Paolo Bonzini wrote:
> On 09/15/2011 01:07 PM, Stefano Lattarini wrote:
>> Maybe I'm missing something fundamental, but this looks defintely wrong
>> to me.  The code guarded by this "if" will now never be run when
>> MBS_SUPPORT is 0, while previously is was run if `match_icase' was 1
>> and MBS_SUPPORT was 0.  Maybe you intended to use like this instead?
>>
>>    if (match_icase&&  (!MBS_SUPPORT || MB_CUR_MAX == 1)) { ... }
>
> Good catch.

Good catch indeed.  I've fixed that.

However, do note that while technically it was a bug (and at
the front of the series!), a later change fixed it (albeit unwittingly),

  maint: dfa: simplify multi-byte-related conditionals

    -      assert(!MBS_SUPPORT || MB_CUR_MAX == 1);
    +      assert(MB_CUR_MAX == 1);

Reply via email to