1. VERY strange problem? HELP!
I have been working on a program when somthing very strange happened.
The server generates a number using "n=time()*100+rand(99)",
then starting a client program using 'n' as a parameter.
After getting strange results, I modified both to display n.
I ran it a few times and got:
# from server 488 567 (501) 467
# from client 488 567 (500) 467
As you can see, at on point the server displayed '501'
and the client displayed '500'.
I tried to create the problem in another program, but got another
strange error:
[ex.err]
...\New Windows App.exw:11 in procedure w_onOpen()
type_check failure, n is 486
self = 3
event = 4
params = {}
o = 6
...
...\New Windows App.exw:
f = 42'*'
n = 486
[end]
In types.e:
constant True = (1=1)
global type int( integer i )
return True
end type
so 'n' is an integer, and 486 is a perfectly leagal value for an integer.
Also, int() always returns True, so an invalid value woulgd cause
an error on 'i'. And the program also displayed a few other numbers just fine.
The example programs are below, put them in thsame place and
run New Windows App.exw
Can somine explain what is happenning here?!?!?!?!
[New Windows App.exw]
include win32lib.ew
without warning
int f,n
constant w=createEx( Window, "", 0, Default, Default, 600, 400, 0, 0 )
procedure w_onOpen (integer self, integer event, sequence params)--params is ()
for o=1 to 10 do
n=time()*100+rand(99)
?n
f=shellExecuteEx(0,
"exw.exe",
sprintf("%s %d",{"p2",n}),
"",--reverse(ms1),
SW_SHOWNORMAL,
0)
end for
end procedure
setHandler( w, w32HClick, routine_id("w_onOpen"))
WinMain(w,Normal)
[end]
[p2.exw]
include win32lib.ew
seq a
a=command_line()
puts(1,a[3])
a[1]=wait_key()
[end]
2. Re: VERY strange problem? HELP!
I think a hacker is messing up my computer;
it is possible that is what is causing this,
but is that even possible?
CoJaBo wrote:
>
> I have been working on a program when somthing very strange happened.
> The server generates a number using "n=time()*100+rand(99)",
> then starting a client program using 'n' as a parameter.
> After getting strange results, I modified both to display n.
> I ran it a few times and got:
>
> # from server 488 567 (501) 467
> # from client 488 567 (500) 467
>
> As you can see, at on point the server displayed '501'
> and the client displayed '500'.
> I tried to create the problem in another program, but got another
> strange error:
>
> [ex.err]
> ...\New Windows App.exw:11 in procedure w_onOpen()
> type_check failure, n is 486
> self = 3
> event = 4
> params = {}
> o = 6
>
> ...
>
> ...\New Windows App.exw:
> f = 42'*'
> n = 486
> [end]
>
> In types.e:
> constant True = (1=1)
>
> global type int( integer i )
> return True
> end type
> so 'n' is an integer, and 486 is a perfectly leagal value for an integer.
> Also, int() always returns True, so an invalid value woulgd cause
> an error on 'i'. And the program also displayed a few other numbers just fine.
>
> The example programs are below, put them in thsame place and
> run New Windows App.exw
> Can somine explain what is happenning here?!?!?!?!
>
> [New Windows App.exw]
> include win32lib.ew
> without warning
>
>
> int f,n
> constant w=createEx( Window, "", 0, Default, Default, 600, 400, 0, 0 )
>
>
> procedure w_onOpen (integer self, integer event, sequence params)--params is
> ()
> for o=1 to 10 do
> n=time()*100+rand(99)
> ?n
> f=shellExecuteEx(0,
> "exw.exe",
> sprintf("%s %d",{"p2",n}),
> "",--reverse(ms1),
> SW_SHOWNORMAL,
> 0)
> end for
> end procedure
> setHandler( w, w32HClick, routine_id("w_onOpen"))
> WinMain(w,Normal)
> [end]
>
> [p2.exw]
> include win32lib.ew
>
> seq a
> a=command_line()
> puts(1,a[3])
> a[1]=wait_key()
> [end]
>
3. Re: VERY strange problem? HELP!
If n displays as 486,
it does not mean that n is precisely
the integer: 486.00000000000000...
n might well be: 485.99999999999999
which will fail the integer() test.
By default, Euphoria displays numbers to
about 10 significant digits.
Try:
? n - 486
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
4. Re: VERY strange problem? HELP!
On Fri, 25 Jun 2004 18:44:00 -0700, CoJaBo <guest at RapidEuphoria.com>
wrote:
>The server generates a number using "n=time()*100+rand(99)",
Try n=floor(time()*100)+rand(99)
Regards,
Pete
5. Re: VERY strange problem? HELP!
On Sat, 26 Jun 2004 14:41:25 -0700, CoJaBo <guest at RapidEuphoria.com>
wrote:
>On Win32, time()*100+rand(99) can only be an integer.
FALSE. Try:
?integer(1.1*100)
IEEE 754 floats are not designed for this handling. This is a hardware
feature, nothing to do with RDS Euphoria, or any other software.
As you have found, if you had defined n as an integer in the first
place, rather than an atom or object, the error would have occurred
immediately, and I think you would have figured it out much quicker.
Regards,
Pete
6. Re: VERY strange problem? HELP!
Thanks! The problem was that it was defined as an ATOM. I thought
it was an INT, but I had been looking at another similirely named
integer (which explains why time()*100+rand(99) seemed to have been
returning an integer). But that still doesn't explain why it ?time()
once displayed -3812688, or why the problem only exists around this
time (it worked fine this morning!). You will be added to the credits
screen for the program (for if I ever release it).
Pete Lomax wrote:
>
> On Sat, 26 Jun 2004 14:41:25 -0700, CoJaBo <guest at RapidEuphoria.com>
> wrote:
>
> >On Win32, time()*100+rand(99) can only be an integer.
> FALSE. Try:
>
> ?integer(1.1*100)
>
> IEEE 754 floats are not designed for this handling. This is a hardware
> feature, nothing to do with RDS Euphoria, or any other software.
>
> As you have found, if you had defined n as an integer in the first
> place, rather than an atom or object, the error would have occurred
> immediately, and I think you would have figured it out much quicker.
>
> Regards,
> Pete
>
>
