CommonLisp标识符统计

最近在尝试实现CommonLisp，想起CL的常用操作符car, cdr, defparameter之类，如果把它们改短些，输入时不就省事么？不过动手前先找准目标比较妥当，于是写了个Python脚本，统计ClozureCL、CLISP中出现的所有标识符，挑出排名靠前的看看。

# 统计CommonLisp实现中，最常用的标识符及其出现的频率

import string, os, sys, re;

def stat_file(file_path, dic):
  max_id_len = 0;
  fp = open(file_path);
  try:
    for line in fp.readlines():
      ids = re.findall("\w+[a-zA-Z0-9-_]*", line);
      for id in ids:
        l = len(id)
        if l > max_id_len:
          max_id_len = l
        dic[id] = dic.get(id,0) + 1;
  except:
    None
  return max_id_len

base_dirs = ["C:\wz\greenware\ccl-1.7" , "C:\wz\greenware\clisp-2.49"]

dic_id_cnt = {}
max_id_len = 0

for base_dir in base_dirs:
  for root, dirs, files in os.walk(base_dir):
    for file_name in files:
      if re.match(".*\.lisp$", file_name) != None and re.search("\.svn", file_name) == None:
        file_path = os.path.join(root, file_name)
        print(file_path);
        tm = stat_file(file_path, dic_id_cnt);
        if tm > max_id_len:
          max_id_len = tm

ml = max_id_len + 1
tuples = sorted(dic_id_cnt.items(), key=lambda d:(99999999 - d[1]));
n = 0
for e in tuples:
  print(e[0].ljust(20) + str(e[1]).rjust(8)); # 虽计算了最长标识，但前100位都较短
  n += 1
  if n > 100:
    break;

#exit();

结果和空想的完全不同——七个核心操作符并不是最热的（如果冒然改解析器就亏大了）：

the                    25622
if                     17399
let                    15485
nil                    14565
and                    12769
0                      12449
defun                  11855
a                      11716
t                      10312
is                      9790
1                       9649
of                      9405
to                      9294
when                    8918
or                      7997
setf                    7791
setq                    7688
name                    7195
s                       6686
in                      6640
i                       6408
declare                 5964
eq                      5714
not                     5586
x                       5573
type                    5558
fixnum                  5395
form                    5379
seg                     4996
stream                  4905
list                    4836
for                     4668
string                  4334
unless                  4329
dest                    4043
car                     4009
value                   3975
imm0                    3949
n                       3657
2                       3598
p                       3594
l                       3582
target                  3580
buffer                  3478
u                       3421
be                      3409
that                    3403
cdr                     3395
class                   3301
start                   3288
as                      3249
arg_z                   3245
S                       3155
end                     3130
ppc                     3093
args                    2991
values                  2970
key                     2948
it                      2947
length                  2783
return                  2777
with                    2753
src                     2745
vreg                    2736
logior                  2722
null                    2701
function                2678
error                   2650
index                   2639
The                     2603
object                  2578
val                     2566
file                    2522
code                    2507
do                      2500
arm                     2475
4                       2468
this                    2407
eql                     2401
rest                    2397
ash                     2394
optional                2366
progn                   2360
format                  2343
y                       2342
x8664                   2339
b                       2310
cond                    2277
8                       2274
body                    2243
apply                   2235
line                    2221
vector                  2218
result                  2209
lambda                  2199
are                     2198
3                       2197
idx                     2188
lisp                    2161
defconstant             2128
x8632                   2121