Recently in Compiling Category
Some problems found while porting sci20 to unix for the first time.
After several tries we give up using the provided "configure.py" scripts and decided to compile pyqt and pyqwt manually.
In the qmake makefile have a "config" syntax that allows to set variables based on the existence of a certain config. But, determining if the compilation is "release" or "debug" is not as trivial as it seems.
The Problem: Running coilib50 tests we find out that our "Reraise" function no longer work!
Describing some of the changes made in order to make it compile using python 2.5.1 and visual studio 2005
Symptom:
The following error while compiling qt4 (4.3.0) in amd64.redhat (using dist-0707, that is, gcc-4.2.0):
/home/users/ama/Shared/amd64.redhat/binutils-2.17/bin/ld:
.obj/release-shared/qscriptecmaboolean.o:
relocation R_X86_64_PC32 against
`QList<QScriptClassInfo*>::append(QScriptClassInfo* const&)'
can not be used when making a shared object;
recompile with -fPIC
/home/users/ama/Shared/amd64.redhat/binutils-2.17/bin/ld:
final link failed:
Bad value
The ETK stands for "ESSS TookKit" and it's a collection of ESSS libraries used in the project development.
The "creation" of ETK is actually the definition of a process where the common user no longer needs to download, compile and tests these libraries in order to work in its project.
The symptom:
The subversion executable (svn) didn't find the correct subversion dynamic libraries even changing the LD_LIBRARY_PATH.



