Subject: Re: [boost] [1.49.0] Release candidates available
From: Barend Gehrels (barend_at_[hidden])
Date: 2012-02-20 17:36:59


Hi Daniel,

On 20-2-2012 22:55, Daniel James wrote:
> On 20 February 2012 21:11, Barend Gehrels<barend_at_[hidden]> wrote:
>> On 20-2-2012 21:03, Daniel James wrote:
>>> On 20 February 2012 18:39, Barend Gehrels<barend_at_[hidden]> wrote:
>>>
>>> It stopped working after changeset 76991, which changed the geometry
>>> script to use a relative path to the b2 executable at the root of
>>> boost, but it isn't there in my setup. I guess this was done to avoid
>>> using trunk b2 with the release branch. The easiest thing to do would
>>> be to change the script to copy b2 into the expected location and run
>>> another documentation build. I'll do that now.
>> Thanks for your answer and for your action.
> I've run the build again, can you check it? (boost-release-docs.7z).

This is the correct one! Thanks!
Great that it is solved so soon.

>
>> I'm very sorry about that - indeed bjam causes big problems at my system on
>> the release branch - it does not work anymore (because of project.jam). I've
>> complained about that on this list but it seems I'm the only one, which I
>> cannot imagine. Running b2 from release branch is fine - which is the cause
>> of the change.
> I've had the same problem.
>
> http://thread.gmane.org/gmane.comp.lib.boost.build/24525/
>
> I think Steven's solution would work, but no one has implemented it. I
> should probably create a ticket.

I see. So indeed we can expect build problems starting from this release...

>
>> Anyway - sorry about breaking it that way, I should have been more careful.
>> Thanks for taking action.
>>
>> What's the best next step? Can it stay like it is or should I revert it?
> If the updated build script worked fine then leave it.
>
> It might be useful to check the return value from the call and exit
> with an error status if appropriate, that will hopefully cause the
> build script to fail so that we'd catch this kind of thing earlier. I
> might also add a check that certain files have been created.

OK, I'll try to include that in the script.

Thanks again,
Barend