Re: [PATCH 6/6] trace-cmd record: Add new parameter --file-version

[Steven Rostedt <rostedt@goodmis.org>]

On Thu, 29 Apr 2021 06:34:14 +0300
Tzvetomir Stoyanov <tz.stoyanov@gmail.com> wrote:

> On Thu, Apr 29, 2021 at 4:26 AM Steven Rostedt <rostedt@goodmis.org> wrote:
> >
> > On Thu, 22 Apr 2021 10:17:18 +0300
> > "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com> wrote:
> >  
> > > Added a new optional parameter to "trace-cmd record", can be used to
> > > select the desired file version of the trace output file.
> > >
> > > Signed-off-by: Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com>  
> >
> > Even with this last patch, I still get:
> >
> > trace-input.c: In function ‘tracecmd_get_file_version’:
> > trace-input.c:4033:15: error: ‘struct tracecmd_input’ has no member named ‘file_version’
> >  4033 |  return handle->file_version;
> >       |               ^~
> > trace-input.c:4034:1: warning: control reaches end of non-void function [-Wreturn-type]
> >  4034 | }
> >
> > So you must have added a change without somehow committing it :-/  
> 
> The confusion is from another dependency - the "[PATCH 0/6] Bump trace
> file version" patchset depends on "[PATCH v2] trace-cmd: Check if file
> version is supported". I wrote that in the cover letter, but maybe

Ah, I didn't read the cover letter, and just pulled the patches directly
from patchwork. I missed the v2 in patch work as well. 

My fault, I should have read the cover letter, but knowing what it was from
conversations I didn't look at it. And because I didn't see the patch that
it depended on in patchwork (it was hidden between Yordan's and my patch) I
didn't think it would have any dependencies, which is where I was confused.

> that patch should be part of the set as well. I'll send the v2 of the
> "Bump trace file version" with this additional patch. Strange, I do
> not see that patchset in patchwork, only in the mailing list.
> 

Yeah, I marked the series as "change requested" which hides it from the
normal view.

-- Steve