Re: [Plplot-general] Build documentation

[Alexis P. <ale...@ma...>]

Le 2013-11-14 20:19, Alan W. Irwin a écrit :
> On 2013-11-13 15:42+0100 Alexis Praga wrote:
> 
>>> Could you just check the Tcl version number in that file tcl.h?
>> 
>> It's the 8.5 version:
>> 
>> #define TCL_VERSION	    "8.5"
>> #define TCL_PATCH_LEVEL	    "8.5.9"
>> 
>> If you know how to activate a more verbose for CMake, I would be happy
>> to do it.
> 
> Hi Alexis:
> 
> I am joining this conversation late because I couldn't participate
> earlier (incoming mail-chain troubles that are now fixed).  I am also
> CCing Orion since he is very experienced at dealing with PLplot issues
> on Fedora. (He maintains the Fedora rpms for PLplot.)
> 
> Coincidentally, I have just finished (see my latest post to the
> plplot-devel list) a big effort on improving the parts of the build
> system having to do with Tcl, Tk, Itcl, Itk, and Iwidgets.  The result
> is more demanding of consistency for all components and also more
> verbose about what it finds.  So I would appreciate you giving the
> plplot svn trunk version a try to see if (a) it works on your
> Fedora system, and (b) it is verbose enough for you.
> 
> That version also has Arjen's fix concerning TclFormatInt.  I am
> virtually positive the fundamental issue there is the PLplot build
> historically relied on "private" Tcl headers. In 2000 or so when I was
> still building PLplot on a RedHat platform, I ran into a lot of such
> issues which Maurice (our Tcl/Tk guru who has mostly retired from
> working on PLplot now) fixed at the time.  But after I switched to
> Debian (whose Tcl devel package includes all private headers) a few of
> these issues may have crept back (because Tcl developers likely made
> different decisions about what was private and what not). So Arjen's
> fix is a good step in the right direction, and if you now try the svn
> trunk version of PLplot on Fedora you may find all is well now with
> the Tcl component of the build.  Or you might run into a few more
> private header issues which we would be happy to fix if you were
> willing to do the necessary build experiments.
> 
> I also have access to a Tcl8.6 build I have done myself which I am
> currently using to test PLplot's Tcl8.6 capabilities.  I notice that
> build has a switch that can be used to turn off the install of private
> headers.  So now you have reminded me of the private header issue with
> Tcl, I plan to use that switch to see if there is anything more to
> deal with.
> 
> 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); the Time
> Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> software package (plplot.sf.net); 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
> __________________________

Hi Alan,

I successfully built and installed the latest SVN (revision 12702). As a 
reference,
I have include the output of plplot-test.sh in share/examples, which 
looks OK
to me (only warnings).
Also, I consider it verbose enough for a normal install, even though the 
verbosity
would be more interesting had the build/install failed.

Finally, this version was validated by an install of PerlDL (and all of 
its testing
suite) and some plotting I did with it.

Cheers,

-- 
Alexis Praga
GPG key : 66F9 4BD4 70EC E5D5 2AA4  9097 0673 7CE2 4671 306F