Re: [Boost-cmake] Progress Report

Subject: Re: [Boost-cmake] Progress Report
From: Michael Jackson (mike.jackson_at_[hidden])
Date: 2008-11-07 10:27:31

On Nov 5, 2008, at 7:56 PM, Michael Jackson wrote:

>
> On Nov 4, 2008, at 2:21 PM, Michael Jackson wrote:
>
>>
>> On Nov 4, 2008, at 2:11 PM, Beman Dawes wrote:
>>
>>> On Tue, Nov 4, 2008 at 12:26 PM, Michael Jackson
>>> <mike.jackson_at_[hidden]> wrote:
>>>> I have been working hard on updating the CMake build system to  
>>>> match the
>>>> Boost trunk repository. As it stands I have been able to compile  
>>>> all the
>>>> libraries in static and dynamic multi-threaded debug modes. I can  
>>>> get about
>>>> 50% of the regression test to compile.
>>>
>>> Thanks for the update!
>>>
>>> What kinds of problems are you running into with the 50% that aren't
>>> compiling? Anything serious?
>>>
>>> --Beman
>>
>> Nothing serious at this point. It just takes time to compile, wait  
>> for compile error (which is all about include paths at this point),  
>> add the proper module dependency, re-run cmake, re-run make, wait  
>> for compile error....
>>
>> There have been some general bug fixes and enhancements to the core  
>> Boost-CMake files which ate up some time. I did add the ability of  
>> the graphviz output to have those libraries that are actually  
>> compiled to have a colored node which makes it easier to see those  
>> libraries in the dependency graph.
>>
>> For those interested, here is the latest dependency graph (which  
>> will change as soon as I do the next compile cycle most likely) but  
>> gives you an idea of what boost is looking like.
>>
>> <http://www.bluequartz.net/dependencies.pdf>
>>
>> _________________________________________________________
>> Mike Jackson                  mike.jackson_at_[hidden]
>>           www.bluequartz.net
>>
>
> Just an update. I was finally able to get the regression tests to  
> compile to 100%. What this means is that if the regression test  
> involved creating an executable then that test was compiled but not  
> yet actually executed. The regression tests that are currently being  
> compiled were left over from the previous CMake files so some newer  
> tests are obviously not included at this point. Also, I have not  
> updated my svn working directory since sometime last week.
>
> I did actually try "make test" and I get into the exception  
> regression tests and I get a hard failure which stops everything.  
> The test that is failing is a test where a source file is actually  
> attempted to be compiled and is failing. I believe I know the  
> problem so I am going to dive into that next.
>
> Again, this is with a static-multi-threaded build only on OS X  
> 10.5.5 with gcc (Xcode 3.1). I have yet to try this on windows. I  
> Have VS express 2008 at my disposal to try sometime.
>
> _________________________________________________________
> Mike Jackson                  mike.jackson_at_[hidden]
>            www.bluequartz.net
>

I am making good progress updating the CMake build system. I have most  
everything building and some of the regression tests passing. I pulled  
the latest from the trunk last night and am doing a recompile at this  
point so that I can do an svn commit later today with all the updated  
and new cmake files.

Mike