Are you OK with working up a patch for using pipes or do you want me to do it?
Greg > On Aug 1, 2014, at 10:33 AM, Zachary Turner <[email protected]> wrote: > > That sounds fine. > > > On Fri, Aug 1, 2014 at 10:26 AM, Greg Clayton <[email protected]> wrote: > I was actually thinking of removing this code in Driver.cpp and replacing > with native pipes. This would avoid the temp file. I didn't really want to > expose the pipe class we made internally. We could use #ifdef's and use > pipe() for non windows and _pipe() for windows. Thoughts? > > My main objection to adding temp file support to the API is the only thing > that really needs it is the Driver and we could just work around this in > Driver.cpp by using #ifdef and native pipes. > > Comments? > > > On Jul 31, 2014, at 1:46 PM, Zachary Turner <[email protected]> wrote: > > > > Sorry, here's a more easily digestable patch. > > > > > > On Thu, Jul 31, 2014 at 1:45 PM, Zachary Turner <[email protected]> wrote: > > mkstemp doesn't exist on Windows. LLVM already has a support function > > which is the same in spirit as mkstemp. This patch adds a method to the > > host layer to create a temporary file, and the implementation of this > > method on the private side delegates to the corresponding LLVM support > > function. > > > > This fixes the build on Windows and should have identical semantics on > > other platforms, but please verify. > > > > <mkstemp.diff> > > _______________________________________________ lldb-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
