On Thursday 16 August 2007 11:59, Chris Lattner wrote:> >> Why not just use the unlocked stdio calls? > > > > Because I'm lazy. :) > > It seems much easier to use a robust and well tested system than invent a > new crazy scheme ;-)I'm wondering if there are issues with static initialization _a_la_ the contortions iostreams goes through to initialize cout and friends. Right now everything I have is initialized in constructors. I haven't looked at this enough to know for sure yet. -Dave
On Wed, 15 Aug 2007, David A. Greene wrote:>>> and thus won't be real useful for Windows. >> >> Why not just use the unlocked stdio calls? > > Because I'm lazy. :)It seems much easier to use a robust and well tested system than invent a new crazy scheme ;-)> No, seriously, because I just haven't had time to do it. And because I'm > lazy. If/when I submit this to LLVM I'll look at going that route. As for > now, I just have to get something working for our needs here.ok! -Chris -- http://nondot.org/sabre/ http://llvm.org/
On Thursday 16 August 2007 11:59, Chris Lattner wrote:> On Wed, 15 Aug 2007, David A. Greene wrote: > >>> and thus won't be real useful for Windows. > >> > >> Why not just use the unlocked stdio calls? > > > > Because I'm lazy. :) > > It seems much easier to use a robust and well tested system than invent a > new crazy scheme ;-)Actually, there's another reason not to use unlocked calls. They require POSIX compliance. To get portability and most of the performance I plan to look at unbuffered stdio. -Dave
On Thu, 16 Aug 2007, David Greene wrote:> Actually, there's another reason not to use unlocked calls. They require > POSIX compliance.Posix is pretty available, what system doesn't have them?> To get portability and most of the performance I plan to look at unbuffered > stdio.Buffering is goodness, no? -Chris -- http://nondot.org/sabre/ http://llvm.org/