If you get the latest sources from CVS, all of these issues with iterators
will go away. If you don't want to do that, then do the following to fix
the problem:
&*theEntry->m_Name.begin()
Wherever there's a problem with the assumption that a vector iterator is a
pointer, this will do the trick. You can always rely on a vector being
implemented as a contiguous array. It's not guaranteed by the C++ standard
as originally published, but was fixed fairly quickly in an erratum.
As far as a new Xalan release is concerned, we'ed like to do one this
month, but we're trying to figure out what the current state of Xerces is.
Right now, we can't do a new Xalan with Xerces 1.5 as there are too many
problems (bugs and huge memory leaks), but we really don't want to do the
new release with Xerces 1.4, only to have a new Xerces release several
weeks later.
Dave
Holger Fl�rke
<floerke@doct To: [EMAIL PROTECTED]
ronic.de> cc: (bcc: David N Bertoni/CAM/Lotus)
Subject: Xalan-C++ and GCC3.0 - New
Release?
06/20/2001
05:28 AM
Please
respond to
xalan-dev
Hi there,
I would like to compile the Xalan-C1_1 release with the new GCC3.0
(released this monday). Before I spent much time to fix some compile
errors, I would like to know whether there will be a new xalan release this
month.
First I found a bad casting in PlatformSupport/AttrListImpl.cpp:174:
// Add the item...
tempVector.push_back(
getNewEntry(
theEntry->m_Name.begin(),
theEntry->m_Type.begin(),
theEntry->m_Value.begin()));
As far as I see, there is a cast from 'vector<XMLCh>::const_iterator' to
'XMLCh*'. Wahh. Can you rely on the std::vector being implemented as an
array and the iterator as a pointer to the elements?
The same thing happens in many other places (FormatterToHTML.cpp:1100,
FormatterToHTML:1110, XalanArrayKeyMap.hpp:227,...)
Is there a good and general way to handle this incompatibilities?
HolgeR
--
holger floerke d o c t r o n i c
email [EMAIL PROTECTED] information publishing + retrieval
phone +49 (0) 2222 9292 90 http://www.doctronic.de
fax +49 (0) 2222 9292 99
