Re: [Boost-build] [git] genrating forward headers

Subject: Re: [Boost-build] [git] genrating forward headers
From: Bjørn Roald (bjorn_at_[hidden])
Date: 2013-07-22 10:18:25

On 07/19/2013 03:45 PM, Beman Dawes wrote:
> Bjørn Roald <bjorn <at> 4roald.org> writes:
>
>> On 07/19/2013 03:26 AM, Beman Dawes wrote:
>>> Can we please get these patches applied to trunk?
>> ...
>>
>> Agree, someone please get this into trunk.  The patches I used are
>> attached here:
>> http://lists.boost.org/boost-build/2013/07/26861.php
> Steven, since you understand what is going on with headers, could you look
> these patches over and then apply them?

That would be appreciated.  I don't know if Steven agree with the 
patches in current form, and he certainly suggested adjustments to fix 
regression testing in the status folder here:

http://lists.boost.org/boost-build/2013/06/26812.php

which JÅ©rgen tested without complete success.  I basically get the same 
result as JÅ©rgen with Jamroot
"alias headers" like this:

alias headers : : <dependency>$(all-headers)-headers 
<dependency>numeric-$(numeric-headers)-headers : : <include>.  ;

I am not sure that is what Steven suggested.  However with this the 
regression tests int the status directory are running successfully for 
the vast majority of libraries.

The patches has a number of tests commented out in status/Jamfile.v2 
which probably is not intended to be applied.  I did not try to 
uncomment them when I tested, I assume there was or are failures there.

>> The problem with execute bit on the unix shell scripts need to be fixed
>> in conversion I guess. In the meantime the following steps on Linux
>> allow testing:
>>
>> chmod 775 bootstrap.sh tools/build/v2/engine/build.sh
>> ./bootstrap.sh
>> ./b2 headers
>> ./b2
> Perhaps submit a ticket? I'm not sure to which tracker it should go to.

fixed by Dave here https://github.com/ryppl/Boost2Git/issues/33.

>> Note I have not tested on windows.
> I've been testing on Windows. After deleting the extra headers in
> type_traits, deleting the boost header hierarchy, and rerunning "b2
> headers", the regression tests for config, system, and filesystem are now
> working.

Ok, it seems to me that lacking b2 support for generating header 
symlinks is the primary blocker for progress with git conversion now.  
Someone with with better knowledge of b2 than me need to step up.  At 
least I need some directions as to what I may try.

--
Bjørn