On Wed, 29 Nov 2017 12:03:14 +0200
Vladislav Valtchev wrote:
> On Tue, 2017-11-28 at 14:35 -0500, Steven Rostedt wrote:
> >
> > Yes. Simply because we lost the fact that we can do it better.
> >
>
> Hi Steven,
> of course we can do better, if we make a step further than just
> a mechanical r
On Tue, 2017-11-28 at 14:35 -0500, Steven Rostedt wrote:
>
> Yes. Simply because we lost the fact that we can do it better.
>
Hi Steven,
of course we can do better, if we make a step further than just
a mechanical refactoring. I'm used to intentionally limit myself
to do such kind of mechanical
On Tue, 28 Nov 2017 20:34:22 +0200
Vladislav Valtchev wrote:
> On Tue, 2017-11-28 at 12:14 -0500, Steven Rostedt wrote:
> > As everything is doing both init_common_record_context() and
> > parse_record_options() why not just move the
> > init_common_record_context() into parse_record_option
On Tue, 2017-11-28 at 12:14 -0500, Steven Rostedt wrote:
> As everything is doing both init_common_record_context() and
> parse_record_options() why not just move the
> init_common_record_context() into parse_record_options(). If you need
> to do something special (like set events = 1 for p
On Thu, 23 Nov 2017 18:33:32 +0200
"Vladislav Valtchev (VMware)" wrote:
> +static void init_common_record_context(struct common_record_context
*ctx,
> +enum trace_cmd curr_cmd)
> +{
> + memset(ctx, 0, sizeof(*ctx));
> + ctx->instance = &top_instance;
>
This simple patch make the above-mentioned commands independent from 'record'.
The point of doing so is to follow the convention of one entry-point per command
that is followed by most of trace-cmd's code. Ultimately that aims to prevent
the complexity to concentrate in a single point, making the c
6 matches
Mail list logo
