Re: [boost] base64?

Subject: Re: [boost] base64?
From: Robert Ramey (ramey_at_[hidden])
Date: 2017-07-24 17:21:53

On 7/24/17 9:54 AM, Roger Leigh via Boost wrote:
> 
> 
> On 24/07/17 16:28, Vinnie Falco via Boost wrote:
>> On Mon, Jul 24, 2017 at 8:27 AM, Vinnie Falco <vinnie.falco_at_[hidden]> 
>> wrote:
>>> On Mon, Jul 24, 2017 at 8:24 AM, Greger via Boost 
>>> <boost_at_[hidden]> wrote:
>>> I also see this note:
>>>
>>>> It is available for a moderate compensation. Let me know by
>>>> email and I will send you the source.
>>
>> Is this library open source or commercial in nature? What license is
>> it furnished under? Also, how is this better than the numerous other
>> base64 libraries (which are free) such as this one:
>>
>> <http://renenyffenegger.ch/notes/development/Base64/Encoding-and-decoding-base-64-with-cpp> 
>>
> 
> Or 
> https://github.com/ome/ome-common-cpp/blob/master/lib/ome/common/base64.h
> 
> which I probably wrote based on a similar source when I realised the 
> Boost base64 support was broken and needed a quick implementation.  It 
> avoids allocation issues by allowing the use of iterators including 
> back_inserter.
> 
> We already have base64 conversion in 
> boost/archive/iterators/*base64*.hpp.  It works but screws up the 
> whitespace 
> (https://stackoverflow.com/questions/10521581/base64-encode-using-boost-throw-exception/10973348#10973348). 
>   It's a shame the various adapters don't work properly together; is 
> there any prospect of this being fixed, or is the existing design simply 
> too inflexible to make a proper base64 implementation?

I'm guessing it would be pretty easy to fix.  I base this on the design 
which is built up as a stack of iterator pairs.  This is documented in 
the serialization library under

http://www.boost.org/doc/libs/1_64_0/libs/serialization/doc/dataflow.html

This is used for the base64 stuff in the serialization library which as 
far as I know is correct.  That doesn't include adding on padding at the 
end - which maybe it should.

In any case making a stand along base64 component would require 
understanding the issues, making a small variation on the one in the 
library, make a test and adding it the test suite and making a 
commitment to support it in the future.  I just haven't had the 
motivation to do that - even though I think the code and design are very 
convenient and flexible.  Result is decent performance as well.  FYI it 
was made before boost ranges and can be considered a demonstration of 
the validity of using the range idea.

Robert Ramey

> 
> 
> Regards,
> Roger
> 
> _______________________________________________
> Unsubscribe & other changes: 
> http://lists.boost.org/mailman/listinfo.cgi/boost
>