Re: address@hidden: libjava build times]

[Ralf Wildenhues]

* Joe Buck wrote on Fri, May 06, 2005 at 01:03:14AM CEST:
> On Thu, May 05, 2005 at 11:25:53PM +0200, Ralf Wildenhues wrote:
> > * Albert Chin wrote on Thu, May 05, 2005 at 09:20:50PM CEST:
> > > On Thu, May 05, 2005 at 03:17:26PM +0200, Ralf Wildenhues wrote:
> > > > 
> > > > Stuff like
> > > >   compile_command="$compile_command $qarg"
> > > > is better written as
> > > >   $ECHO " $qarg" >&FD_COMPILE_COMMAND
> > > >   compile_command=`$SED 's/^ //' <"$compile_command_file" | $NL2SP`
> > > 
> > > Can you explain how the latter is faster? From inspection, I'd believe
> > > the first to be faster.
> > 
> > The former scales quadratically with some shells (all I tested on
> > GNU/Linux), presumably because they do repeated reallocs.
> 
> Wow.  One of my favorite job interview questions is how to implement
> a string class or (for C types) dynamic string library that can grow
> strings without producing quadratic behavior.

Oh well.

> Are you saying that bash fails the test?

GNU bash, version 3.00.14(1)-release (x86_64-redhat-linux-gnu)
Copyright (C) 2004 Free Software Foundation, Inc.

ash-0.3.8

PD KSH v5.2.14 99/07/13.2

all fail the test, with slightly different crossover times, of course,
but even with constant string addition.  I believe zsh does, too, but
have not tested it now.

Regards,
Ralf

Test (uses unportable `seq'):
# --- a.sh ---
a=
: ${n=1000}
for i in `seq 1 $n`
do
        a="$a 1"
done