From: John Carter <john.carter@...>
Date: 2008-03-27T10:50:21+09:00
Subject: Re: "loop do IO.select([io], nil, nil)" eats 95% of CPU,any other  	way?

--Boundary_(ID_7jYW9YVuvyxVA4b4l+OBuQ)
Content-type: TEXT/PLAIN; format=flowed; charset=X-UNKNOWN
Content-transfer-encoding: QUOTED-PRINTABLE

On Thu, 27 Mar 2008, I=C3=B1aki Baz Castillo wrote:

> ...
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> select(4, [3], [], [], {0, 0})          =3D 0 (Timeout)  <---
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0
> sigprocmask(SIG_BLOCK, NULL, [])        =3D 0

My guess is if you include the read statements you get something like=
 this...
   read(3, "", 1000)                       =3D 0
Which means (according to "man 2 read")
   On success, the number of bytes read is returned (zero indicates e=
nd of file)

If you read "man select_tut"
        9.     If the functions read(2), recv(2), write(2), and send(=
2) fail  with  errors  other  than  those
               listed  in 7., or one of the input functions returns 0=
, indicating end of file, then you should
               not pass that descriptor to select() again.  In the ab=
ove example, I close the descriptor imme=E2=80=90
               diately, and then set it to -1 to prevent it being inc=
luded in a set.

Thus...

mkfifo foofi
ruby -w test.rb&
echo bah > foofi

=3D=3Dtest.rb=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
loop do
    begin
       open( "foofi") do |f|
          loop do
             p select([f],nil,nil,500000)
             p f.sysread(1000)
          end
       end
    rescue EOFError =3D> details
       p details
    end
end
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D
Works as expected...

But...

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D
open( "foofi") do |f|
    loop do
      p select([f],nil,nil,500000)
      p f.sysread(1000)
    end
end
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D

Bombs out with EOFError

but curiously
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D
open( "foofi") do |f|
    loop do
       p select([f],nil,nil,500000)
       p f.read(1000)
    end
end
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D
Exhibits exactly the behaviour you describe.

Looking at the strace....
open("foofi", O_RDONLY|O_LARGEFILE)     =3D 9
select(10, [9], NULL, NULL, {500000, 0}) =3D 1 (in [9], left {500000,=
 0})
fstat64(1, {st_dev=3Dmakedev(0, 11), st_ino=3D37, st_mode=3DS_IFCHR|0=
620, st_nlink=3D1, st_uid=3D1001, st_gid=3D5, st_blksize=3D1024, st_b=
locks=3D0, st_rdev=3Dmakedev(136, 35), st_atime=3D2008/03/27-14:45:52=
, st_mtime=3D2008/03/27-14:45:52, st_ctime=3D2008/03/27-14:45:52}) =
=3D 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1=
, 0) =3D 0xb7db2000
write(1, "[[#<File:foofi>], [], []]\n", 26) =3D 26
fstat64(9, {st_dev=3Dmakedev(8, 3), st_ino=3D2375703, st_mode=3DS_IFI=
FO|0644, st_nlink=3D1, st_uid=3D1001, st_gid=3D65534, st_blksize=3D40=
96, st_blocks=3D0, st_size=3D0, st_atime=3D2008/03/27-14:43:56, st_mt=
ime=3D2008/03/27-14:45:54, st_ctime=3D2008/03/27-14:45:54}) =3D 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1=
, 0) =3D 0xb7db1000
read(9, "bah\n", 4096)                  =3D 4
read(9, "", 4096)                       =3D 0
write(1, "\"bah\\n\"\n", 8)             =3D 8
select(10, [9], NULL, NULL, {500000, 0}) =3D 1 (in [9], left {500000,=
 0})
write(1, "[[#<File:foofi>], [], []]\n", 26) =3D 26
write(1, "nil\n", 4)                    =3D 4
select(10, [9], NULL, NULL, {500000, 0}) =3D 1 (in [9], left {500000,=
 0})
write(1, "[[#<File:foofi>], [], []]\n", 26) =3D 26
write(1, "nil\n", 4)                    =3D 4

The curiosity is why doesn't the f.read throw an EOFError

According to ri IO#read...

      At end of file, it returns nil or "" depend on length.
      ios.read() and ios.read(nil) returns "".
      ios.read(positive-integer) returns nil.

So it all behaves according to plan... just not your plan... :-))

John Carter                             Phone : (64)(3) 358 6639
Tait Electronics                        Fax   : (64)(3) 359 4632
PO Box 1645 Christchurch                Email : john.carter@tait.co.n=
z
New Zealand

--Boundary_(ID_7jYW9YVuvyxVA4b4l+OBuQ)--