On Fri, Apr 23, 2021 at 11:24 AM Martin Liška <mli...@suse.cz> wrote: > > On 4/23/21 8:49 AM, Richard Biener wrote: > > On Thu, Apr 22, 2021 at 9:47 PM Andi Kleen via Gcc-patches > > <gcc-patches@gcc.gnu.org> wrote: > >> > >> Martin Liška <mli...@suse.cz> writes: > >> > >>> Hey. > >>> > >>> I/O buffering in gcov seems duplicite to what modern C library can > >>> provide. > >>> The patch is a simplification and can provide easier interface for system > >>> that don't have a filesystem and would like using GCOV. > >>> > >>> I'm going to install the patch after 11.1 if there are no objections. > >>> > >>> Patch can bootstrap on x86_64-linux-gnu and survives regression tests. > >> > >> What happens if someone compiles the C library with gcov? > > Haven't tried that.. > > > > > Yeah, I think this is the wrong direction - we're already having > > issues with using > > malloc, this makes it much worse. > > I don't see where problem with my patch? It's not adding usage of any > additional > system routines.
True. So the only impact should be more calls to libc (not necessarily more I/O since libc should do buffering). "The patch is a simplification and can provide easier interface for system that don't have a filesystem and would like using GCOV." Can you explain? > Martin > > > > > Richard. > > > >> Being as self contained as possible (only system calls) would seem > >> safer. > >> > >> -Andi >
