Re: [PATCH v2 14/14] perf tools: Move subcommand framework and related utils to libapi

[Josh Poimboeuf]

On Tue, Dec 08, 2015 at 07:27:32PM -0300, Arnaldo Carvalho de Melo wrote:
> Em Tue, Dec 08, 2015 at 03:48:25PM -0600, Josh Poimboeuf escreveu:
> > On Tue, Dec 08, 2015 at 04:40:26PM -0300, Arnaldo Carvalho de Melo wrote:
> > > Em Tue, Dec 08, 2015 at 01:17:00PM -0600, Josh Poimboeuf escreveu:
> > > > On Tue, Dec 08, 2015 at 04:09:31PM -0300, Arnaldo Carvalho de Melo wrote:
> > > > > Em Tue, Dec 08, 2015 at 12:49:53PM -0600, Josh Poimboeuf escreveu:
> > > > > > On Tue, Dec 08, 2015 at 07:16:26PM +0100, Jiri Olsa wrote:
> > > > > > > On Mon, Dec 07, 2015 at 10:21:52PM -0600, Josh Poimboeuf wrote:
> > > > > > > > The perf subcommand framework is needed for other tools.  Move
> > > > > > > > parse-options.c and its dependencies over to libapi.
> > > > > > > > 
> > > > > > > > Any function names with 'perf' have been renamed to something more
> > > > > > > > generic.
> > > > > > > > 
> > > > > > > > Also created a util_cfg struct for passing perf-specific configuration
> > > > > > > > to the library.  Specifying the configuration at runtime allows the same
> > > > > > > > binary to be shared by multiple tools without having to recompile it.
> > > > > > > 
> > > > > > > this patch is too big.. IMO it needs to be split into 3 parts
> > > > > > > as described in above 3 paragraphs
> > > > > > 
> > > > > > Ok, will do.
> > > > > 
> > > > > Also please rename this util_cfg struct to something more expressive,
> > > > > breaking down the patch may help in finding a better name, I guess.
> > > > 
> > > > I'm certainly open to doing so, but I'm having trouble coming up with a
> > > > better name.  The current name makes sense to me, because the struct
> > > > contains various configuration options needed by the libapi "util" code.
> > > > 
> > > > Would 'libapi_util_config' be better?  Or do you have any other
> > > > suggestions?
> > > 
> > > Please break it up into multiple pieces, as suggested by Jiri, in doing
> > > so you may find some better name.
> > >
> > > But since several are related to command environment setup, perhaps
> > > 'struct cmd_exec_env'?
> > 
> > IMO, 'struct cmd_exec_env' doesn't describe the struct accurately.  I
> > think it tangentially describes some features of some of the fields, but
> > not all of them.  That seems more confusing to me.
> 
> So do not try to keep in a single struct unrelated stuff, create two.
> :-)
> 
> > Is your complaint that the name is too vague?  If so, that's actually by
> > design, because the struct is meant to be a generic interface for
> > providing various unrelated configuration variables to libapi.
> > 
> > I've split the patch up into the above 3 paragraphs as Jiri suggested.
> > But I still don't have any ideas for a name better than 'util_cfg'
> > (other than something more verbose like 'libapi_util_config').
> > 
> > Instead of a single struct, we could consider splitting it into multiple
> > structs (e.g., one for exec_cmd.c, one for parse-options.c, and one for
> > pager.c).  But the 'exec_name' field is used by multiple files, so it
> 
> Yeah, got the same conclusion some lines above :)

Ok.  Instead of per-file structs, if there are no objections, I think
I'll go with per-file init functions.  For example:

void exec_cmd_init(const char *exec_name, const char *prefix,
		   const char *exec_path, const char *exec_path_env);

void pager_init(const char *pager_env);

I think that would be a clearer and less surprising interface.

> > wouldn't necessarily be a clean split.  It would also possibly create
> > more room for error for the users of libapi, since there would then be
> > three config interfaces instead of one.
> 
> Humm, and now that you talk... libapi was supposed to be just sugar
> coating kernel APIs, perhaps we need to put it somewhere else in
> tools/lib/ than in tools/lib/api/?

Ah, I didn't realize libapi was a kernel API abstraction library.  Shall
we put it in tools/lib/util instead?

> Borislav, ideas?
> 
> - Arnaldo

-- 
Josh
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/