=rsc
DELTA=186  (181 added, 0 deleted, 5 changed)
OCL=34203
CL=34203

for decoding closures.  There is now no notion of a current OS
thread, though that needs to come back in the form of a
current Go thread.  As a result, Process now implements Peek
and Poke and maps them to any stopped OS thread, since they
all share the address space anyways.

R=rsc
APPROVED=rsc
DELTA=322  (310 added, 3 deleted, 9 changed)
OCL=34136
CL=34201

R=rsc
DELTA=44  (10 added, 30 deleted, 4 changed)
OCL=34195
CL=34200