On Mon, 2004-08-16 at 09:23 +0900, GOTO Masanori wrote: > At Sun, 15 Aug 2004 03:02:47 -0400, > Andres Salomon wrote: > > (Not filing as wishlist because, well, it's a bug.) > > If sendfile64() fails because the underlying OS doesn't implement it (as > > linux 2.4.x fails to do for a number of architectures), it currently > > just breaks. If large file support is turned on, then there's no way > > for user applications to fall back to normal sendfile(). Since glibc > > already takes care of replacing sendfile() calls w/ sendfile64(), it > > should also handle calling sendfile() if sendfile64() isn't implemented > > by the OS. > > Do you have a test case? > > Regards, > -- gotom > Currently, apache2 is our test case. Adam Conrad or myself will send you a smaller test case, post-sarge; in the meantime, this bug can wait. -- Andres Salomon <dilinger@voxel.net>
Attachment:
signature.asc
Description: This is a digitally signed message part