[ltt-dev] LTTV 0.12.28

[Benjamin Poirier]

Benjamin Poirier wrote:
> Mathieu Desnoyers wrote:
>   
>> * Benjamin Poirier (benjamin.poirier at polymtl.ca) wrote:
>>     
>>> Mathieu Desnoyers wrote:
>>>       
>>>> Not much more to say about 0.12.28 than: same goal as 0.12.[26,27], but
>>>> this time it builds and run.
>>>>
>>>> Mathieu
>>>>
>>>> * Mathieu Desnoyers (compudj at krystal.dyndns.org) wrote:
>>>>         
>>>>> Hi,
>>>>>
>>>>> A small note: LTTV 0.12.27 fixes the package build. I had to temporarily
>>>>> disable the new distributed network synchronization module, because it
>>>>> failed to build on 64-bit and did not package properly with "make dist".
>>>>> I'll re-enable it when these problems are fixed.
>>>>>
>>>>> Benjamin: Please reconsider inclusion of include/linux/if_ether.h, as it
>>>>> brings dependencies on the specific kernel headers used, and lttv does
>>>>> not depend on it. If we really need this, then we have to require that
>>>>> these headers are included on the user systemes, and it will become very
>>>>> important to test the build _and_ execution on a wide range of
>>>>> architectures (x86 32/64, powerpc at the very least) and on a wide range
>>>>> of kernel header versions. The second point is that the structure of the
>>>>> lttv/sync/ makefiles is a bit weird: the .o are built in the upper-level
>>>>> lttv/ (should be in the lttv/sync/ subdirectory instead). I have
>>>>>           
>>> That was my first approach, build inside sync/ as a static library and
>>> link to it from below. I don't remember the specifics but I ran into
>>> troubles, I think with libtool because I had to specify manually which
>>> symbols to export. The automake manual says its ok to do it as it is
>>> now too:
>>> http://sources.redhat.com/automake/automake.html#Alternative
>>>
>>>       
>> Well, the point of putting sync/ files in a subdirectory is to separate
>> it from the rest, right ?
>>
>> As we build the files from the upper-level makefile, it:
>>
>> a) pollutes the upper level makefile
>> b) pollutes the build tree with unrelated files in the upper level dir.
>>
>> What were these issues you ran into exactly ?
>>
>>     
At the moment the sync functions are linked into the lttv binary but
used from modules. This is the same thing for the rest of the "core
functions" under lttv/lttv.

If the code under sync/ is built as a static library and linked into the
lttv binary, nothing is linked in because the symbols are only
referenced in shared libraries. The --whole-archive option to ld is
needed, but that is not supported by libtool. This is the trouble I was
talking about.
>>>>> temporarily added the lttv/sync/*.h headers as noinst_HEADERS so the
>>>>> tarball created by make dist builds.
>>>>>
>>>>> Commits:
>>>>> c73a133267f0de33bd7303de34a8f3e0ef90422a
>>>>> ccf1db700d7477f4cd825ac435e2632443ee51f2
>>>>>
>>>>> Are the ones that disable the network sync module and add the
>>>>> noinst_HEADERS.
>>>>>           
>>> This should fix both problems.
>>>
>>> Please pull from
>>> 	git://hal.drap.polymtl.ca/lttv-sync.git for-pmf
>>>
>>>       
>> I'll wait until the build structure discussion settles before pulling.
>> There is nothing like discussing problems openly to get them solved in
>> the best interest of everyone. ;)
>>     
>
> I'll look into cleaning up the makefiles. Why not merge those patches
> now, they fix the other problems?
>
>   
>> Thanks,
>>
>> Mathieu
>>
>>     
>>> Benjamin Poirier (3):
>>>       Partial revert "Disable sync/"
>>>       Add header files in SOURCES
>>>       Remove dependency on kernel headers
>>>
>>>  lttv/lttv/Makefile.am                              |   47 +++++++++++------
>>>  lttv/lttv/sync/Makefile.am                         |   57 ++++++++------------
>>>  lttv/lttv/sync/event_processing_lttng_standard.c   |    6 ++-
>>>  lttv/modules/gui/lttvwindow/lttvwindow/callbacks.c |    5 +--
>>>  .../gui/lttvwindow/lttvwindow/init_module.c        |    3 +-
>>>  lttv/modules/text/batchAnalysis.c                  |    6 +--
>>>  6 files changed, 63 insertions(+), 61 deletions(-)
>>>
>>>       
>>>>> Thanks,
>>>>>
>>>>> Mathieu
>>>>>
>>>>>
>>>>> * Mathieu Desnoyers (compudj at krystal.dyndns.org) wrote:
>>>>>           
>>>>>> Hi everyone,
>>>>>>
>>>>>> I just released LTTV 0.12.26, which fixes a rare bug in the detailed
>>>>>> event list "backward" scroll. See commit
>>>>>> 0903a18630f5ec2087c84e2575bf3e7f2d938210 for details.
>>>>>>
>>>>>> Mathieu
>>>>>>
>>>>>> -- 
>>>>>> Mathieu Desnoyers
>>>>>> OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68
>>>>>>
>>>>>> _______________________________________________
>>>>>> ltt-dev mailing list
>>>>>> ltt-dev at lists.casi.polymtl.ca
>>>>>> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>>>>>>
>>>>>>             
>>>>> -- 
>>>>> Mathieu Desnoyers
>>>>> OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68
>>>>>
>>>>> _______________________________________________
>>>>> ltt-dev mailing list
>>>>> ltt-dev at lists.casi.polymtl.ca
>>>>> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>>>>>
>>>>>           
>>> _______________________________________________
>>> ltt-dev mailing list
>>> ltt-dev at lists.casi.polymtl.ca
>>> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>>>
>>>       
>
>
> _______________________________________________
> ltt-dev mailing list
> ltt-dev at lists.casi.polymtl.ca
> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>