This takes 5 seconds using Windows 98 SE

[Mike Doty]

Why would the TCP OPEN take 5 seconds using Windows 98 SE?
Immediate using XP. Changed port number and host for security reasons.

'snoop.bas
#COMPILE EXE
#DIM ALL
FUNCTION PBMAIN () AS LONG
  'Send information to your running monitor or just write to a page on your server
  LOCAL nsocket AS LONG, PortNumber AS LONG, TheServer AS STRING ,sLine AS STRING
 
  PortNumber = 999
  TheServer = "99.99.99.99"
  'Method 2:        Send to a server program running on your desktop
  nsocket = FREEFILE
  TCP OPEN PORT PortNumber AT TheServer AS nSocket 'TIMEOUT 250
  IF ERR = 0 THEN
    sLine = "TEST"
    TCP SEND nSocket, sLine
    TCP CLOSE nSocket
  ELSE
    ? "Unable to connect to " + TheServer+ " on port" + STR$(PortNumber)
  END IF
END FUNCTION

[Michael Mattias]

Five seconds would be the average of how many runs under what load conditions (computer) and what level of network trafffic using how many different internet connection methods with what settings for networks?

Sometimes it just takes five seconds, ya know?

Or I could cop out and tell you you are lucky you connected at all when using an unsupported operating system.

[Marco Pontello]

Maybe a different kind of (personal) firewall / antivirus / anti-something on the two system?

Bye!

[Joe Byrne]

Or the fact that the IP stack was an add-on in Win9.x and not 100% standard. There are lots of quirks with TCP on the Win9.x platform.

[Mike Stefanik]

Or the fact that the IP stack was an add-on in Win9.x and not 100% standard. There are lots of quirks with TCP on the Win9.x platform.

Absolutely. Writing code so that SocketTools would behave consistently for both Windows 9x and Windows NT/2K/XP was a huge hassle. The day that Microsoft finally pulled the plug on Windows 98, I had a beer or three in celebration of its passing. And ripping that ugly compatibility code out of SocketTools 5.0 was a joyous event in of itself.