Status of trunk / main-3.3 comparison

[Cantor, Scott]

I've completed the comparison in both directions and nothing really new has popped up as a concern, just the two trunk changes I identified (repeated here for emphasis):

r1517488 (XERCESC-2016)
r1528170 (XERCESC-2019)

There are a number of fixes to port up to the trunk, at least one needing a deeper dive because of a reported regression on the branch, but that was in my queue to do anyway.

And there are a few things that need to be done regardless and the biggest one [1] should be simpler to do on the trunk anyway. If anybody who knows the DOM impl could weigh in on what the heck that code in DOMCasts.hpp/cpp was doing and why, I'd welcome the perspective.

I'll be starting to commit to trunk soon, likely this week, and try and get some time to review the CMake patch proposal.

-- Scott

[1] https://issues.apache.org/jira/browse/XERCESC-2088


---------------------------------------------------------------------
To unsubscribe, e-mail: c-dev-***@xerces.apache.org
For additional commands, e-mail: c-dev-***@xerces.apache.org

[Boris Kolpackov]

Hi Scott,

Post by Cantor, Scott
If anybody who knows the DOM impl could weigh in on what the heck that
code in DOMCasts.hpp/cpp was doing and why, I'd welcome the perspective.

Yeah, infamous Xerces-C++ brain-death ;-)

Try to replace this (and if it works other similar) function with:

template <typename T>
static inline DOMNodeImpl*
castToNodeImpl(const T* p)
{
return &p->fNode;
}

There might still be some cases you you will have to cast the argument
to appropriate DOM*Impl type.

Boris

---------------------------------------------------------------------
To unsubscribe, e-mail: c-dev-***@xerces.apache.org
For additional commands, e-mail: c-dev-***@xerces.apache.org

[Cantor, Scott]

Post by Boris Kolpackov
template <typename T>
static inline DOMNodeImpl*
castToNodeImpl(const T* p)
{
return &p->fNode;
}

Yeah, I was considering but wasn't sure if we wanted to add templates in, are we using them anywhere else in the code base? Anyway, that seems like a reasonable suggestion certainly, and is one reason I was preferring to do this on trunk.

Post by Boris Kolpackov
There might still be some cases you you will have to cast the argument
to appropriate DOM*Impl type.

I think so, yeah, there are quite a few of these casting helpers.

Thanks,
-- Scott



---------------------------------------------------------------------
To unsubscribe, e-mail: c-dev-***@xerces.apache.org
For additional commands, e-mail: c-dev-***@xerces.apache.org