Re: [Boost-users] Documentation for Boost Range

Subject: Re: [Boost-users] Documentation for Boost Range
From: Christian Aichinger (Greek0_at_[hidden])
Date: 2016-05-17 17:47:41

On 17.05.2016 21:08, Neil Groves wrote:
> Missing specification of headers or return types would be defects in
> the reference documentation. I've fixed a few. I did some spot checks
> and found the expected information. If you have specifics that would
> be very helpful as they should all be corrected.
>
> I have spent a few minutes looking through the examples and they look
> fabulous. Adding examples can only be a good thing. I'd like to ensure
> that the reference documentation is not deficient as one effort, and
> to incorporate your examples as another. I have noted on the thread
> that there was concern the work might be rejected. This is not the
> case. So long as the work is correct we'll get it in. My preference is
> to have a git fork that I can simply merge in.
>
> I already have a number of examples included in the bjam that are
> tested with the regression tests. My examples look nowhere near as
> beautiful. It would be ideal if I also migrated these onto the tooling
> you have built. I'd like to have all of the output looking as
> beautiful.
>
> Thank you very much for your hard work. I'll try and assist where I
> can, but at the very least keep out of the way so that I'm not
> becoming a problem.

Thank you very much for your kind words, Neil!

The way the documentation is built is "Examples to Markdown to HTML". 
The first step is done by a Make/Shell/Python abomination, the second 
step is done by my homepage publishing system (Pelican).

To integrate it with the Boost Range doc system, I would suggest 
changing that to "Examples -> QuickBook -> *". But that would still 
impose a Python dependency on the documentation build. Is that OK for you?

Essentially, what I'm doing here can't really be done inside Bjam or 
QuickBook alone TTBOMK, so I'm wondering what external requirements are 
OK for you (e.g. Python, common UNIX tooling such such as diff or md5sum 
to compare expected and found output).

Regards,
Chris