Re: ... linker problems

[Antonio Coralles]

Ralf Wildenhues wrote:

> * Antonio Coralles wrote on Sat, Jan 22, 2005 at 12:08:32AM CET:
[snip]

> > /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3.5/../../../../i686-pc-linux-gnu/bin/ld: 
>
> > `.gnu.linkonce.t._ZN5boost9date_time23gregorian_calendar_baseINS0_19year_month_day_baseINS_9gregorian9greg_yearENS3_10greg_monthENS3_8greg_dayEEEmE16end_of_month_dayES4_S5_' 
>
> > referenced in section `.rodata' of testUtils/testutils.o: defined in
> > discarded section
> > `.gnu.linkonce.t._ZN5boost9date_time23gregorian_calendar_baseINS0_19year_month_day_baseINS_9gregorian9greg_yearENS3_10greg_monthENS3_8greg_dayEEEmE16end_of_month_dayES4_S5_' 
>
> > of testUtils/testutils.o
> > ...
> >
> > This messages don't appear with my normal makefiles ...

[snip]

Well this problem is allready solved. I can't tell what it was exactly, 
but I still get this error when I try to mix object or *.a files 
compiled with different flags. The problems I'm experiencing now are 
explained in "strange problems".

> Well.  You don't tell us what you did before and what link command line
> Automake issues.  How are we to know the difference then?
Ok, here is my old link command line:

g++ -Wall -g `pkg-config gtkmm-2.4 --libs` 
../../../objectFiles/messageMem.o -lsqlite -lboost_date_time 
../../../objectFiles/worker.o ../../../objectFiles/schedule.o 
../../../objectFiles/dataStorage.o ../../../objectFiles/timePeriod.o 
../../../objectFiles/judge.o ../../../objectFiles/judgeRules.o 
../../../objectFiles/overtimeExpert.o 
../../../objectFiles/overtimeRules.o ../../../objectFiles/calendar.o 
../../../objectFiles/organizer.o ../../../objectFiles/exceptionHelpers.o 
../../../objectFiles/simpleUtils.o ../../../objectFiles/extendedUtils.o 
../../../objectFiles/treaty.o ../../../objectFiles/poolLabour.o 
../../../objectFiles/parsers.o ../../../objectFiles/coreUtilsPrivate.o 
../../../objectFiles/dataWriters.o ../../../objectFiles/htmlExport.o 
../../../objectFiles/misc.o ../../../objectFiles/mainWindow.o 
../../../objectFiles/shiftTable.o ../../../objectFiles/shiftTableBox.o 
../../../objectFiles/monthBox.o ../../../objectFiles/tableField.o 
../../../objectFiles/textField.o ../../../objectFiles/coloredTextField.o 
../../../objectFiles/workerScheduleField.o 
../../../objectFiles/baseView.o ../../../objectFiles/shiftsView.o 
../../../objectFiles/shiftsViewConnected.o 
../../../objectFiles/attributeWindow.o 
../../../objectFiles/shiftAttributeWindow.o 
../../../objectFiles/shiftAttributeWindowConnected.o 
../../../objectFiles/treatiesView.o 
../../../objectFiles/treatiesViewConnected.o 
../../../objectFiles/treatyAttributeMainWindow.o 
../../../objectFiles/treatyOvertimeWindow.o 
../../../objectFiles/treatyOvertimeWindowConnected.o 
../../../objectFiles/treatyRulesWindow.o 
../../../objectFiles/treatyRulesWindowConnected.o 
../../../objectFiles/needsWindow.o ../../../objectFiles/workersView.o 
../../../objectFiles/workersViewConnected.o 
../../../objectFiles/workerAttributeWindow.o 
../../../objectFiles/invalidCharacters.o 
../../../objectFiles/shiftPrefWindow.o 
../../../objectFiles/weekdayPrefWindow.o 
../../../objectFiles/statFields.o ../../../objectFiles/dateFields.o 
../../../objectFiles/poolScheduleField.o ../../../objectFiles/main.o -o 
mainTest; \

And here the new one:
g++  -g -Wall   -o dpm  dpm-main.o interface/mainWindow/lib_mainwindow.a 
interface/listView/lib_listview.a 
interface/attributeWindow/lib_attributewindow.a 
interface/shiftTable/lib_shifttable.a interface/misc/lib_misc.a 
htmlExport/lib_htmlexport.a messageMem/lib_messagemem.a 
dataStorage/lib_datastorage.a utility/lib_utils.a 
organizer/lib_organizer.a manpower/lib_manpower.a 
labourLaw/lib_labourlaw.a schedule/lib_schedule.a 
helperFunctions/lib_helperfunctions.a shared/lib_shared.a  
-Wl,--export-dynamic -lgtkmm-2.4 -lgdkmm-2.4 -latkmm-1.6 -lgtk-x11-2.0 
-lpangomm-1.4 -lglibmm-2.4 -lsigc-2.0 -lgdk-x11-2.0 -latk-1.0 
-lgdk_pixbuf-2.0 -lm -lpangoxft-1.0 -lpangox-1.0 -lpango-1.0 
-lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0   -L/usr/lib 
-lboost_date_time-mt-d -lsqlite

As one may notice: I switched from -lboost_date_time to 
-lboost_date_time_mt-d, but it didn't work without the "_mt-d" either. 
I've also partly reorganized my code, but only after experiencing 
problems, because I thought that may help. The various *.a files are 
created like this:

make[2]: Entering directory `/home/antonio/coding/dpm/10/src/shared'
source='timePeriod.cc' object='timePeriod.o' libtool=no \
depfile='.deps/timePeriod.Po' tmpdepfile='.deps/timePeriod.TPo' \
depmode=gcc3 /bin/sh ../../depcomp \
g++ -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" 
-DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE=\"dpm\" 
-DVERSION=\"0.0.42\"  -I. -I.     -Wall -g -c -o timePeriod.o `test -f 
timePeriod.cc || echo './'`timePeriod.cc
source='calendar.cc' object='calendar.o' libtool=no \
depfile='.deps/calendar.Po' tmpdepfile='.deps/calendar.TPo' \
depmode=gcc3 /bin/sh ../../depcomp \
g++ -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" 
-DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE=\"dpm\" 
-DVERSION=\"0.0.42\"  -I. -I.     -Wall -g -c -o calendar.o `test -f 
calendar.cc || echo './'`calendar.cc
source='simpleUtils.cc' object='simpleUtils.o' libtool=no \
depfile='.deps/simpleUtils.Po' tmpdepfile='.deps/simpleUtils.TPo' \
depmode=gcc3 /bin/sh ../../depcomp \
g++ -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" 
-DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE=\"dpm\" 
-DVERSION=\"0.0.42\"  -I. -I.     -Wall -g -c -o simpleUtils.o `test -f 
simpleUtils.cc || echo './'`simpleUtils.cc
rm -f lib_shared.a
ar cru lib_shared.a timePeriod.o calendar.o simpleUtils.o
ranlib lib_shared.a
make[2]: Leaving directory `/home/antonio/coding/dpm/10/src/shared'

> Please don't start a new thread if you are still trying to solve the
> same problem:

Well, my problems have changed ....

> * Antonio Coralles wrote on Mon, Jan 24, 2005 at 05:12:30PM CET:
> > May there be a differnece between ranlib and  i686-pc-linux-gnu-ranlib
> > and/or ar and i686-pc-linux-gnu-ar ?
>
> That depends quite on what particular system you are on.  If that
> matches i686-pc-linux-gnu, there should be little difference.
> But you did not tell us that either.

Yes, it matches i686-pc-linux-gnu.

> > and if yes, how can i force
> > automake/autoconf to choose a particular - or better - in any case the
> > correct version ?
>
> Use --host=i686-pc-linux-gnu.  But I don't think you need this.
>
> Regards,
> Ralf

I hope that made things clearer ...
Regards,
antonio