Re: boost.url review

From: Rainer Deyke (rdeyke_at_[hidden])
Date: 2022-08-21 14:14:52

Here is my quick review of Boost.URL.  I vote to ACCEPT this library. 
It may not be perfect, but it's a good enough solution for a real problem.

I do have some concerns:

   - The lack of IRI support is unfortunate.  It's 2022; we should all 
be writing software with Unicode support by default.  However, this can 
be built on top of Boost.URL, and isn't needed in all cases.

   - The "string+offsets" implementation means that a lot of characters 
need to be copied every time something at the left end of a long URL is 
modified - for example when changing an http URL to use https.  This is 
unlikely to cause a problem in most circumstances, but it's something to 
be aware of.

On 13.08.22 04:29, Klemens Morgenstern via Boost wrote:

> Some questions to consider:
> 
>   - Does this library bring real benefit to C++ developers for real world
> use-case?

Yes.

>   - Do you have an application for this library?

Yes.

>   - Does the API match with current best practices?

It's good enough.

>   - Is the documentation helpful and clear?

Yes, but I did find a few errors that should be fixed.  One of the 
examples on https://master.url.cpp.al/url/containers/authority.html does 
not match the text.  I have to admit I skipped over the customization 
section.

>   - Did you try to use it? What problems or surprises did you encounter?

No.

>   - What is your evaluation of the implementation?

Like many Boost libraries, I found the implementation fairly opaque. 
Class url delegates to url_base, which delegates to url_view_base, which 
delegates to detail::url_impl, with much of the functionality spread out 
over various view classes.  This isn't a problem per se, but it makes 
casual reading difficult.

-- 
Rainer Deyke (rainerd_at_[hidden])