d081473c   
   From: grahn+nntp@snipabacken.se   
      
   On Sat, 2009-11-14, David Schwartz wrote:   
   > On Nov 14, 2:54 am, mockturtle  wrote:   
   >   
   >> Well, I would not really say broken, but with a different design   
   >> choice... :-)   
   >   
   > The problem is, the design choice is incompatible with TCP's   
   > documented segmentation behavior.   
   >   
   >>  I was able to get some more information about the   
   >> functions we are using.  They are part of a large library to write   
   >> "generalized structures" over "generalized streams" (files, network   
   >> connections, and so on...).  An older version of the library just   
   >> iterated over the "components" of the "structure" and wrote them one   
   >> at time.  Since the components of the string are its characters, they   
   >> were written  one at time.  [Technically, this is coherent with the   
   >> documented interface of the function: they grant you that the   
   >> structure will be written, but how this is done is left to the   
   >> implementation]   
   >   
   > If this is documented behavior, then connecting this directly to a TCP   
   > connection is broken behavior.   
      
   "Stupid behavior" or "an unbelievable waste of bandwidth", yes. And in   
   my opinion a big blinking "We are morons! Don't use our library!" neon   
   sign. But it's not really broken, is it? In the sense that the TCP   
   stacks or the other end may not process it properly?   
      
   /Jorgen   
      
   --   
     // Jorgen Grahn    O  o   .   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)