kicad-developers team mailing list archive

[Wayne Stambaugh <stambaughw@xxxxxxxxxxx>]

The reason you don't see either of these problems after the first build
is that boost and the source files created by the make_lexer() cmake
macro are already created so all of the dependencies are satisfied and
the build completes without issues.  You would only see the errors on a
new source build and possibly after running `make clean`.  Although I'm
not sure all of these files get remove on `make clean`.  It's hard to
believe that we have been bit by the missing files generated by
make_lexer() before now because none of the generated files are marked
as dependencies when building their respective targets.  I should have a
fix later today but I have some testing to do.

On 10/22/2014 1:59 AM, Bernhard Stegmaier wrote:
> Hi,
> 
> I didn’t use any other version of boost… always use the custom version that still comes with KiCad.
> I also did a lot of builds recently, mostly with -j4 and never had problems with that…
> 
> 
> Regards,
> Bernhard
> 
> On 21.10.2014, at 21:41, Wayne Stambaugh <stambaughw@xxxxxxxxxxx> wrote:
> 
>> Adam,
>>
>> Try applying the attached patch.  It looks like on osx builds, boost is
>> not being made a dependency so the boost build is not complete before
>> the common library is getting built on parallel builds.  I think
>> Bernhard was using a patched version of boost 1.56 which he has
>> installed on his system so he would not have seen this problem since he
>> wasn't using our custom 1.54 version of boost.  If this doesn't fix it,
>> maybe Bernhard can take a look at it.
>>
>> Wayne
>>
>> On 10/21/2014 3:07 PM, Adam Wolf wrote:
>>> I don't get these errors when building with -j1 *after* a clean.
>>>
>>> Adam Wolf
>>>
>>> On Tue, Oct 21, 2014 at 1:56 PM, Bob Gustafson <bobgus@xxxxxxx
>>> <mailto:bobgus@xxxxxxx>> wrote:
>>>
>>>    Do you get that error when building with -j1 ?
>>>
>>>
>>>    On 10/21/2014 01:14 PM, Adam Wolf wrote:
>>>>    Hi folks,
>>>>
>>>>    I'm doing a *lot* of builds prepping these nightly builds, and
>>>>    especially on OS X, I'm seeing a lot of issues with parallel
>>>>    builds, where something like this happens:
>>>>
>>>>    In file included from /Users/jenkins/remoteroot/workspace/KiCadMacBuild/kicad/common/draw_panel_gal.cpp:33:
>>>>    /Users/jenkins/remoteroot/workspace/KiCadMacBuild/kicad/include/view/view.h:30:10: fatal error: 'boost/unordered/unordered_map.hpp' file not found
>>>>    #include <boost/unordered/unordered_map.hpp>
>>>>
>>>>    I have a little logic in there where I first try to build with
>>>>    -j4, and if it fails, i rerun make with -j1, but that doesn't seem
>>>>    to help.
>>>>
>>>>    Any thoughts, folks?
>>>>
>>>>    Adam Wolf
>>>>    Cofounder and Engineer
>>>>    W&L, LLC
>>>>
>>>>
>>>
>>>
>>>    _______________________________________________
>>>    Mailing list: https://launchpad.net/~kicad-developers
>>>    Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
>>>    <mailto:kicad-developers@xxxxxxxxxxxxxxxxxxx>
>>>    Unsubscribe : https://launchpad.net/~kicad-developers
>>>    More help   : https://help.launchpad.net/ListHelp
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Mailing list: https://launchpad.net/~kicad-developers
>>> Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
>>> Unsubscribe : https://launchpad.net/~kicad-developers
>>> More help   : https://help.launchpad.net/ListHelp
>>>
>>
>> <osx-boost-build-fix.patch>_______________________________________________
>> Mailing list: https://launchpad.net/~kicad-developers
>> Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~kicad-developers
>> More help   : https://help.launchpad.net/ListHelp
> 
>