Hi Masahiro, On 18/02/2020 00:37, Masahiro Yamada wrote: > Hi, > [...] >>> >> >> Could you please elaborate a bit more? > > > I am just not familiar with this usage of /dev/zero. > > > The behavior of /dev/null is clear. > The read/write to /dev/null is described in the standard: > https://pubs.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap10.html > > I just do not know whether the write to /dev/zero is > the defined-behavior or not. > Even if /dev/null is more common for this use, in linux /dev/zero has the same behavior for what concerns the write operation. In particular all write operations to /dev/zero succeed with no other effects. You do not find it at the link you provided because existence of /dev/zero is not required by POSIX as far as I know, but the behavior is standardized by Linux ABI hence I do not think it is going to change. Please have a look at drivers/char/mem.c +818 where the operation is defined. Based on this I propose to put a comment on why we are not using /dev/null in this case accordingly with what you stated in your previous email, but to leave the rule as is. > > Masahiro Yamada > > >> >>> >>> Does this work? >>> >>> as-option = $(success, cat /dev/null | $(CC) $(CLANG_FLAGS) $(1) -c -x >>> assembler -o /dev/null -) >>> >> [...] -- Regards, Vincenzo
Attachment:
pEpkey.asc
Description: application/pgp-keys