Re: [PATCH v2 2/7] parse-options: add precision handling for PARSE_OPT_CMDMODE

[Patrick Steinhardt <ps@pks.im>]

On Wed, Jul 09, 2025 at 05:05:35PM +0200, René Scharfe wrote:
> On 7/9/25 3:58 PM, Patrick Steinhardt wrote:
> > On Wed, Jul 09, 2025 at 11:45:14AM +0200, René Scharfe wrote:
> >>
> >> Call the function that does the raw casting do_get_int_value() to
> >> reserve the name get_int_value() for a more friendly wrapper we're
> >> going to introduce in one of the next patches.
> 
> >> diff --git a/parse-options.c b/parse-options.c
> >> index 68ff494492..ddac008a5e 100644
> >> --- a/parse-options.c
> >> +++ b/parse-options.c
> >> @@ -68,6 +68,26 @@ static char *fix_filename(const char *prefix, const char *file)
> >>  		return prefix_filename_except_for_dash(prefix, file);
> >>  }
> >>  
> >> +static int do_get_int_value(const void *value, size_t precision, intmax_t *ret)
> > 
> > Nit: after the fourth patch we have `do_get_int_value()` and
> > `get_int_value()`, where the major difference is that the latter dies if
> > we failed to parse the value. It might be easier to discern which is
> > which if we called them `get_int_value()` and `get_int_value_or_die()`.
> 
> That would be misleading because get_int_value() doesn't die() like a
> function from write-or-die.c, it BUGs instead.  I don't think it makes
> sense to advertise the presence of assertions in a function's name.
> But we do have a tradition of using a prefix of "do_" with wrapped
> functions that have a more raw interface and do the actual work.
> 
> Nit: They don't parse, but cast a void pointer to the appropriate type
> and dereference it.

Fair enough. I don't mind it much either way, thanks!

Patrick