Currently I only did some tests to have a very fast and extensible API
whet it should name or were it should belongs to is debatable later ;)
@jeremiah
You post hasn't been accepted. I only noticed your commend by nabble.
>I'm am just relaying this comment for a colleague who is not on the
list. This is outside my expertise. Here is his comment:
>
>Performance is much less important for handling UTF-8 than knowing the
limitations of mbstring especially. mbstring is faster because it only
handles a small set of European languages plus some common >Japanese
characters. And why are you not comparing against the gold standard
which is the intl extension.
Performance is very important because if you have to handle with
different character sets you need to wrap each string function if you
don't won't to hard code on one extension.
The intl extension is for internationalization thats not the same as
working with different character sets.
The mbstring extension doesn't handle languages it handles character
sets and it supports some different character sets that iconv.
Greetings
Marc
On 17.06.2012 22:45, Denis Portnov wrote:
> Hi Marc,
> I think you did a great job, but now it looks like full featured
> component, so not sure it belongs to Stdlib anymore.
> Whether or not this component will be accepted, I think there should
> be unified way to work with miltibyte strings throughout ZF2
>
> Denis
>
> On 16.06.2012 2:02, Marc Bennewitz wrote:
>> Hi all,
>>
>> I did some test implementation working more simple with different
>> character sets using iconv / mbstring and native code.
>>
>> The goal was too let users decide which of the php extension use to
>> handle different character sets.
>> The test implementation did a very simple, fast and expansible wrapper
>> for iconv/mbstring/native.
>> (
https://github.com/marc-mabe/zf2/blob/string/library/Zend/Stdlib/StringUtils.php)
>>
>>
>> I also did a simple benchmark that shows the mbstring adapter is faster
>> as iconv even if wrapped with an adapter:
>> (
https://gist.github.com/2938899)
>>
>> $ php stringutils-bench.php
>> native (▒): 0.0067460536956787
>> NativeAdapter (): 0.035496950149536
>> IconvNative (ß): 0.03082799911499
>> IconvAdapter (ß): 0.038977146148682
>> MbStringNative (ß): 0.0065720081329346
>> MbStringAdapter (ß): 0.010815858840942
>>
>>
>> Example 1:
>> $stringAdapter = StringUtils::getAdapterByCharset("UTF-8");
>> $stringAdapter->strlen("ß");
>> // ...
>>
>> Example 2: (Fallback to ASCII if single byte charset)
>> try {
>> StringUtils::getAdapterByCharset($charset)->strlen($str);
>> } catch (Exception $e) {
>> if (StringUtils::isSingleByteCharset($charset)) {
>> StringUtils::getAdapterByCharset("ASCII")->strlen($str);
>> }
>> }
>>
>> What do you think about this ?
>>
>> Greetings
>> Marc
>>
>
>
>
Locked
