Documentation request: If a user wants to make his own analysis area (using cvs), what layout should he create? And, of course, how does he neatly switch between CVSROOTs?
I took a look at this script and I think it is being run with
new packages, but it does not provide the information
I am looking for. It is what the physicist should do next,
after this level of preparation. The sort of information I'm looking for
is things like, for an effort referred to as "abc" should I put my
linkdef files in
rlc/dict/abc
or in
rlc/abc/dict
Should my include files be in
rlc/rlc/abc
or in
rlc/inc/abc
or in
rlc/abc/inc
or does it matter? What do I want in the makefiles if I want
the dictionaries objects in the same library or a different library?
How do I get my shared object to link with root? What goes
into makefile to run rootcint? What path does rootcint use?
How do LINK_SHARED and USESHLIBS interact? What do ROOTLIBS
LIBCCFILES sharedlib_o_dir and SUBDIRS do? When do I need vpath?
Of course to really explain all this you need
a huge document, and I bet some documents exist somewhere,
and at some level. I may have even read a few at some point
and I think I found they were out of date.
But my request is just for a somewhat
non-trivial example analysis package with a suggested organization.
Something that has a Stntuple analysis .so, a topntuple analysis .so,
and an AC++ exe, for a trigger study, for example. With some
comments in the makefiles.
And recall my request was to put this on the wish list -
we have survived long enough to show it is not urgent.
At the same time, I have passed my package as an example to
dozens of people and maybe we should have a thought-out example...
Ray
Modified 23-November-2010 at 14:29
http://hep.physics.wisc.edu/~jnb/cdfcode/23Nov2010
| Previous notes | Next notes | Main slide directory |