[ksh93-integration-discuss] [osol-code] extremly badperformance of Solaris regex

[Roland Mainz]

Jens Elkner wrote:
[Please keep ksh93-integration-discuss at opensolaris.org in the CC: field
for AST-related topics since not all people there are subscribed to all
other opensolaris-org lists, too]
> On Wed, Oct 31, 2007 at 01:04:55PM -0400, Glenn Fowler wrote:
> > > Jens Elkner wrote:
> > > >                  regex.c                          RegexTest.java    MHz
> > > > Solaris sparc    13.12u  0.01s  0:13.17  99.6%    5569 ms           1503
> > > > Solaris x86       6.28u  0.00s  0:06.29  99.8%    2571 ms           2813
> > > > Linux i686        0.704u 0.004s 0:00.70 100.0%    5587 ms           2079.593
> >
> > I found RegexTest.java but it takes as input a file of patterns/string "regex.txt"
> 
> Ahh - good.
> 
> > and it looks like that file is user-specified
> 
> Yes - the RE is hardcoded and as file (first arg to each prog) one may
> choose any file [which yields the worst case aka no match]. For the
> numbers above I've choosen an email of my SUN contact, which contains
> matching grant infos wrt. FUL aka EDU - so I guess not really public stuff.
> 
> elkner.q ~/tmp > wc xt
>     6617    6883  427011 xt
> 
> Anyway, produced the numbers again for those, which wanna fetch
> http://iws.cs.uni-magdeburg.de/~elkner/regex/mk.log
> 
> elkner.q ~/tmp > wc /tmp/mk.log
>     2633   27847  325535 /tmp/mk.log
> 
>                 regex.c                         RegexTest.java   MHz
> Solaris sparc   12.14u  0.00s  0:12.25  99.1%   5324 ms          1503
> Solaris x86      5.85u  0.00s  0:05.86  99.8%   2190 ms          2813
> Linux            0.676u 0.020s 0:00.69 100.0%   4447 ms          2079.593

Mhhh... it would be nice to get seperate numbers for LC_ALL=C and
LC_ALL=en_US.UTF-8 ... and maybe for LC_ALL=zh_CN.GB18030 , too (since
some GNU stuff has hardcoded hacks for UTF-8 to speed-up things (while
ignoring anything related to "standards" ... ;-( ))

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) roland.mainz at nrubsig.org
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 7950090
 (;O/ \/ \O;)