1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
|
* Thu Aug 20 14:56:44 1998
agrep -v fails to write anything unless at least one line (possibly, a
particular line) of the input matches the pattern that we're trying to
mismatch. The non-glimpse agrep does not have this problem. E.g.:
thalia[189]$ for p in O e n w o r ; do echo "Pattern '${p}':" ; echo "One@Two@Three" | tr '@' '\012' | agrep -v ${p} ; done
Pattern 'O':
Pattern 'e':
Two
Pattern 'n':
Pattern 'w':
One
Pattern 'o':
One
Pattern 'r':
One
Two
thalia[190]$ for p in O e n w o r ; do echo "Pattern '${p}':" ; echo "One@Two@Three" | tr '@' '\012' | agrep-2.04 -v ${p} ; done
Pattern 'O':
Two
Three
Pattern 'e':
Two
Pattern 'n':
Two
Three
Pattern 'w':
One
Three
Pattern 'o':
One
Three
Pattern 'r':
One
Two
The non-glimpse agrep uses bitap to do the search; the glimpse one uses bm.
Some pre-condition is unsatisfied in the call to bm, because it overruns the
input text buffer by a huge amount in attempting to find a pattern match.
It isn't obvious to me where this bug is arising, or how to fix it.
|
