1. New rtrim()
Dear Eu users,
There is the rtrim() function in
the putsxy.e lib.
This function removes zeros from
the end of a sequence.
New rtrim_() function makes the
same thing about 30% faster
and is more simple.
--
function rtrim_(sequence s)
for i=length(s) to 1 by -1 do
if s[i] then
return s[1..i] end if end for
return {}
end function
--
Use it please if you need such a function.
Regards,
Igor Kachan
kinz at peterlink.ru
2. Re: New rtrim()
Colin,
Igor's unattributed brilliant idea *is* my version. Pity I did not
copyright the bloody thing as seems to be the trend these days.
But then I do not really care at all, I just hate confusion...
jiri
----- Original Message -----
From: <cetaylor at compuserve.com>
To: "EUforum" <EUforum at topica.com>
Subject: RE: New rtrim()
>
> Igor,
>
> Thanks for the suggestion. I tried it in putsxy.e and didn't see
any
> improvement in performance, so for the moment I'll stick with Jiri's
> original version.
>
> Colin Taylor
>
>
> Igor Kachan wrote:
> > Dear Eu users,
> >
> > There is the rtrim() function in
> > the putsxy.e lib.
> >
> > This function removes zeros from
> > the end of a sequence.
> >
> > New rtrim_() function makes the
> > same thing about 30% faster
> > and is more simple.
> >
> > --
> > function rtrim_(sequence s)
> > for i=length(s) to 1 by -1 do
> > if s[i] then
> > return s[1..i] end if end for
> > return {}
> > end function
> > --
> >
> > Use it please if you need such a function.
> >
> > Regards,
> > Igor Kachan
> > kinz at peterlink.ru
> >
> >
>
>
3. Re: New rtrim()
Colin Taylor wrote:
> Igor,
>
> Thanks for the suggestion. I tried it in putsxy.e and didn't see any
> improvement in performance, so for the moment I'll stick with Jiri's
> original version.
>
> Colin Taylor
>
>
> Igor Kachan wrote:
> > Dear Eu users,
> >
> > There is the rtrim() function in
> > the putsxy.e lib.
> >
> > This function removes zeros from
> > the end of a sequence.
> >
> > New rtrim_() function makes the
> > same thing about 30% faster
> > and is more simple.
> >
> > --
> > function rtrim_(sequence s)
> > for i=length(s) to 1 by -1 do
> > if s[i] then
> > return s[1..i] end if end for
> > return {}
> > end function
> > --
> >
> > Use it please if you need such a function.
> >
Yes, Colin, this function is not the noticeable
one for the putsxy.e perfomance.
But try please the test below.
--program test.ex
function rtrim_(sequence s)
for i=length(s) to 1 by -1 do
if s[i] then
return s[1..i] end if end for
return {}
end function -- rtrim_()
function rtrim(sequence s)
integer i
i=length(s)
while i do
if s[i] then exit end if
i=i-1
end while
return s[1..i]
end function -- rtrim()
sequence S,s
S=repeat(repeat(0,1000),1000)
atom T
T=time()
for i=1 to 1000 do
s=rtrim_(S[i])
end for
T=time()-T
? T
T=time()
for i=1 to 1000 do
s=rtrim(S[i])
end for
T=time()-T
? T
--end of program
On my a 386 machine, I have
T = 10.27 with rtrim_()
T = 14.28 with rtrim()
On the Pentium, you must replace
1000 by 10000, or so, to get
the correct results.
So, there may be the cases when
rtrim_() is useful enough.
I use rtrim_() function now,
if I need to remove zeros from the
end of a sequence.
Regards,
Igor Kachan
kinz at peterlink.ru
4. Re: New rtrim()
Colin,
>
> Igor Kachan wrote:
>
> >
> > Yes, Colin, this function is not the noticeable
> > one for the putsxy.e perfomance.
> >
> > But try please the test below.
> >
>
> Igor,
>
> Your test program always returns an empty sequence
> instead of a slice,
I have tested
S=repeat({1,1}&repeat(0,1000),1000)
just now.
Result is
10.27 with rtrim_()
14.34 with rtrim()
> because your test sequences contain only zeros.
> This isn't a realistic
> test, and it gives you an artificially fast score
> for your version of
> rtrim().
I think, this test is just the hardest
for such the program.
<snip>
Don't worry, use, thing is good enough
and very well tested now 
Regards,
Igor Kachan
kinz at peterlink.ru
