Re: [PATCH v6 43/44] builtin/apply: add a cli option for be_silent

On Fri, Jun 10, 2016 at 10:59 PM, René Scharfe [off-list ref] wrote:

Am 10.06.2016 um 22:11 schrieb Christian Couder:

quoted

Let's make it possible to request a silent operation on the
command line.

Signed-off-by: Christian Couder <redacted>
---
  builtin/apply.c | 2 ++
  1 file changed, 2 insertions(+)

diff --git a/builtin/apply.c b/builtin/apply.c
index ddd61de..93744f8 100644
--- a/builtin/apply.c
+++ b/builtin/apply.c

@@ -74,6 +74,8 @@ int cmd_apply(int argc, const char **argv, const char

*prefix)
                OPT_BOOL(0, "allow-overlap", &state.allow_overlap,
                        N_("allow overlapping hunks")),
                OPT__VERBOSE(&state.apply_verbosely, N_("be verbose")),
+               OPT_BOOL(0, "silent", &state.be_silent,
+                       N_("do not print any output")),
                OPT_BIT(0, "inaccurate-eof", &options,
                        N_("tolerate incorrectly detected missing new-line
at the end of file"),
                        APPLY_OPT_INACCURATE_EOF),

Why not -q/--quiet as for most other commands?

First as I say in the cover letter, I am going to discard this patch.
That is because it is not necessary, and it appeared a bit
controversial whether I should use -q/--quiet or not in the v2
discussions.

Furthermore, you could use OPT__VERBOSITY, which causes -v and -q to update
the same variable variable and thus lets parseopt handle their interaction.
Perhaps verbosity == 1 could mean verbose, 0 normal, -1 no infos, -2 no
warnings and -3 no errors?

Yeah, that could be done.

And if you add the ability to silence the apply functions before using them
you don't have to export and unexport dup_devnull().

Yeah, I will do something like that in the next version.

Thanks,
Christian.