On 2009-04-24 09:32+0200 Werner Smekal wrote:

>>> 
>>> I would prefer the following logic instead:
>>> 
>>> if(PL_DOUBLE AND NOT DEFINED LIB_TAG)
>>> set(LIB_TAG "d")
>>> endif(PL_DOUBLE AND NOT DEFINED LIB_TAG)
>>> 
>>> In other words, if somebody specifies a library name suffix on the command
>>> line including the null string, let us assume they know what they want so
>>> that is exactly what they get.  However, if -DLIB_TAG is not specified at
>>> all by the user, then the suffix will be "d" for the usual (default)
>>> PL_DOUBLE case just like it is now.
>>> 
>>> Does any developer here mind this proposed change to the build-system
>>> logic?
>>> 
>> 
>> Hi Alan,
>> 
>> I think that is an excellent idea: it makes that feature much clearer
>> and more useful - it allows for people to develop a set of
>> configurations (should they need that), besides configurations that are
>> useful for debugging, also for use with static device drivers and so on.
>
> Many users are puzzled by the 'd' at the end if they compile a release build. 
> So this would be a good idea. But: The LIB_TAG is only used for the plplot 
> libraries, but not for nn, csa, cd .... So setting LIB_TAG helps the user to 
> have several versions of plplot side by side, but there are still conflicts 
> for the nn, csa, cd library. We must use the custom LIB_TAG in the 
> corresponding CMakeLists.txt files as well (but without the 'd').

The history here is "d" was introduced years ago simply as a reminder that
PLFLT was DOUBLE for the library in question.  But none of the libplplot
enhancement libraries (i.e., csironn, csirocsa, nistcd, or qsastime) use
PLFLT so LIB_TAG has not been used as a suffix for their names. So I agree
the present status is only a partial solution for what Vadim wanted.

Probably the above logic should be changed to

set(LIB_TAG_ENHANCED ${LIB_TAG})
if(PL_DOUBLE AND NOT DEFINED LIB_TAG)
   set(LIB_TAG "d")
endif(PL_DOUBLE AND NOT DEFINED LIB_TAG)

where LIB_TAG_ENHANCED is used for the name suffix of all the enhancement
libraries (requiring CMakeLists.txt changes a number of places as well as
making all xxxdll.h files for the enhancement libraries configurable).  The
result would be the enhancement libraries have the same suffix as libplplot
and other core libraries if LIB_TAG was specified by the user.  Otherwise
libplot and other core libraries would (by default) have a "d" suffix and
the enhancement libraries no suffix at all just like now.

As outlined above, addressing the library tag issue for the enhancement
libraries is straightforward but needs care since it touches quite a few
files.  If somebody else wants to deal with this issue before this release,
I would prefer it be done in the next day or so to insure thorough
pre-release testing next week. It's unlikely I will have time to implement
library tags for the enhancement libraries myself in the next day or so
since I am trying to finish up another PLplot project in the next day or so
for the same pre-release testing reason.

Vadim, while waiting for library tags for the enhancement libraries to be
dealt with (most likely after this forthcoming release), an excellent
alternative is simply to build and install your various types of PLplot
libraries using independent installation prefixes.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to