1. Eu 2.3 trace BUG!
- Posted by akusaya at gmx.net
Apr 11, 2002
This is my code:
------------------ start ------------
function cekPrima(sequence bilPrima, atom cekangka)
atom isilist
for j=1 to length(bilPrima) do
isilist=bilPrima[j]
if (cekangka/isilist - floor(cekangka/isilist))=0 then
--bukan bil prima
return 0
else
--bilprima
if power(isilist,2)>cekangka then
return 1
end if
end if
end for
return 1
end function
sequence p
p = {2, 3}
for i = 3 to 10000 by 2 do
if cekPrima(p, i)=1 then
p &= i
end if
end for
? p
integer pos
integer jum
with trace
trace(1)
for i=2 to 100000000 do
pos = 1
jum = 1
while 1 do
if p[pos] > sqrt(i) then
exit
elsif p[pos] = sqrt(i) then
jum += p[pos]
else
pos += 1
if integer( i/p[pos] ) then
jum += i/p[pos] + p[pos]
end if
end if
end while
if jum = i then
? i
end if
end for
------------------------- end ----------------
Execute, and it will bring a trace window.
Highlighted: for i=2 to 100000000 do
Press DOWN. (don't press ENTER)
Press ?i
"i=9999" is displayed on left side??? it should be 2.
Press ENTER many times until "end for"
Press ENTER one time more.
"i=9999" in the left disappears, "i=3" appears in right side.
Press ?i
"i=3" in the right disappears, "i=9999" appears again in left side.
Is this really a bug?!
2. Re: Eu 2.3 trace BUG!
On 11 Apr 2002, at 21:18, akusaya at gmx.net wrote:
>
> This is my code:
>
> ------------------ start ------------
>
> function cekPrima(sequence bilPrima, atom cekangka)
> atom isilist
>
> for j=1 to length(bilPrima) do
> isilist=bilPrima[j]
> if (cekangka/isilist - floor(cekangka/isilist))=0 then
> --bukan bil prima
> return 0
> else
> --bilprima
> if power(isilist,2)>cekangka then
> return 1
> end if
> end if
> end for
>
> return 1
>
> end function
>
> sequence p
> p = {2, 3}
>
> for i = 3 to 10000 by 2 do
> if cekPrima(p, i)=1 then
> p &= i
> end if
> end for
>
> ? p
>
> integer pos
> integer jum
>
> with trace
> trace(1)
>
> for i=2 to 100000000 do
> pos = 1
> jum = 1
>
> while 1 do
> if p[pos] > sqrt(i) then
> exit
> elsif p[pos] = sqrt(i) then
> jum += p[pos]
> else
> pos += 1
> if integer( i/p[pos] ) then
> jum += i/p[pos] + p[pos]
> end if
> end if
> end while
>
> if jum = i then
> ? i
> end if
> end for
>
> ------------------------- end ----------------
>
> Execute, and it will bring a trace window.
>
> Highlighted: for i=2 to 100000000 do
>
> Press DOWN. (don't press ENTER)
Press ENTER. (Don't press DOWN)
> Press ?i
>
> "i=9999" is displayed on left side??? it should be 2.
No, because the display will not update if you do not press ENTER. The i
was 9999 from lines before.
> Press ENTER many times until "end for"
>
> Press ENTER one time more.
>
> "i=9999" in the left disappears, "i=3" appears in right side.
Because you pressed ENTER.
> Press ?i
>
> "i=3" in the right disappears, "i=9999" appears again in left side.
>
>
> Is this really a bug?!
No. You should press ENTER.
Kat
3. Re: Eu 2.3 trace BUG!
Aku writes:
> Is this really a bug?!
Yes. There's a known bug, originally reported by Ricardo Forno,
where the ? command in the trace facility can sometimes
get confused when you have multiple for-loops that use
the same loop variable.
Instead of reporting the value for the currently active loop variable
"i", it might report the final value attained by "i" in some previous
loop in the same routine, or perhaps both loops are at the top-level.
The previous "i" should have been forgotten, but it's actually
still in the symbol table, for various reasons. This can cause
the interactive ? command in trace to "see" the wrong variable.
Execution of programs is not affected, just the value you see in
the trace window. I'll try to fix this for the next release.
Thanks for the example code.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
