Geometry Boost List: Re: [geometry] Testing the development branch

Subject: Re: [geometry] Testing the development branch
From: Mateusz Loskot (mateusz_at_[hidden])
Date: 2016-03-06 16:26:05

On 6 March 2016 at 17:42, Adam Wulkiewicz <adam.wulkiewicz_at_[hidden]> wrote:
> Mateusz Loskot wrote:
>>
>> In this section of the wiki [1]
>>
>> cd libs/geometry
>> git checkout develop
>> git branch -vv
>> git pull
>
>
> AFAIU you're not talking about [1] but about [3], section "Testing the
> development branch of Boost.Geometry", is that correct?

Adam,

You are right. I copied wrong URL. Thanks for correcting me.

>> shouldn't the `cd` command read
>>
>> cd modular-boost/libs/geometry
>>
>> as presented in typical commands sequence [2]  for a Boost library
>>
>> At first glance, the section [1] reads like it is an alternative
>> solution to build Boost.Geometry develop branch against
>> e.g. Boost installed from binaries package *without* need to git
>> pull the modular boost base project, and the only requirement
>> is that b2 is in PATH. I don't think that is even possible.
>>
>
> I assumed that after pulling modular boost a user is already in BOOST_ROOT,
> so in modular-boost directory. I have nothing against changing it. Indeed it
> could be more intuitive this way, so to use the same path as TryModBoost.

My question is more about: does this section [2] talk about trying fork
inside modular Boost source tree or inside a standalone geometry-only folder?

> Regarding the section [2] or actually subsection [4]. This is an optional
> step indeed presenting a way to check newly created fork outside BOOST_ROOT
> directory, without running tests or anything, only for testing if a fork works.

Right. I think, "fork works" was confusing to me.
As it is about git capabilities, not about "does my forked geometry source
tree build and test?"

I don't propose to remove it. I just did not get its actual intentions
after first reading :)

Best regards,
-- 
Mateusz Loskot, http://mateusz.loskot.net