Re: [PATCH] move socket-related client parts into separate file
Derek Robert Price <[email protected]>
| Newsgroups | gmane.comp.version-control.cvs.bugs |
|---|---|
| Organization | Ximbiot <http://ximbiot.com> |
| Message-ID | <[email protected]> |
Alexey Mahotkin wrote: >>>>>>"DRP" == Derek Robert Price <[email protected]> writes: >>>>>> >>>>>> > > DRP> Have you tried compiling this on Windows? Windows and OS2 are the > DRP> only platforms that define NO_SOCKET_TO_FD that I know of and this > DRP> will break there. At the very least, cvsnt.mak will need to be > DRP> modified to find the new source file, but it is possible that this > DRP> code needs some includes from cvs.h and client.h. Don't worry about > DRP> OS2 so much, but we still try and keep support on Windows up to date. > > DRP> Okay, I just defined NO_SOCKET_TO_FD by hand on UNIX and attempted to > DRP> compile and discovered that, at the least, socket-client.c needs some > DRP> function prototypes from cvs.h. I also got some "dereferencing > DRP> pointer to incomplete type" errors I didn't bother to track down. > > DRP> I'm cc'ing bug-cvs in the interest of additional input. > >That's a lose-lose scenario IMHO. That means that there are ~300 lines >which could not be checked by most of the developers. > Which part is lose-lose? Supporting Windows? > Btw, is that really >still a case for modern Win32, that sockets are not convertible to fds? > Don't know about WinXP, but I'm still trying to keep up Win95 support and it is probably still true of Win98. >Ok, I turned on NO_SOCKET_TO_FD temporarily, and I think that this >compile-fixes it. > Passes muster on my machine too. >2003-04-24 Alexey Mahotkin <[email protected]> > > >Fixes for NO_SOCKET_TO_FD > Committed, on top of Alexey's previous socket-client patch. Derek -- *8^) Email: [email protected] Get CVS support at <http://ximbiot.com>! -- You want my advice? Go back to Bulgaria. - Humphrey Bogart as Rick, _Casablanca_