On 06/03/2014 08:22 AM, Amos Jeffries wrote:
> On 4/06/2014 1:08 a.m., Alex Rousskov wrote:
>> On 06/03/2014 04:46 AM, Amos Jeffries wrote:
>>
>>> This replaces the SBuf::trim() to use CharacterSet instead of an SBuf()
>>> list of characters and memchr()
>>>
>>> It seems to be faster for CharacterSet lookup than repeated memchr
>>> calls, but Im not certain of that. It is certainly makes simpler parser
>>> code with trim and a predefined CharacterSet than static SBuf set of chars.
>>
>> I agree that CharacterSet membership test should be faster than repeated
>> memchr() calls.
>>
>> No objections to this patch, although I suspect that any code calling
>> SBuf::trim() should actually use Tokenizer instead; we may be optimizing
>> code that should not be used in the first place.
>
> The use-case that brought this up is scanning mime header lines.
>
> Tokenizer tok(...);
> SBuf line;
> while (tok.prefix(line, CharacterSet::LF)) {
> // drop optional trailing CR* sequence
> line.trim(CharacterSet::CR, false, true);
The above does not make sense to me: After tok.prefix(), "line" will
contain LF characters only. There will be no CR characters to trim.
Alex.
