On Tue, Nov 20, 2018 at 12:49:36AM +0000, Ramsay Jones wrote:
>
>
> On 20/11/2018 00:18, Luc Van Oostenryck wrote:
> > On Mon, Nov 19, 2018 at 08:52:42PM +0000, Ramsay Jones wrote:
> >>
> >> The current -dD option outputs the macro definitions, in addition to the
> >> pre-processed text. In contrast, the -dM option outputs only the macro
> >> definitions.
> >>
> >> Signed-off-by: Ramsay Jones <ramsay@xxxxxxxxxxxxxxxxxxxx>
> >> ---
> >> lib.c | 41 +++++++++++++++++-----
> >> lib.h | 1 +
> >> validation/preprocessor/dump-macros-only.c | 36 +++++++++++++++++++
> >> validation/preprocessor/dump-macros.c | 6 ++++
> >> 4 files changed, 75 insertions(+), 9 deletions(-)
> >> create mode 100644 validation/preprocessor/dump-macros-only.c
> >>
> >> diff --git a/lib.c b/lib.c
> >> index 07a5b9c..23c3d27 100644
> >> --- a/lib.c
> >> +++ b/lib.c
> >> @@ -287,6 +287,7 @@ int Wunknown_attribute = 0;
> >> int Wvla = 1;
> >>
> >> int dump_macro_defs = 0;
> >> +int dump_macros_only = 0;
> >>
> >> int dbg_compound = 0;
> >> int dbg_dead = 0;
> >> @@ -794,16 +795,34 @@ static char **handle_switch_v(char *arg, char **next)
> >> return next;
> >> }
> >>
> >> -static struct flag dumps[] = {
> >> - { "D", &dump_macro_defs},
> >> -};
> >> -
> >> static char **handle_switch_d(char *arg, char **next)
> >> {
> >> - char ** ret = handle_onoff_switch(arg, next, dumps, ARRAY_SIZE(dumps));
> >> - if (ret)
> >> - return ret;
> >> + char *arg_char = arg + 1;
> >>
> >> + /*
> >> + * -d<CHARS>, where <CHARS> is a sequence of characters, not preceded
> >> + * by a space. If you specify characters whose behaviour conflicts,
> >> + * the result is undefined.
> >> + */
> >> + while (*arg_char) {
> >> + switch (*arg_char) {
> >> + case 'M': /* dump just the macro definitions */
> >> + dump_macros_only = 1;
> >> + dump_macro_defs = 0;
> >
> > I prefer a small change in the logic to have dump_macro_defs set
> > here too.
>
> Hmm, I don't quite follow.
Sorry, I didn't saw this mail yesterday.
I just meant that I would prefer to let dump_macro_defs set to 1 here
and then later the test can just be:
if (dump_macro_defs) {
...
}
> >
> >> + break;
> >> + case 'D': /* like 'M', but also output pre-processed text */
> >> + dump_macro_defs = 1;
> >> + dump_macros_only = 0;
> >> + break;
> >
> >
> >> + case 'N': /* like 'D', but only output macro names not bodies */
> >> + break;
> >> + case 'I': /* like 'D', but also output #include directives */
> >> + break;
> >> + case 'U': /* like 'D', but only output expanded macros */
> >> + break;
> >
> > Do you have a specific reason to add those three too?
> > Now if someone use 'gcc -d[NIU] ...' will silently succeed, right?
>
> Heh, yeah, I was going to see how much effort it would take to
> implement all of the options that gcc provides, so I put these
> here as a reminder/place-holder ... but didn't get back to it! :-D
Yes :) -dN should be easy to add but -dI and -dU would need much
more work.
Best regards,
-- Luc
