On Mon, May 19, 2014 at 11:15 PM, Thomas Braun <thomas.braun@xxxxxxxxxxxxxxx> wrote: > Am 19.05.2014 21:33, schrieb Jonathan Nieder: > >> Hi, >> >> Thomas Braun wrote: >> >>> pushing over the dump git protocol with a windows git client. >> >> >> I've never heard of the dump git protocol. Do you mean the git >> protocol that's used with git:// URLs? > > > You are right I mean the protocol involving git:// URLs. But unfortunately I > got it wrong as according to [1] the git:// is one of the so-called smart > protocols. That was also the source where I read that there are smart and > dump protocols. > > [1]: http://git-scm.com/book/en/Git-Internals-Transfer-Protocols > > >> [...] >>> >>> Alternative approaches considered but deemed too invasive: >>> - Rewrite read/write wrappers in mingw.c in order to distinguish between >>> a file descriptor which has a socket behind and a file descriptor >>> which has a file behind. >> >> >> I assume here "too invasive" means "too much engineering effort"? >> >> It sounds like a clean fix, not too invasive at all. But I can >> understand wanting a stopgap in the meantime. > > > No actually I meant too invasive in the sense of "requiring large rewrites > which only benefit git on windows and hurt all others". > > The two fixes I can think of either involve: > - In a read *and* write wrapper the need to check if the fd is a socket, if > yes use send/recv if no use read/write. According to Erik's comments this > should be possible. But I would deem the expected performance penalty quite > large as that will be done in every call. You clearly haven't stepped through MSVCRT's read and write implementations :P I wouldn't worry too much about this, at least not until the numbers are in. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html