index.html

<br>
<hr>
<a name="core"></a>
<h3>Core</h3>
<ul>
<li>
The ROOT source directory has been drastically reorganized and simplified.
Each source directory containing the code for a single library or plugin,
were all in the same top level directory, without logical grouping.
This made it difficult to see which libraries were related and belonged
together. In the new structure we've added a set of meta directories
that are used to group the sources of related libraries, and that also
reflects the ROOT team work package structure.
Note, the name and number of libraries has not been changed.
This new structure also facilitates the maintaining of the release notes
and other documentation items per meta package.
</li>
<li>
On Linux, MacOS X and Windows, there is no need anymore to define the
environment variable ROOTSYS. Internally ROOTSYS is set depending on the
location of the ROOT libraries. ROOTSYS was never needed when ROOT was
configured using --prefix.
</li>
<li>
On MacOS X when configure'ing with --enable-rpath (and not specifying
--prefix) the installation does not need (DY)LD_LIBRARY_PATH to be set
anymore. The installation is completely relocatable. The (DY)LD_LIBRARY_PATH
is determined relative to the location of the root executable.
</li>
<li>
On Windows .root files are now associated with the most recently executed
ROOT installation, i.e. run ROOT once and .root files open with ROOT.
</li>
<li>
Class TMessageHandler derives now from TQObject and does emit signals.
This allows for easier usage of this class.
</li>
<li>
In this release xrootd and libAfterImage are managed in Subversion
via so called "vendor branches". This is completely transparent
except for people who do directly use svn. You will get the message:
<pre>
   $ svn up
   svn: Failed to add directory 'xrootd/src/xrootd': object of the same name already exists
   $ rm -rf xrootd/src/xrootd
   $ svn up
   svn: Failed to add directory 'asimage/src/libAfterImage': object of the same name already exists
   $ rm -rf asimage/src/libAfterImage
   $ svn up
</pre>
</li>
<li>
Port to gcc 4.3.1. This version of gcc is much stricker with respect to
implicit header files so in many source files &lt;stdlib.h&gt; and &lt;string.h&gt;
had to be added.
</li>
<li>
TPRegexp
<ul>
<li>
Modularized Match() and Substitute() functions so that the low-level work
is done by MatchInternal() and SubstituteInternal().
</li>
<li>
Added function TString GetModifiers() that translates from bit-flag
options back to perl-style option characters.
</li>
<li>
For all functions that do actual matching replaced the default value
of 'Int_t nMatchMax' argument from 30 to 10. PCRE internals require
the index buffer to be 3-times the number of allowed matches. This
multiplication is now performed in individual functions and nMatchMax
is really the number of allowed matches. Two function calls in 
TString passing this number explicitly have been updated.
</li>
</ul>
</li>
<li>
TPMERegexp
<ul>
<li>
New sub-class of TPRegexp with API closer to PERL.
Supports main Perl operations using regular expressions (Match,
Substitute and Split). To retrieve the results one can simply use
operator[] returning a TString.
</li>
</ul>
</li>
<li>
New tutorial regexp_pme.C.
</li>
</ul>
<h4>Meta</h4>
<ul>
<li>Add a new TClass interface to be able to trigger the "auto addition of object
to a TDirectory object".  If a class has a member function:
<pre>        DirectoryAutoAdd(TDirectory*)</pre>
it will now be accessible (when the dictionary is generated via rootcint 
for now) via TClass::GetDirectoryAutoAdd which return a wrapper with the 
signature:<pre>
        void (*)(void *this_obj,TDirectory *where_to_add)
</pre>
</li><li>Extend <b><i>#pragma create TClass;</i></b> to namespaces.</li>
<li>Enable autoloading of the cintdlls.</li>
<li>rlibmap now handles <b><i>#pragma create TClass;</i></b> statements.</li>
</ul>
<h4>Cont</h4>
<ul>
<li>
Support for std algorithms and iterators for ROOT collection classes by Anar.
The background is the following. While working on the PROOF code I found
that enumerating TList is an inconvenient and a long operation, I had to
write the same code all over the place and make duplications. I tried
to use STD algorithms with it, namely std::for_each, and failed.
I therefore decided to enable std::for_each algorithm for ROOT 
Containers/Iterators by making as few as possible changes, without rewriting
iterators at all. Now with only two simple lines of code one is able to
iterate through a container:
<pre>
   TIter iter(&amp;list);
   for_each(iter.Begin(), TIter::End(), SEnumFunctor());
</pre>
or
<pre>
   for_each(iter.Begin(), inter_end, SEnumFunctor());
</pre>
where iter_end could be an iterator to a middle of the container.
<p>
After I had changed Iterators so that they could be used with std::for_each,
I decided to go further and did some more changes. As a result,
<ul>
<li>
   I have updated CINT implementation of some algorithms (they look now
   more or less better in terms of the standard),
</li>
<li>
   All collections can be now used with std::for_each, std::find_if,
   std::count_if (probably with some more algorithms. I've listed here
   only what has been *checked* by me).
</li>
</ul>
<li>
A test program has been added: $ROOTSYS/test/stressIterators.cxx
</li>
<li>
A tutorial macro has been added: $ROOTSYS/tutorials/cont/TListAndSTL.C
</li>
</ul>