Bugs for Sorcery are reported here

[bugzilla-daemon AT bugs.sourcemage.org]

http://bugs.sourcemage.org/show_bug.cgi?id=13356





------- Additional Comments From jakakranjc AT email.si  2007-01-05 14:19 -------
Created an attachment (id=6578)
 --> (http://bugs.sourcemage.org/attachment.cgi?id=6578&action=view)
IFSless new real_list_find and extended unit test

performance wise it is only slower than the current implementation in my "100
string lenghts, one nonmatching search term" (17s vs 6s) and the "100 string
lenghts, five nonmatching search terms" worst-case test (57s vs 22s). Not too
bad, remember they are agregate times of 200 runs.
It shines in the last one (5s vs 86s).

It even comes out faster in the net-tools test - 14s vs 17s (the awk one) vs
62s (the original).

Then I tried something heavier, lots of queries and depends:
cast net-tools nano xorg-server libdrm

the averages of 3 hot runs:
60s vs 119s (new vs original)

So it looks pretty good. It seems all the small queries' speedup weights out
the big ones' slowdown and more.

But it actually didn't even get called with anything heavy, here is an uniq -c
on the list of calls with number of fields in $input and search tokens:
      2 --1*0--
      4 --1*1--
    106 --1*18--
    239 --1*41--
     30 --15*1--
     14 --15*2--
     76 --1*7--
     12 --2*1--
     71 --36*1--
     35 --36*2--

In the patch there are also extended unit tests - I fixed the aforementioned
two and added two more, testing newline and tab as delimiters.

-- 
Configure bugmail: http://bugs.sourcemage.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.