Re: [PATCH 3/3] tools/lib/traceevent: Implement new traceevent APIs for accessing struct tep_handler fields

[Steven Rostedt <rostedt@goodmis.org>]

On Fri, 22 Mar 2019 11:33:47 +0000
Tzvetomir Stoyanov <tstoyanov@vmware.com> wrote:


> > >
> > > int tep_host_bigendian(void);  
> >
> > Rename to: bool tep_is_bigendian()  
> 
> Steven, int tep_host_bigendian(void) actually checks the endianness of
> the host -
> I think it makes sens "host" to be part of API's name. The other one:
> int tep_is_host_bigendian(struct tep_handle *pevent) checks the
> endianness stored
> in the tep_handle, I think we should remove "host" from its name.
> 

I hate the word "host" here. Because what would you get if you run this
on a guest? You get the "guest" endianess (which may possibly be not
the same as the host (if emulated)).


I want to keep this as "tep_is_bigendian()" as it doesn't take a
parameter, and is obvious to what it is returning.

> > > int tep_file_bigendian(struct tep_handle *pevent);  
> >
> > Rename to: bool tep_is_file_bigendian()
> >  
> > > void tep_set_file_bigendian(struct tep_handle *pevent, enum tep_endian endian);
> > > int tep_is_host_bigendian(struct tep_handle *pevent);  
> >
> > Rename to: bool tep_is_host_bigendian()
> >  

Let's rename this to:

void tep_set_local_bigendian()
bool tep_is_local_bigendian()

It can be argued that this could be for the above (test this machine),
but since the "test this machine" doesn't take a parameter, but "test
the pevent bigendian" does, I want it to be consistent with
tep_is_file_bigendian(), which also takes a parameter.

   bool tep_is_bigendian(void);
   bool tep_is_file_bigendian(struct tep_handle *tep);
   bool tep_is_local_bigendian(struct tep_handle *tep);

seems more consistent than:

   bool tep_is_local_bigendian(void);
   bool tep_is_file_bigendian(struct tep_handle *tep);
   bool tep_is_bigendian(struct tep_handle *tep);

So lets go with the first one.

Make sense?

-- Steve