Re: [PATCH] eal: add helper to skip whitespaces

[Bruce Richardson <bruce.richardson@intel.com>]

On Thu, Feb 15, 2024 at 10:08:40AM +0100, David Marchand wrote:
> On Wed, Feb 14, 2024 at 2:07 PM Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > On Wed, Feb 14, 2024 at 01:12:19PM +0100, David Marchand wrote:
> > > Reduce code duplication by providing a simple inline helper.
> > >
> > > Signed-off-by: David Marchand <david.marchand@redhat.com>
> > > ---
> > >  app/graph/utils.c                          | 13 ++------
> > >  app/test-eventdev/parser.c                 | 14 ++++-----
> > >  app/test-eventdev/parser.h                 |  8 -----
> > >  app/test-mldev/parser.c                    | 17 ++++++-----
> > >  app/test-mldev/parser.h                    |  8 -----
> > >  app/test-pmd/cmdline_tm.c                  | 13 ++------
> > >  app/test/test_string_fns.c                 | 35 ++++++++++++++++++++++
> > >  examples/fips_validation/fips_validation.c | 16 +++-------
> > >  examples/ip_pipeline/parser.c              | 14 ++++-----
> > >  examples/ip_pipeline/parser.h              |  8 -----
> > >  examples/pipeline/cli.c                    | 13 ++------
> > >  lib/eal/include/rte_string_fns.h           | 27 +++++++++++++++++
> > >  12 files changed, 98 insertions(+), 88 deletions(-)
> > >
> > > diff --git a/app/graph/utils.c b/app/graph/utils.c
> > > index c7b6ae83cf..1f5ee68273 100644
> > > --- a/app/graph/utils.c
> > > +++ b/app/graph/utils.c
> > > @@ -9,17 +9,10 @@
> > >  #include <string.h>
> > >
> > >  #include <rte_common.h>
> > > +#include <rte_string_fns.h>
> > >
> > >  #include "module_api.h"
> > >
> > > -#define white_spaces_skip(pos)                       \
> > > -({                                           \
> > > -     __typeof__(pos) _p = (pos);             \
> > > -     for ( ; isspace(*_p); _p++)             \
> > > -             ;                               \
> > > -     _p;                                     \
> > > -})
> > > -
> > >  static void
> > >  hex_string_to_uint64(uint64_t *dst, const char *hexs)
> > >  {
> > > @@ -47,7 +40,7 @@ parser_uint64_read(uint64_t *value, const char *p)
> > >       char *next;
> > >       uint64_t val;
> > >
> > > -     p = white_spaces_skip(p);
> > > +     p = rte_str_skip_whitespaces(p);
> > >       if (!isdigit(*p))
> > >               return -EINVAL;
> > >
> > > @@ -73,7 +66,7 @@ parser_uint64_read(uint64_t *value, const char *p)
> > >               break;
> > >       }
> > >
> > > -     p = white_spaces_skip(p);
> > > +     p = rte_str_skip_whitespaces(p);
> >
> > I like the idea of this. However, a question on naming. Do we want to make
> > it clear in the name that it only skips leading whitespace, and doesn't
> > e.g. trim whitespace off the end? How about:
> >
> > rte_str_skip_leading_spaces()
> >
> > It's longer, but I wonder if it's a bit clearer.
> 
> It is clearer.
> 
> 
> >
> > Also, might it be worthwhile providing an "rte_str_ltrim()" function along
> > with this, basically combining the skip_spaces call, with an memmove? Most
> > string handling in DPDK is not perf-sensitive and so a function with
> > similar name and behaviour to the python ltrim function may be appreciated
> > for its simplicity.
> 
> Well, most intree users won't touch the data (and simply pass around
> const char * pointers).
> This new API you propose is probably fine, but I suspect it won't be
> of much use for now.
> 

Ok, good point. I'd suggest otherwise just going with your patch, perhaps
with the tweaked name.

/Bruce