On 06/12/2018 11:15, Jan Beulich wrote: >>>> On 06.12.18 at 11:01, <[email protected]> wrote: >> On 06/12/2018 10:50, Jan Beulich wrote: >>>>>> On 06.12.18 at 09:06, <[email protected]> wrote: >>>> @@ -477,7 +478,8 @@ unsigned long long parse_size_and_unit(const char *s, >> const char **ps) >>>> s1++; >>>> break; >>>> default: >>>> - ret <<= 10; /* default to kB */ >>>> + if ( *s1 && *s1 != no_size ) >>>> + ret <<= 10; /* default to kB */ >>>> break; >>> >>> So did you figure anything wrong with simply special casing '%' >>> here? '%' is a form of "unit", after all. Perhaps demanding "ps" >>> to be non-NULL (to have some form of indication the caller will >>> check the suffix char) might be reasonable. >> >> I thought a more general approach would be better. I can use the simple >> '%' special case with the ps check in case you prefer that. > > I'd indeed prefer that, first and foremost because that other > function's name becomes too long for my taste, but give others > a chance to voice differing opinions.
In case the function name is your main concern I could rename it to parse_size_or_int() or similar. I'm fine either way, so lets wait for some days and act then according to the comments. Juergen _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
