[geos-devel] GEOS libraries
pramsey at refractions.net
Fri Aug 30 14:00:10 EDT 2002
I have a CVS sandbox with Normans changes integrated ready to commit. I
have not integrated his changes to the actual code, which commented out
alot of code which failed to compile under UNIX. I assume we'll fix that
in a different pass.
Is there any reason I should not commit up these changes?
They do two things: (a) add a bunch of Makefile.am files to the
directory structure (no harm) and (b) move all the testing files out of
the /source directory and into /source/test
Speak now or forever hold your peace...
Yury A. Bychkov wrote:
> Hi everyone,
> Me and Martin have just finished discussing it and here is what we think:
>>I have a quick question
>>about the library structure. The most natural source tree configuration
>>for automake is one library per directory - based on this I would
>>produce three libraries, libgsgeo (from source in source/geom), libgsgrp
>>(source/graph) and libgsio (source/io). I assume the code in
>>subdirectories of these (source/io/markup, source/graph/index) would
>>link into the corresponding upper-level library. Everyone ok with this
>>(names and arrangements)?
> There are way to many rather complex interdependencies for it to work (even
> 'graph/index' is called from outside of 'graph'). It seems to us that everything
> should be put in one library.
>>Also, is the code in util/ needed into any of the libs? If so, which one?
> 'util' contains Assert class that is used almost everywhere.
>>Yury - do you want me to fix the trivial quirks about the code (like the
>>need for including stdio.h for sprintf)?
> I'd prefer if you'll just send me an email and tell me what you want me to
> geos-devel mailing list
> geos-devel at geos.refractions.net
| Paul Ramsey
| Refractions Research
| Email: pramsey at refractions.net
| Phone: (250) 885-0632
More information about the geos-devel