now that OCOMP exists

R=ken
OCL=22216
CL=22216

R=r
OCL=22215
CL=22215

R=rsc
DELTA=3  (1 added, 0 deleted, 2 changed)
OCL=22195
CL=22207

- now have struct, array, slice, and map literals

DELTA=34  (13 added, 6 deleted, 15 changed)
OCL=22180
CL=22204

no more surprises - all caught up

R=rsc
DELTA=4  (0 added, 0 deleted, 4 changed)
OCL=22194
CL=22194

R=r
OCL=22187
CL=22187

R=r
OCL=22185
CL=22185

R=r
OCL=22182
CL=22182

  (this code doesn't match the existing language at this point,
  but it's a large code base which compiles - will eventually go
  away)
- enable compilation of it again in run.bash

R=r
DELTA=1147  (534 added, 311 deleted, 302 changed)
OCL=22176
CL=22176

(the text still needs fixing)
add the tutorial programs to the test run.

R=rsc
DELTA=41  (6 added, 0 deleted, 35 changed)
OCL=22174
CL=22174

R=ken
OCL=22173
CL=22173

R=r
OCL=22171
CL=22171

R=rsc
OCL=22170
CL=22170

R=ken
OCL=22168
CL=22168

R=r
OCL=22166
CL=22166

R=r
OCL=22160
CL=22160

  (will help produce html output where we can click on identifiers and
  get to the declaration)
- snapshot before changing back to old new

R=r
OCL=22159
CL=22159

new for pointers

R=r
OCL=22158
CL=22158

indexes, and a global boolean function for quick action.

R=rsc
DELTA=152  (127 added, 12 deleted, 13 changed)
OCL=22140
CL=22142

R=rsc
DELTA=40  (18 added, 22 deleted, 0 changed)
OCL=22134
CL=22139

- adding "init()"
- fixing some bugs with slice documentation

DELTA=118  (45 added, 7 deleted, 66 changed)
OCL=22084
CL=22136

,s/int32/long/g in 8a, 8c, 8l.
delete dead code.
move enam.c, 8.out.h, mkenam from 8c to 8l.

R=r
DELTA=1850  (581 added, 983 deleted, 286 changed)
OCL=22119
CL=22129

remove unused second parameter.

R=ken
OCL=22126
CL=22126

R=r
DELTA=3562  (0 added, 3560 deleted, 2 changed)
OCL=22112
CL=22114

ask for bison explicitly in cc, gc to try to
avoid problems with other yaccs that
might be installed.

R=r
DELTA=29  (10 added, 2 deleted, 17 changed)
OCL=22110
CL=22113

R=r
DELTA=19539  (19539 added, 0 deleted, 0 changed)
OCL=22109
CL=22109

R=ken
OCL=22078
CL=22080

R=rsc
DELTA=10  (1 added, 2 deleted, 7 changed)
OCL=22033
CL=22033

now passes even under loaded conditions on r45.

R=r
DELTA=2  (0 added, 0 deleted, 2 changed)
OCL=22019
CL=22022

was incrementing lineno twice for
the \n after a // comment.

R=r
DELTA=3  (0 added, 2 deleted, 1 changed)
OCL=21984
CL=22021

Instead of trying to make open arrays look like arrays (which they aren't,
they just look like them), I decided to call them "slice types". As a result
they have value semantics, and one can really talk about what they really are.

Note:
- There are lots of missing details
- There are probably lots of mistakes

Looking for some initial feedback.

R=r
DELTA=444  (180 added, 189 deleted, 75 changed)
OCL=21769
CL=22020

R=ken
DELTA=20  (20 added, 0 deleted, 0 changed)
OCL=21889
CL=21929

has to do with printing
patched it out

R=r
OCL=21921
CL=21923

receivers in methods.
put back original code for hilbert/bignum

R=r
OCL=21910
CL=21920

rvalue promoted to method receiver.
fixes to bignum that failed.

R=r
OCL=21827
CL=21827

TBR=rsc
OCL=21766
CL=21766

TBR=rsc
DELTA=34  (33 added, 0 deleted, 1 changed)
OCL=21718
CL=21718

it's 25% faster and runs with 40% less memory allocation than before

R=rsc
DELTA=20  (15 added, 0 deleted, 5 changed)
OCL=21690
CL=21690

lines marked BUG are rewrites working around 6g bug.

R=rsc
DELTA=161  (42 added, 2 deleted, 117 changed)
OCL=21689
CL=21689

make sure type of expr is T not just []int

R=ken
OCL=21688
CL=21688