Re: [lmi] Patching severl xml-library defects

[Vadim Zeitlin]

On Mon, 23 Oct 2006 02:29:02 +0000 Greg Chicares <address@hidden> wrote:

GC> (1) libxslt-1.1.17/libxslt/xsltInternals.h
GC> 
GC> -     XSLT_FUNC_EXTENSION,
GC> +     XSLT_FUNC_EXTENSION
GC>   #ifdef XSLT_REFACTORED
GC> +                        ,
GC>       XSLT_FUNC_OTHERWISE,
GC>       XSLT_FUNC_FALLBACK,
GC>       XSLT_FUNC_MESSAGE,
GC>       XSLT_FUNC_INCLUDE,
GC>       XSLT_FUNC_ATTRSET,
GC>       XSLT_FUNC_LITERAL_RESULT_ELEMENT,
GC> -     XSLT_FUNC_UNKOWN_FORWARDS_COMPAT,
GC> +     XSLT_FUNC_UNKOWN_FORWARDS_COMPAT
GC>   #endif

 I'd suggest to fix this one differently: add a dummy XSLT_FUNC_MAX just
before the closing brace. This is somewhat easier to maintain than
staggered commas in the patch above.

GC> (2) libxml++/parsers/parser.h
GC> 
GC>   #ifdef WIN32 //TODO: Why do we do this? murrayc.
GC> - #pragma warning (disable : 4786)
GC> + //#pragma warning (disable : 4786)
GC>   #endif
GC> 
GC> Obviously the maintainers would be more likely to accept a
GC> patch that adds appropriate conditionals:
GC>   #ifdef MSVC_SOMETHING
GC>   ...
GC>   #endif // MSVC_SOMETHING

 This would be "#ifdef _MSC_VER" (instead of #ifdef WIN32). It's probably
also worth adding a comment stating what this warning is ("identifier was
truncated to 255 characters in the debug information") and why do we need
to disable it (because it's useless and is due to compiler limitation and
there is nothing we can do about it).

 Regards,
VZ