I don't really have the time, but here's some debugging code that'll maybe help diagnosing why/if WinSock is failing to start up: - compile the attached initws.c via ghc, foo$ ghc -c initws.c - run some tests foo$ ghc -package net IWS.hs initws.o -e "initWS 1 1 >>= print" foo$ ghc -package net IWS.hs initws.o -e "initWS 2 2 >>= print" foo$ ghc -package net IWS.hs initws.o -e "initWS 1 1 >> Network.BSD.getProtocolByName \"tcp\" >>= print" foo$ ghc -package net IWS.hs initws.o -e "initWS 2 2 >> Network.BSD.getProtocolByName \"tcp\" >>= print" initWS returns 0 on success. The GHC networking support is ultra-conservative wrt WinSock versioning, sticking with 1.1 for max portability. Perhaps that's working against us here? Needless to say, this is working for me (on a pair of XP boxes.) --sigbjorn ----- Original Message ----- From: "Joel Reymont" <joelr1@gmail.com> To: "Simon Marlow" <simonmar@microsoft.com> Cc: "Haskell Cafe" <haskell-cafe@haskell.org>; <sof@galois.com> Sent: Friday, October 21, 2005 06:31 Subject: Re: [Haskell-cafe] Re: "getServiceEntry: does not exist" on Windows
It stopped happening for me on WinXP but it's still happening for my customer on Win2K.
On Oct 21, 2005, at 1:45 PM, Simon Marlow wrote:
Nope, actually it also happens for me with the updated installer.
Sigbjorn - any ideas? It doesn't happen with my local STABLE build, but I'm still using gcc 3.2.x.