Hi Christian,
I did ci Antonio's proposed patch, but if you have other ideas let me know.
Also the doc generation script would be nice to have. Can you contribute it?
Cheers, Fons.
Christian Holm Christensen wrote:
> Hi Antonio,
>
> On Tue, 2005-04-05 at 18:58 +0100, Antonio Amorim wrote:
>
>>Dear Colleagues,
>>
>> I have successfully installed all new root debian packages (4-3-3 from
>>the cvs head) using also the cernlib and generator packages from Kevin
>>http://borex.princeton.edu/~kmccarty/ to solve the package dependencies.
>> I have found two small problems that are easy to fix:
>>1- the libmysqlclient-dev dependency in the file
>>build/package/lib/makebuilddepend.sh should be replaced by :
>>libmysqlclient-dev | libmysqlclient12-dev | libmysqlclient14-dev
>>
>>For example, I am using MySql 4.1 and have libmysqlclient14-dev instaled.
>
>
> Hmm. libmysqlclient12-dev provides libmysqlclient-dev, and I think it's
> a bug that libmysqlclient14-dev doesn't. Of course, the versioned
> dependency will work too. However, I'm not sure it's what the policy
> dictates. Did you try to run lintian or linda on the generated
> packages?
>
>
>>2- The building of the ROOT class reference manual.
>>I know that a tar file is available for this manual but, since I am
>>compiling from the cvs head, I wanted to avoid any inconsistencies by
>>make the class reference html files from the source.
>> make html
>>invokes the script build/unix/makehtml.sh that fails and I think is
>>incomplete.
>>One has to load all the shared libraries to create the reference manual.
>>I have created a file that at least runs without problems. (The root
>>developers can certainly figure out the dependencies much better).
>>
>>I include the command bellow:
>>
>>bin/root -b <<makedoc
>>gSystem.Load("libHistPainter");
>>gSystem.Load("libTreePlayer");
>>gSystem.Load("libPhysics");
>>gSystem.Load("libThread");
>>gSystem.Load("libEG");
>>gSystem.Load("libEGPythia6");
>>gSystem.Load("libHist.so");
>>gSystem.Load("libRLDAP.so");
>>gSystem.Load("libHtml.so");
>>gSystem.Load("libSrvAuth.so");
>>gSystem.Load("libKrb5Auth.so");
>>gSystem.Load("libMatrix.so");
>>gSystem.Load("libMySQL.so");
>>gSystem.Load("libPgSQL.so");
>>gSystem.Load("libTree.so");
>>gSystem.Load("libMinuit.so");
>>gSystem.Load("libFumili.so");
>>gSystem.Load("libXMLIO.so");
>>gSystem.Load("libGX11.so");
>>gSystem.Load("libGX11TTF.so");
>>gSystem.Load("libGui.so");
>>gSystem.Load("libGuiBld.so");
>>gSystem.Load("libGraf.so");
>>gSystem.Load("libGraf3d.so");
>>gSystem.Load("libASImage.so");
>>gSystem.Load("libRint.so");
>>gSystem.Load("libGed.so");
>>gSystem.Load("libGedOld.so");
>>gSystem.Load("libGeom.so");
>>gSystem.Load("libGeomPainter.so");
>>gSystem.Load("libTreeViewer.so");
>>gSystem.Load("libMLP.so");
>>gSystem.Load("libX3d.so");
>>gSystem.Load("libNew.so");
>>gSystem.Load("libGpad.so");
>>gSystem.Load("libXMLParser.so");
>>gSystem.Load("libPostscript.so");
>>gSystem.Load("libXrdSeckrb4.so");
>>gSystem.Load("libXrdSeckrb5.so");
>>gSystem.Load("libXrdSec.so");
>>gSystem.Load("libRGL.so");
>>gSystem.Load("libHbook.so");
>>gSystem.Load("libNetx.so");
>>gSystem.Load("libVMC.so");
>>gSystem.Load("libPyROOT.so");
>>gSystem.Load("libQuadp.so");
>>
>>THtml html;
>>html.MakeAll();
>>.q
>>makedoc
>
>
> Perhaps this should be generated by `configure' based on the
> `enable_<foo>' variables? So that the script would be something like
>
> // Skeleton HTML doc file - html.C.in
> ...
> Bool_t hasPyRoot = @enable_python@;
> ...
> void html()
> {
> ...
> if (hasPyRoot) gSystem->Load("libPyRoot.so");
> ...
> THtml htmlGenerator;
> htmlGenerator.MakeAll();
> }
> // EOF
>
> and then do the substitutions like
>
> sed ... \
> -e "s,@enable_python@,`test x$enable_python = "xyes" && echo kTRUE
> || echo kFALSE"` \
> ...
>
> Or the script could simply loop over all libraries in the `lib'
> directory and load them.
>
>
>>Also it would be nice to have a flag that avoids creating the pdf
>>Inheritance files to speed up the process.
>
>
> How come these inheritance file are created as PDFs rather than HTML?
> And, the isn't the inheritance graphs in the beginning of the class
> documentation enough?
>
> Yours,
>
--
Org: CERN, European Laboratory for Particle Physics.
Mail: 1211 Geneve 23, Switzerland
E-Mail: Fons.Rademakers_at_cern.ch Phone: +41 22 7679248
WWW: http://www.rademakers.org/fons/ Fax: +41 22 7679480
Received on Wed Apr 06 2005 - 14:43:00 MEST