[Python-Dev] noreply@sourceforge.net: [Python-bugs-list] [Bug #111620] lots of use of send() without verifyi ng amount of d (original) (raw)

Fredrik Lundh Fredrik Lundh" <effbot@telia.com
Sat, 12 Aug 2000 16:29:57 +0200

gordon wrote:

Fredrik wrote: =20 > fwiw, if "send" may send less than the full buffer in blocking > mode on some platforms (despite what the specification implies), > it's quite interesting that nobody has ever noticed before... =20 I noticed, but I expected it, so had no reason to comment. The=20 Linux man pages are the only specification of send that I've=20 seen that don't make a big deal out it. And clearly I'm not the=20 only one, otherwise there would never have been a bug report=20 (he didn't experience it, he just noticed sends without checks).

I meant "I wonder why my script fails" rather than "that piece of code looks funky".

::: fwiw, I still haven't found a single reference (SUSv2 spec, man- pages, Stevens, the original BSD papers) that says that a blocking socket may do anything but sending all the data, or fail.

if that's true, I'm not sure we really need to "fix" anything here...